We present BootComp, a novel framework based on text-to-image diffusion models for controllable human image generation with multiple reference garments. Here, the main bottleneck is data acquisition for training: collecting a large-scale dataset of high-quality reference garment images per human subject is quite challenging, i.e., ideally, one needs to manually gather every single garment photograph worn by each human. To address this, we propose a data generation pipeline to construct a large synthetic dataset, consisting of human and multiple-garment pairs, by introducing a model to extract any reference garment images from each human image. To ensure data quality, we also propose a filtering strategy to remove undesirable generated data based on measuring perceptual similarities between the garment presented in human image and extracted garment. Finally, by utilizing the constructed synthetic dataset, we train a diffusion model having two parallel denoising paths that use multiple garment images as conditions to generate human images while preserving their fine-grained details. We further show the wide-applicability of our framework by adapting it to different types of reference-based generation in the fashion domain, including virtual try-on, and controllable human image generation with other conditions, e.g., pose, face, etc.

Efficient tokenization of videos remains a challenge in training vision models that can process long videos. One promising direction is to develop a tokenizer that can encode long video clips, as it would enable the tokenizer to leverage the temporal coherence of videos better for tokenization. However, training existing tokenizers on long videos often incurs a huge training cost as they are trained to reconstruct all the frames at once. In this paper, we introduce CoordTok, a video tokenizer that learns a mapping from coordinate-based representations to the corresponding patches of input videos, inspired by recent advances in 3D generative models. In particular, CoordTok encodes a video into factorized triplane representations and reconstructs patches that correspond to randomly sampled (x, y, t) coordinates. This allows for training large tokenizer models directly on long videos without requiring excessive training resources. Our experiments show that CoordTok can drastically reduce the number of tokens for encoding long video clips. For instance, CoordTok can encode a 128-frame video with 128 × 128 resolution into 1280 tokens, while baselines need 6144 or 8192 tokens to achieve similar reconstruction quality. We further show that this efficient video tokenization enables memory-efficient training of a diffusion transformer that can generate 128 frames at once.

Aligning text-to-image (T2I) diffusion models with preference optimization is valuable for human-annotated datasets, but the heavy cost of manual data collection limits scalability. Using reward models offers an alternative, however, current preference optimization methods fall short in exploiting the rich information, as they only consider pairwise preference distribution. Furthermore, they lack generalization to multi-preference scenarios and struggle to handle inconsistencies between rewards. To address this, we present Calibrated Preference Optimization (CaPO), a novel method to align T2I diffusion models by incorporating the general preference from multiple reward models without human annotated data. The core of our approach involves a reward calibration method to approximate the general preference by computing the expected win-rate against the samples generated by the pretrained models. Additionally, we propose a frontier-based pair selection method that effectively manages the multi-preference distribution by selecting pairs from Pareto frontiers. Finally, we use regression loss to fine-tune diffusion models to match the difference between calibrated rewards of a selected pair. Experimental results show that CaPO consistently outperforms prior methods, such as Direct Preference Optimization (DPO), in both single and multi-reward settings validated by evaluation on T2I benchmarks, including GenEval and T2I-Compbench.

Learning the visual representation via masked auto-encoder (MAE) training has been proven to be a powerful technique. Transferring the pre-trained vision transformer (ViT) to downstream tasks leads to superior performance compared to conventional task-by-task supervised learning. Recent research works on MAE focus on large-sized vision transformers (>50 million parameters) with outstanding performance. However, improving the generality of the under-parametrized lightweight model has been widely ignored. In practice, downstream applications are commonly intended for resource-constrained platforms, where large-scale ViT cannot easily meet the resource budget. Current lightweight MAE training heavily relies on knowledge distillation with a pre-trained teacher, whereas the root cause behind the poor performance remains under-explored. Motivated by that, this paper first introduces the concept of "closest neighbor patch" to characterize the local semantics among the input tokens. Our discovery shows that the lightweight model failed to distinguish different local information, leading to aliased understanding and poor accuracy. Motivated by this finding, we propose NoR-MAE, a novel MAE training algorithm for lightweight vision transformers. NoR-MAE elegantly repels the semantic aliasing between patches and their closest neighboring patch (semantic centroid) with negligible training cost overhead. With the ViT-Tiny model, NoR-MAE achieves up to 7.22%/3.64% accuracy improvements on ImageNet-100/ImageNet-1K datasets, as well as up to 5.13% accuracy improvements in tested downstream tasks. https://github.com/SeoLabCornell/NoR-MAE

Personalization in language models aims to tailor model behavior to individual users or user groups.Prompt-based methods incorporate user preferences into queries, while trainingbased methods encode them into model parameters.Model merging has also been explored for personalization under limited data.However, existing methods often fail to directly optimize task-specific utility and lack explicit mechanisms for privacy preservation.To address the limitations, we propose Privacy-Preserving Model Merging via Evolutionary Algorithms (PriME), a novel personalization approach that employs gradient-free methods to directly optimize utility while reducing privacy risks.By integrating privacy preservation into the optimization objective, PriME creates personalized modules that effectively capture target user preferences while minimizing privacy risks for data-sharing users.Experiments on the LaMP benchmark show that PriME consistently outperforms a range of baselines, achieving up to a 45% improvement in task performance.Further analysis demonstrates that PriME achieves a superior privacy-utility trade-off compared to a prior state-of-the-art, with enhanced robustness to membership inference attacks and greater utility in capturing user preferences.

Woomin Song, Saket Dingliwal, Sai Muralidhar Jayanthi, Bhavana Ganesh, Jinwoo Shin, Aram Galstyan, Sravan Babu Bodapati. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Recent advances in diffusion models have introduced a new era of text-guided image manipulation, enabling users to create realistic edited images with simple textual prompts. However, there is significant concern about the potential misuse of these methods, especially in creating misleading or harmful content. Although recent defense strategies, which introduce imperceptible adversarial noise to induce model failure, have shown promise, they remain ineffective against more sophisticated manipulations, such as editing with a mask. In this work, we propose DiffusionGuard, a robust and effective defense method against unauthorized edits by diffusion-based image editing models, even in challenging setups. Through a detailed analysis of these models, we introduce a novel objective that generates adversarial noise targeting the early stage of the diffusion process. This approach significantly improves the efficiency and effectiveness of adversarial noises. We also introduce a mask-augmentation technique to enhance robustness against various masks during test time. Finally, we introduce a comprehensive benchmark designed to evaluate the effectiveness and robustness of methods in protecting against privacy threats in realistic scenarios. Through extensive experiments, we show that our method achieves stronger protection and improved mask robustness with lower computational costs compared to the strongest baseline. Additionally, our method exhibits superior transferability and better resilience to noise removal techniques compared to all baseline methods. Our source code is publicly available at https://choi403.github.io/diffusionguard.

Reinforcement Learning (RL) agents have demonstrated their potential across various robotic tasks. However, they still heavily rely on human-engineered reward functions, requiring extensive trial-and-error and access to target behavior information, often unavailable in real-world settings. This paper introduces REDS: REward learning from Demonstration with Segmentations, a novel reward learning framework that leverages action-free videos with minimal supervision. Specifically, REDS employs video demonstrations segmented into subtasks from diverse sources and treats these segments as ground-truth rewards. We train a dense reward function conditioned on video segments and their corresponding subtasks to ensure alignment with ground-truth reward signals by minimizing the Equivalent-Policy Invariant Comparison distance. Additionally, we employ contrastive learning objectives to align video representations with subtasks, ensuring precise subtask inference during online interactions. Our experiments show that REDS significantly outperforms baseline methods on complex robotic manipulation tasks in Meta-World and more challenging real-world tasks, such as furniture assembly in FurnitureBench, with minimal human intervention. Moreover, REDS facilitates generalization to unseen tasks and robot embodiments, highlighting its potential for scalable deployment in diverse environments.

Existing realistic blur datasets provide insufficient variety in scenes and blur patterns to be trained, while expanding data diversity demands considerable time and effort due to complex dual-camera systems. To address the challenge, data augmentation can be an effective way to artificially increase data diversity. However, existing methods on this line are typically designed to estimate motions from a 2D perspective, e.g., estimating 2D non-uniform kernels disregarding 3D aspects of blur modeling, which leads to unrealistic motion patterns due to the fact that camera and object motions inherently arise in 3D space. In this paper, we propose a 3D-aware blur synthesizer capable of generating diverse and realistic blur images for blur data augmentation. Specifically, we estimate 3D camera positions within the motion blur interval, generate the corresponding scene images, and aggregate them to synthesize a realistic blur image. Since the 3D camera positions projected onto the 2D image plane inherently lie in 2D space, we can represent the 3D transformation as a combination of 2D transformation and projected 3D residual component. This allows for 3D transformation without requiring explicit depth measurements, as the 3D residual component is directly estimated via a neural network. Furthermore, our blur synthesizer allows for controllable blur data augmentation by modifying blur magnitude, direction, and scenes, resulting in diverse blur images. As a result, our method significantly improves deblurring performance, making it more practical for real-world scenarios.

Aligning large language models (LLMs) with human preferences becomes a key component to obtaining state-of-the-art performance, but it yields a huge cost to construct a large human-annotated preference dataset. To tackle this problem, we propose a new framework, Spread Preference Annotation with direct preference judgment (SPA), that boosts the alignment of LLMs using only a very small amount of human-annotated preference data. Our key idea is leveraging the human prior knowledge within the small (seed) data and progressively improving the alignment of LLM, by iteratively generating the responses and learning from them with the self-annotated preference data. To be specific, we propose to derive the preference label from the logits of LLM to explicitly extract the model's inherent preference. Compared to the previous approaches using external reward models or implicit in-context learning, we observe that the proposed approach is significantly more effective. In addition, we introduce a noise-aware preference learning algorithm to mitigate the risk of low quality within generated preference data. Our experimental results demonstrate that the proposed framework significantly boosts the alignment of LLMs. For example, we achieve superior alignment performance on AlpacaEval 2.0 with only 3.3% of the ground-truth preference labels in the Ultrafeedback data compared to the cases using the entire data or state-of-the-art baselines.

Recent advances in large language models (LLMs) have led to a growing interest in developing LLM-based agents for automating web tasks. However, these agents often struggle with even simple tasks on real-world websites due to their limited capability to understand and process complex web page structures. In this work, we introduce LCoW, a framework for Learning language models to Contextualize complex Web pages into a more comprehensible form, thereby enhancing decision making by LLM agents. LCoW decouples web page understanding from decision making by training a separate contextualization module to transform complex web pages into comprehensible format, which are then utilized by the decision-making agent. We demonstrate that our contextualization module effectively integrates with LLM agents of various scales to significantly enhance their decision-making capabilities in web automation tasks. Notably, LCoW improves the success rates of closed-source LLMs (e.g., Gemini-1.5-flash, GPT-4o, Claude-3.5-Sonnet) by an average of 15.6%, and demonstrates a 23.7% average improvement in success rates for open-source LMs (e.g., Llama-3.1-8B, Llama-3.1-70B) on the WorkArena benchmark. Moreover, the Gemini-1.5-flash agent with LCoW achieves state-of-the-art results on the WebShop benchmark, outperforming human experts. The relevant code materials are available at our project page: https://lcowiclr2025.github.io.

Recent studies have shown that the denoising process in (generative) diffusion models can induce meaningful (discriminative) representations inside the model, though the quality of these representations still lags behind those learned through recent self-supervised learning methods. We argue that one main bottleneck in training large-scale diffusion models for generation lies in effectively learning these representations. Moreover, training can be made easier by incorporating high-quality external visual representations, rather than relying solely on the diffusion models to learn them independently. We study this by introducing a straightforward regularization called REPresentation Alignment (REPA), which aligns the projections of noisy input hidden states in denoising networks with clean image representations obtained from external, pretrained visual encoders. The results are striking: our simple strategy yields significant improvements in both training efficiency and generation quality when applied to popular diffusion and flow-based transformers, such as DiTs and SiTs. For instance, our method can speed up SiT training by over 17.5$\times$, matching the performance (without classifier-free guidance) of a SiT-XL model trained for 7M steps in less than 400K steps. In terms of final generation quality, our approach achieves state-of-the-art results of FID=1.42 using classifier-free guidance with the guidance interval.

Data imbalance remains a fundamental challenge in real-world machine learning. However, most existing work has focused on classification, leaving imbalanced regression underexplored despite its importance in many applications. To address this gap, we propose PRIME, a framework that leverages learnable proxies to construct a balanced and well-ordered feature space for imbalanced regression. At its core, PRIME arranges proxies to be uniformly distributed in the feature space while preserving the ordinal structure of regression targets, and then aligns each sample feature to its corresponding proxy. By using proxies as reference points, PRIME induces the desired structure of learned representations, promoting better generalization, especially in underrepresented target regions. Moreover, since proxy-based alignment resembles classification, PRIME enables the seamless application of class imbalance techniques to regression, facilitating more balanced feature learning. Extensive experiments demonstrate the effectiveness and broad applicability of PRIME, achieving state-of-the-art performance on four real-world regression benchmark datasets across diverse target domains.

Selecting a layer normalization (LN) strategy that stabilizes training and speeds convergence in Transformers remains difficult, even for today’s large language models (LLM). We present a comprehensive analytical foundation for understanding how different LN strategies influence training dynamics in large-scale Transformers. Until recently, Pre-LN and Post-LN have long dominated practices despite their limitations in large-scale training. However, several open-source models have recently begun silently adopting a third strategy without much explanation. This strategy places normalization layer peripherally around sublayers, a design we term Peri-LN. While Peri-LN has demonstrated promising performance, its precise mechanisms and benefits remain almost unexplored. Our in-depth analysis delineates the distinct behaviors of LN strategies, showing how each placement shapes activation variance and gradient propagation. To validate our theoretical insight, we conduct extensive experiments on Transformers up to $3.2$B parameters, showing that Peri-LN consistently achieves more balanced variance growth, steadier gradient flow, and convergence stability. Our results suggest that Peri-LN warrants broader consideration for large-scale Transformer architectures, providing renewed insights into the optimal placement of LN.

Deep learning models perform poorly when domain shifts exist between training and test data. Test-time adaptation (TTA) is a paradigm to mitigate this issue by adapting pre-trained models using only unlabeled test samples. However, existing TTA methods can fail under severe domain shifts, while recent active TTA approaches requiring full-class labels are impractical due to high labeling costs. To address this issue, we introduce a new setting of TTA with binary feedback, which uses a few binary feedbacks from annotators to indicate whether model predictions are correct, thereby significantly reducing the labeling burden of annotators. Under the setting, we propose BiTTA, a novel dual-path optimization framework that leverages reinforcement learning to balance binary feedback-guided adaptation on uncertain samples with agreement-based self-adaptation on confident predictions. Experiments show BiTTA achieves substantial accuracy improvements over state-of-the-art baselines, demonstrating its effectiveness in handling severe distribution shifts with minimal labeling effort.

Self-awareness, i.e., the ability to assess and correct one's generation, is a fundamental aspect of human intelligence, making its replication in large language models (LLMs) an important yet challenging task. Previous works tackle this by employing extensive reinforcement learning or relying on large external verifiers. In this work, we propose Refine via Intrinsic Self-Verification (ReVISE), an efficient and effective framework that enables LLMs to self-correct their outputs through self-verification. The core idea of ReVISE is to enable LLMs to verify their reasoning processes and continually rethink reasoning trajectories based on its verification. To implement this efficiently, we introduce a structured curriculum based on preference learning. Specifically, as ReVISE involves two challenging tasks (i.e., self-verification and reasoning correction), we tackle each task sequentially using curriculum learning, collecting both failed and successful reasoning paths to construct preference pairs for efficient training. During inference, our approach enjoys natural test-time scaling by integrating self-verification and correction capabilities, further enhanced by our proposed confidence-aware decoding mechanism. Our experiments on various reasoning tasks demonstrate that ReVISE achieves efficient self-correction and significantly improves the reasoning performance of LLMs.

Learning robust representations from data often requires scale, which has led to the success of recent zero-shot models such as CLIP. However, the obtained robustness can easily be deteriorated when these models are fine-tuned on other downstream tasks (e.g., of smaller scales). Previous works often interpret this phenomenon in the context of domain shift, developing fine-tuning methods that aim to preserve the original domain as much as possible. However, in a different context, fine-tuned models with limited data are also prone to learning features that are spurious to humans, such as background or texture. In this paper, we propose StarFT (Spurious Textual Alignment Regularization), a novel framework for fine-tuning zero-shot models to enhance robustness by preventing them from learning spuriosity. We introduce a regularization that aligns the output distribution for spuriosity-injected labels with the original zero-shot model, ensuring that the model is not induced to extract irrelevant features further from these descriptions. We leverage recent language models to get such spuriosity-injected labels by generating alternative textual descriptions that highlight potentially confounding features. Extensive experiments validate the robust generalization of StarFT and its emerging properties: zero-shot group robustness and improved zero-shot classification. Notably, StarFT boosts both worst-group and average accuracy by 14.30% and 3.02%, respectively, in the Waterbirds group shift scenario, where other robust fine-tuning baselines show even degraded performance.

Large language models (LLMs) are increasingly used in sensitive domains, where their ability to infer personal data from seemingly benign text introduces emerging privacy risks. While recent LLM-based anonymization methods help mitigate such risks, they often rely on proprietary models (e.g., GPT-4), raising concerns about cost and the potential exposure of sensitive data to untrusted external systems. To address this, we introduce $\textit{SElf-refining Anonymization with Language model}$ (SEAL), a novel distillation framework for training small language models (SLMs) to perform effective anonymization without relying on external models at inference time. SEAL leverages adversarial interactions between an LLM anonymizer and an inference model to collect trajectories of anonymized texts and inferred attributes, which are then used to distill anonymization and critique capabilities into SLMs through supervised fine-tuning and preference learning. The resulting models learn both to anonymize text and to evaluate their outputs, enabling iterative improvement of anonymization quality via self-refinement. Experiments on SynthPAI, a dataset of synthetic personal profiles and text comments, demonstrate that SLMs trained with SEAL achieve substantial improvements in anonymization capabilities. Notably, 8B models attain a privacy-utility trade-off comparable to that of the GPT-4 anonymizer and, with self-refinement, even surpass it in terms of privacy protection. These results highlight the effectiveness of our adversarial distillation framework for training SLMs as efficient anonymizers.

Large Vision-Language Models (LVLMs) have recently shown great promise in advancing robotics by combining embodied reasoning with robot control. A common approach involves training on embodied reasoning tasks related to robot control using Supervised Fine-Tuning (SFT). However, SFT datasets are often heuristically constructed and not explicitly optimized for improving robot control. Furthermore, SFT often leads to issues such as catastrophic forgetting and reduced generalization performance. To address these limitations, we introduce Robot-R1, a novel framework that leverages reinforcement learning to enhance embodied reasoning specifically for robot control. Robot-R1 learns to predict the next keypoint state required for task completion, conditioned on the current scene image and environment metadata derived from expert demonstrations. Inspired by the DeepSeek-R1 learning approach, Robot-R1 samples reasoning-based responses and reinforces those that lead to more accurate predictions. Our experiments show that models trained with Robot-R1 outperform SFT methods on embodied reasoning tasks. Despite having only 7B parameters, Robot-R1 even surpasses GPT-4o on reasoning tasks related to low-level action control, such as spatial and movement reasoning.

Agents based on large language models (LLMs) for machine learning engineering (MLE) can automatically implement ML models via code generation. However, existing approaches to build such agents often rely heavily on inherent LLM knowledge and employ coarse exploration strategies that modify the entire code structure at once. This limits their ability to select effective task-specific models and perform deep exploration within specific components, such as experimenting extensively with feature engineering options. To overcome these, we propose MLE-STAR, a novel approach to build MLE agents. MLE-STAR first leverages external knowledge by using a search engine to retrieve effective models from the web, forming an initial solution, then iteratively refines it by exploring various strategies targeting specific ML components. This exploration is guided by ablation studies analyzing the impact of individual code blocks. Furthermore, we introduce a novel ensembling method using an effective strategy suggested by MLE-STAR. Our experimental results show that MLE-STAR achieves medals in 64% of the Kaggle competitions on the MLE-bench, significantly outperforming the best alternative.

Understanding the 3D world from 2D monocular videos is a crucial ability for AI. Recently, to tackle this underdetermined task, end-to-end 3D geometry priors have been sought after, such as pre-trained point map models at scale. These models enable robust 3D understanding from casually taken videos, providing accurate object shapes disentangled from uncertain camera parameters. However, they still struggle when affected by object deformation and dynamics, failing to establish consistent correspondence over the frames. Furthermore, their architectures are typically limited to pairwise frame processing, which is insufficient for capturing complex motion dynamics over extended sequences. To address these limitations, we introduce Track3R, a novel framework that integrates a new architecture and task to jointly predict point map and motion trajectories across multiple frames from video input. Specifically, our key idea is modeling two disentangled trajectories for each point: one representing object motion and the other camera poses. This design not only can enable understanding of the 3D object dynamics, but also facilitates the learning of more robust priors for 3D shapes in dynamic scenes. In our experiments, Track3R demonstrates significant improvements in a joint point mapping and 3D motion estimation task for dynamic scenes, such as 25.8% improvements in the motion estimation, and 15.7% in the point mapping accuracy.

As large language models increasingly gain popularity in real-world applications, processing extremely long contexts, often exceeding the model’s pre-trained context limits, has emerged as a critical challenge. While existing approaches to efficient long-context processing show promise, recurrent compression-based methods struggle with information preservation, whereas random access approaches require substantial memory resources. We introduce REFORM, a novel inference framework that efficiently handles long contexts through a two-phase approach. First, it incrementally processes input chunks while maintaining a compressed KV cache, constructs cross-layer context embeddings, and utilizes early exit strategy for improved efficiency. Second, it identifies and gathers essential tokens via similarity matching and selectively recomputes the KV cache. Compared to baselines, REFORM achieves over 50% and 27% performance gains on RULER and BABILong respectively at 1M context length. It also outperforms baselines on ∞-Bench, RepoEval, and MM-NIAH, demonstrating flexibility across diverse tasks and domains. Additionally, REFORM reduces inference time by 30% and peak memory usage by 5%, achieving both efficiency and superior performance.

Federated learning (FL) enables distributed training while preserving data privacy, but stragglers—slow or incapable clients can significantly slow down the total training time and degrade performance. To mitigate the impact of stragglers, system heterogeneity, including heterogeneous computing and network bandwidth, has been addressed. While previous studies have addressed system heterogeneity by splitting models into submodels, they offer limited flexibility in model architecture design, without considering potential inconsistencies arising from training multiple submodel architectures. We proposenested federated learning (NeFL), a generalized framework that efficiently divides deep neural networks into submodels using both depthwise and widthwise scaling. To address theinconsistencyarising from training multiple submodel architectures, NeFL decouples a subset of parameters from those being trained for each submodel. An averaging method is proposed to handle these decoupled parameters during aggregation. NeFL enables resource-constrained devices to effectively participate in the FL pipeline, facilitating larger datasets for model training. Experiments demonstrate that NeFL achieves performance gain, especially for the worst-case submodel compared to baseline approaches (7.63% improvement on CIFAR-100). Furthermore, NeFL aligns with recent advances in FL, such as leveraging pre-trained models and accounting for statistical heterogeneity.

As recent advances in mobile camera technology have enabled the capability to capture high-resolution images, such as 4K images, the demand for an efficient deblurring model handling large motion has increased. In this paper, we discover that the image residual errors, i.e., blur-sharp pixel differences, can be grouped into some categories according to their motion blur type and how complex their neighboring pixels are. Inspired by this, we decompose the deblurring (regression) task into blur pixel discretization (pixel-level blur classification) and discrete-to-continuous conversion (regression with blur class map) tasks. Specifically, we generate the discretized image residual errors by identifying the blur pixels and then transform them to a continuous form, which is computationally more efficient than naively solving the original regression problem with continuous values. Here, we found that the discretization result, i.e., blur segmentation map, remarkably exhibits visual similarity with the image residual errors. As a result, our efficient model shows comparable performance to state-of-the-art methods in realistic benchmarks, while our method is up to 10 times computationally more efficient.

Addressing biases in computer vision models is crucial for real-world AI deployments. However, mitigating visual biases is challenging due to their unexplainable nature, often identified indirectly through visualization or sample statistics, which necessitates additional human supervision for interpretation. To tackle this issue, we propose the Bias-to-Text (B2T) framework, which interprets visual biases as keywords. Specifically, we extract common keywords from the captions of mispredicted images to identify potential biases in the model. We then validate these keywords by measuring their similarity to the mispredicted images using a vision-language scoring model. The keyword explanation form of visual bias offers several advantages, such as a clear group naming for bias discovery and a natural extension for debiasing using these group names. Our experiments demonstrate that B2T can identify known biases, such as gender bias in CelebA, background bias in Waterbirds, and distribution shifts in ImageNet-R/C. Additionally, B2T uncovers novel biases in larger datasets, such as Dollar Street and ImageNet. For example, we discovered a contextual bias between “bee” and “flower” in ImageNet. We also highlight various applications of B2T keywords, including debiased training, CLIP prompting, and model comparison.11Code: https://github.com/alinlab/b2t

Active learning, a paradigm within machine learning, aims to select and query unlabeled data to enhance model performance strategically. A crucial selection strategy leverages the model's predictive uncertainty, reflecting the informativeness of a data point. While the sample's distance to the decision boundary intuitively measures predictive uncertainty, its computation becomes intractable for complex decision boundaries formed in multiclass classification tasks. This paper introduces the *least disagree metric* (LDM), the smallest probability of predicted label disagreement. We propose an asymptotically consistent estimator for LDM under mild assumptions. The estimator boasts computational efficiency and straightforward implementation for deep learning models using parameter perturbation. The LDM-based active learning algorithm queries unlabeled data with the smallest LDM, achieving state-of-the-art *overall* performance across various datasets and deep architectures, as demonstrated by the experimental results.

Fine-tuning text-to-image models with reward functions trained on human feedback data has proven effective for aligning model behavior with human intent. However, excessive optimization with such reward models, which serve as mere proxy objectives, can compromise the performance of fine-tuned models, a phenomenon known as reward overoptimization. To investigate this issue in depth, we introduce the Text-Image Alignment Assessment (TIA2) benchmark, which comprises a diverse collection of text prompts, images, and human annotations. Our evaluation of several state-of-the-art reward models on this benchmark reveals their frequent misalignment with human assessment. We empirically demonstrate that overoptimization occurs notably when a poorly aligned reward model is used as the fine-tuning objective. To address this, we propose TextNorm, a simple method that enhances alignment based on a measure of reward model confidence estimated across a set of semantically contrastive text prompts. We demonstrate that incorporating the confidence-calibrated rewards in fine-tuning effectively reduces overoptimization, resulting in twice as many wins in human evaluation for text-image alignment compared against the baseline reward models.

Large language models (LLMs) have made significant advancements in various natural language processing tasks, including question answering (QA) tasks. While incorporating new information with the retrieval of relevant passages is a promising way to improve QA with LLMs, the existing methods often require additional fine-tuning which becomes infeasible with recent LLMs. Augmenting retrieved passages via prompting has the potential to address this limitation, but this direction has been limitedly explored. To this end, we design a simple yet effective framework to enhance open-domain QA (ODQA) with LLMs, based on the summarized retrieval (SuRe). SuRe helps LLMs predict more accurate answers for a given question, which are well-supported by the summarized retrieval that could be viewed as an explicit rationale extracted from the retrieved passages. Specifically, SuRe first constructs summaries of the retrieved passages for each of the multiple answer candidates. Then, SuRe confirms the most plausible answer from the candidate set by evaluating the validity and ranking of the generated summaries. Experimental results on diverse ODQA benchmarks demonstrate the superiority of SuRe, with improvements of up to 4.6\% in exact match (EM) and 4.0\% in F1 score over standard prompting approaches. SuRe also can be integrated with a broad range of retrieval methods and LLMs. Finally, the generated summaries from SuRe show additional advantages to measure the importance of retrieved passages and serve as more preferred rationales by models and humans.

Recent progress in text-to-3D generation has been achieved through the utilization of score distillation methods: they make use of the pre-trained text-to-image (T2I) diffusion models by distilling via the diffusion model training objective. However, such an approach inevitably results in the use of random timesteps at each update, which increases the variance of the gradient and ultimately prolongs the optimization process. In this paper, we propose to enhance the text-to-3D optimization by leveraging the T2I diffusion prior in the generative sampling process with a predetermined timestep schedule. To this end, we interpret text-to-3D optimization as a multi-view image-to-image translation problem, and propose a solution by approximating the probability flow. By leveraging the proposed novel optimization algorithm, we design DreamFlow, a practical three-stage coarse-to-fine text-to-3D optimization framework that enables fast generation of high-quality and high-resolution (i.e., 1024×1024) 3D contents. For example, we demonstrate that DreamFlow is 5 times faster than the existing state-of-the-art text-to-3D method, while producing more photorealistic 3D contents.

Large language models (LLMs) have shown remarkable performance in various natural language processing tasks. However, a primary constraint they face is the context limit, i.e., the maximum number of tokens they can process. Previous works have explored architectural changes and modifications in positional encoding to relax the constraint, but they often require expensive training or do not address the computational demands of self-attention. In this paper, we present Hierarchical cOntext MERging (HOMER), a new training-free scheme designed to overcome the limitations. HOMER uses a divide-and-conquer algorithm, dividing long inputs into manageable chunks. Each chunk is then processed collectively, employing a hierarchical strategy that merges adjacent chunks at progressive transformer layers. A token reduction technique precedes each merging, ensuring memory usage efficiency. We also propose an optimized computational order reducing the memory requirement to logarithmically scale with respect to input length, making it especially favorable for environments with tight memory restrictions. Our experiments demonstrate the proposed method's superior performance and memory efficiency, enabling the broader use of LLMs in contexts requiring extended context. Code is available at https://github.com/alinlab/HOMER.

Video diffusion models have recently made great progress in generation quality, but are still limited by the high memory and computational requirements. This is because current video diffusion models often attempt to process high-dimensional videos directly. To tackle this issue, we propose content-motion latent diffusion model (CMD), a novel efficient extension of pretrained image diffusion models for video generation. Specifically, we propose an autoencoder that succinctly encodes a video as a combination of a content frame (like an image) and a low-dimensional motion latent representation. The former represents the common content, and the latter represents the underlying motion in the video, respectively. We generate the content frame by fine-tuning a pretrained image diffusion model, and we generate the motion latent representation by training a new lightweight diffusion model. A key innovation here is the design of a compact latent space that can directly utilizes a pretrained image diffusion model, which has not been done in previous latent video diffusion models. This leads to considerably better quality generation and reduced computational costs. For instance, CMD can sample a video 7.7$\times$ faster than prior approaches by generating a video of 512$\times$1024 resolution and length 16 in 3.1 seconds. Moreover, CMD achieves an FVD score of 238.3 on WebVid-10M, 18.5% better than the previous state-of-the-art of 292.4.

Self-supervised learning of image representations by predicting future frames is a promising direction but still remains a challenge. This is because of the under-determined nature of frame prediction; multiple potential futures can arise from a single current frame. To tackle this challenge, in this paper, we revisit the idea of stochastic video generation that learns to capture uncertainty in frame prediction and explore its effectiveness for representation learning. Specifically, we design a framework that trains a stochastic frame prediction model to learn temporal information between frames. Moreover, to learn dense information within each frame, we introduce an auxiliary masked image modeling objective along with a shared decoder architecture. We find this architecture allows for combining both objectives in a synergistic and compute-efficient manner. We demonstrate the effectiveness of our framework on a variety of tasks from video label propagation and vision-based robot learning domains, such as video segmentation, pose tracking, vision-based robotic locomotion, and manipulation tasks. Code is available on the project webpage: https://sites.google.com/view/2024rsp.

Developing an effective molecular generation framework even with a limited number of molecules is often important for its practical deployment, e.g., drug discovery, since acquiring task-related molecular data requires expensive and time-consuming experimental costs. To tackle this issue, we introduce Hierarchical Textual Inversion for Molecular Generation (HI-Mol), a novel data-efficient molecular generation method. HI-Mol is inspired by the importance of hierarchical information, e.g., both coarse- and fine-grained features, in understanding the molecule distribution. We propose to use multi-level embeddings to reflect such hierarchical features based on the adoption of the recent textual inversion technique in the visual domain, which achieves data-efficient image generation. Compared to the conventional textual inversion method in the image domain using a single-level token embedding, our multi-level token embeddings allow the model to effectively learn the underlying low-shot molecule distribution. We then generate molecules based on the interpolation of the multi-level token embeddings. Extensive experiments demonstrate the superiority of HI-Mol with notable data-efficiency. For instance, on QM9, HI-Mol outperforms the prior state-of-the-art method with 50x less training data. We also show the effectiveness of molecules generated by HI-Mol in low-shot molecular property prediction. Code is available at https://github.com/Seojin-Kim/HI-Mol.

Diffusion models have demonstrated remarkable success in various domains, including molecular generation. However, conditional molecular generation remains a fundamental challenge due to an intrinsic trade-off between targeting specific chemical properties and generating meaningful samples from the data distribution. In this work, we present Time-Aware Conditional Synthesis (TACS), a novel approach to conditional generation on diffusion models. It integrates adaptively controlled plug-and-play "online" guidance into a diffusion model, driving samples toward the desired properties while maintaining validity and stability. A key component of our algorithm is our new type of diffusion sampler, Time Correction Sampler (TCS), which is used to control guidance and ensure that the generated molecules remain on the correct manifold at each reverse step of the diffusion process at the same time. Our proposed method demonstrates significant performance in conditional 3D molecular generation and offers a promising approach towards inverse molecular design, potentially facilitating advancements in drug discovery, materials science, and other related fields.

Text-to-image (T2I) diffusion models, when fine-tuned on a few personal images, can generate visuals with a high degree of consistency. However, such fine-tuned models are not robust; they often fail to compose with concepts of pretrained model or other fine-tuned models. To address this, we propose a novel fine-tuning objective, dubbed Direct Consistency Optimization, which controls the deviation between fine-tuning and pretrained models to retain the pretrained knowledge during fine-tuning. Through extensive experiments on subject and style customization, we demonstrate that our method positions itself on a superior Pareto frontier between subject (or style) consistency and image-text alignment over all previous baselines; it not only outperforms regular fine-tuning objective in image-text alignment, but also shows higher fidelity to the reference images than the method that fine-tunes with additional prior dataset. More importantly, the models fine-tuned with our method can be merged without interference, allowing us to generate custom subjects in a custom style by composing separately customized subject and style models. Notably, we show that our approach achieves better prompt fidelity and subject fidelity than those post-optimized for merging regular fine-tuned models.

The remarkable capabilities and easy accessibility of large language models (LLMs) have significantly increased societal risks (e.g., fake news generation), necessitating the development of LLM-generated text (LGT) detection methods for safe usage. However, detecting LGTs is challenging due to the vast number of LLMs, making it impractical to account for each LLM individually; hence, it is crucial to identify the common characteristics shared by these models. In this paper, we draw attention to a common feature of recent powerful LLMs, namely the alignment training, i.e., training LLMs to generate human-preferable texts. Our key finding is that as these aligned LLMs are trained to maximize the human preferences, they generate texts with higher estimated preferences even than human-written texts; thus, such texts are easily detected by using the reward model (i.e., an LLM trained to model human preference distribution). Based on this finding, we propose two training schemes to further improve the detection ability of the reward model, namely (i) continual preference fine-tuning to make reward model prefer aligned LGTs even further and (ii) reward modeling of Human/LLM mixed texts (a rephrased texts from human-written texts using aligned LLMs), which serves as a median preference text corpus between LGTs and human-written texts to learn the decision boundary better. We provide an extensive evaluation by considering six text domains across twelve aligned LLMs, where our method demonstrates state-of-the-art results.

In tabular prediction tasks, tree-based models combined with automated feature engineering methods often outperform deep learning approaches that rely on learned representations. While these feature engineering techniques are effective, they typically depend on a pre-defined search space and primarily use validation scores for feature selection, thereby missing valuable insights from previous experiments. To address these limitations, we propose a novel tabular learning framework that utilizes large language models (LLMs), termed Optimizing Column feature generator with decision Tree reasoning (OCTree). Our key idea is to leverage the reasoning capabilities of LLMs to identify effective feature generation rules without manually specifying the search space and provide language-based reasoning information highlighting past experiments as feedback for iterative rule improvements. We use decision trees to convey this reasoning information, as they can be easily represented in natural language, effectively providing knowledge from prior experiments (i.e., the impact of the generated features on performance) to the LLMs. Our empirical results demonstrate that OCTree consistently enhances the performance of various prediction models across diverse benchmarks, outperforming competing automated feature engineering methods. Code is available at https://github.com/jaehyun513/OCTree.

Tracking points in video frames is essential for understanding video content. However, the task is fundamentally hindered by the computation demands for brute-force correspondence matching across the frames. As the current models down-sample the frame resolutions to mitigate this challenge, they fall short in accurately representing point trajectories due to information truncation. Instead, we address the challenge by pruning the search space for point tracking and let the model process only the important regions of the frames without down-sampling. Our first key idea is to identify the object instance and its trajectory over the frames, then prune the regions of the frame that do not contain the instance. Concretely, to estimate the instance’s trajectory, we track a group of points on the instance and aggregate their motion trajectories. Furthermore, to deal with the occlusions in complex scenes, we propose to compensate for the occluded points while tracking. To this end, we introduce a unified framework that jointly performs point tracking and segmentation, providing synergistic effects between the two tasks. For example, the segmentation results enable a tracking model to avoid the occluded points referring to the instance mask, and conversely, the improved tracking results can help to produce more accurate segmentation masks. Our framework can be easily incorporated with various tracking models, and we demonstrate its efficacy for enhanced point tracking throughout extensive experiments. For example, on the recent TAP-Vid benchmark, our framework consistently improves all baselines, e.g., up to 13.5% improvement on the average Jaccard metric.

Due to the rapid generation and dissemination of information, large language models (LLMs) quickly run out of date despite enormous development costs. To address the crucial need to keep models updated, online learning has emerged as a critical tool when utilizing LLMs for real-world applications. However, given the ever-expanding corpus of unseen documents and the large parameter space of modern LLMs, efficient adaptation is essential. To address these challenges, we propose Memory of Amortized Contexts (MAC), an efficient and effective online adaptation framework for LLMs with strong knowledge retention. We propose a feature extraction and memory-augmentation approach to compress and extract information from new documents into compact modulations stored in a memory bank. When answering questions, our model attends to and extracts relevant knowledge from this memory bank. To learn informative modulations in an efficient manner, we utilize amortization-based meta-learning, which substitutes an otherwise required optimization process with a single forward pass of the encoder. Subsequently, we learn to choose from and aggregate selected documents into a single modulation by conditioning on the question, allowing us to adapt a frozen language model during test time without requiring further gradient updates. Our experiment demonstrates the superiority of MAC in multiple aspects, including online adaptation performance, time, and memory efficiency. In addition, we show how MAC can be combined with and improve the performance of popular alternatives such as retrieval augmented generations (RAGs). Code is available at: https://github.com/jihoontack/MAC.

Any classifier can be "smoothed out" under Gaussian noise to build a new classifier that is provably robust to l2-adversarial perturbations, viz., by averaging its predictions over the noise via randomized smoothing. Under the smoothed classifiers, the fundamental trade-off between accuracy and (adversarial) robustness has been well evidenced in the literature: i.e., increasing the robustness of a classifier for an input can be at the expense of decreased accuracy for some other inputs. In this paper, we propose a simple training method leveraging this trade-off to obtain robust smoothed classifiers, in particular, through a sample-wise control of robustness over the training samples. We make this control feasible by using "accuracy under Gaussian noise" as an easy-to-compute proxy of adversarial robustness for an input. Specifically, we differentiate the training objective depending on this proxy to filter out samples that are unlikely to benefit from the worst-case (adversarial) objective. Our experiments show that the proposed method, despite its simplicity, consistently exhibits improved certified robustness upon state-of-the-art training methods. Somewhat surprisingly, we find these improvements persist even for other notions of robustness, e.g., to various types of common corruptions. Code is available at https://github.com/alinlab/smoothing-catrs.

The success of NLP systems often relies on the availability of large, high-quality datasets.However, not all samples in these datasets are equally valuable for learning, as some may be redundant or noisy.Several methods for characterizing datasets based on modeldriven meta-information (e.g., model's confidence) have been developed, but the relationship and complementary effects of these methods have received less attention.In this paper, we introduce infoVerse, a universal framework for dataset characterization, which provides a new feature space that effectively captures multidimensional characteristics of datasets by incorporating various model-driven metainformation.infoVerse reveals distinctive regions of the dataset that are not apparent in the original semantic space, hence guiding users (or models) in identifying which samples to focus on for exploration, assessment, or annotation.Additionally, we propose a novel sampling method on infoVerse to select a set of data points that maximizes informativeness.In three real-world applications (data pruning, active learning, and data annotation), the samples chosen on infoVerse space consistently outperform strong baselines in all applications.Our code and demo are publicly available.

We study contextual linear bandit problems under feature uncertainty; they are noisy with missing entries. To address the challenges of the noise, we analyze Bayesian oracles given observed noisy features. Our Bayesian analysis finds that the optimal hypothesis can be far from the underlying realizability function, depending on the noise characteristics, which are highly non-intuitive and do not occur for classical noiseless setups. This implies that classical approaches cannot guarantee a non-trivial regret bound. Therefore, we propose an algorithm that aims at the Bayesian oracle from observed information under this model, achieving $\tilde{O}(d\sqrt{T})$ regret bound when there is a large number of arms. We demonstrate the proposed algorithm using synthetic and real-world datasets.

Neural networks are often prone to bias toward spurious correlations inherent in a dataset, thus failing to generalize unbiased test criteria. A key challenge to resolving the issue is the significant lack of bias-conflicting training data (i. e., samples without spurious correlations). In this paper, we propose a novel data augmentation approach termed Bias-Adversarial augmentation (BiasAdv) that supplements bias-conflicting samples with adversarial images. Our key idea is that an adversarial attack on a biased model that makes decisions based on spurious correlations may generate syn-thetic bias-conflicting samples, which can then be used as augmented training data for learning a debiased model. Specifically, we formulate an optimization problem for gen-erating adversarial images that attack the predictions of an auxiliary biased model without ruining the predictions of the desired debiased model. Despite its simplicity, we find that BiasAdv can generate surprisingly useful synthetic bias-conflicting samples, allowing the debiased model to learn generalizable representations. Furthermore, BiasAdv does not require any bias annotations or prior knowledge of the bias type, which enables its broad applicability to existing debiasing methods to improve their performances. Our extensive experimental results demonstrate the superiority of BiasAdv, achieving state-of-the-art performance on four popular benchmark datasets across various bias domains.

In practical scenarios where training data is limited, many predictive signals in the data can be rather from some biases in data acquisition (i.e., less generalizable), so that one cannot prevent a model from co-adapting on such (so-called) “shortcut” signals: this makes the model fragile in various distribution shifts. To bypass such failure modes, we consider an adversarial threat model under a mutual information constraint to cover a wider class of perturbations in training. This motivates us to extend the standard information bottleneck to additionally model the nuisance information. We propose an autoencoder-based training to implement the objective, as well as practical encoder designs to facilitate the proposed hybrid discriminative-generative training concerning both convolutional- and Transformer-based architectures. Our experimental results show that the proposed scheme improves robustness of learned representations (remarkably without using any domain-specific knowledge), with respect to multiple challenging reliability measures. For example, our model could advance the state-of-the-art on a recent challenging OBJECTS benchmark in novelty detection by 78.4% → 87.2% in AUROC, while simultaneously enjoying improved corruption, background and (certified) adversarial robustness. Code is available at https://github.com/jh-jeong/nuisance_ib.

Despite the remarkable progress in deep generative models, synthesizing high-resolution and temporally coherent videos still remains a challenge due to their high-dimensionality and complex temporal dynamics along with large spatial variations. Recent works on diffusion models have shown their potential to solve this challenge, yet they suffer from severe computation and memory-inefficiency that limit the scalability. To handle this issue, we propose a novel generative model for videos, coined projected latent video diffusion model (PVDM), a probabilistic diffusion model which learns a video distribution in a low-dimensional latent space and thus can be efficiently trained with high-resolution videos under limited resources. Specifically, PVDM is composed of two components: (a) an autoencoder that projects a given video as 2D-shaped latent vectors that factorize the complex cubic structure of video pixels and (b) a diffusion model architecture specialized for our new factorized latent space and the training/sampling procedure to synthesize videos of arbitrary length with a single model. Experiments on popular video generation datasets demonstrate the superiority of PVDM compared with previous video synthesis methods; e.g., PVDM obtains the FVD score of 639.7 on the UCF-101 long video (128 frames) generation benchmark, which improves 1773.4 of the prior state-of-the-art.

Vision-language (VL) pre-training has recently gained much attention for its transferability and flexibility in novel concepts (e.g., cross-modality transfer) across various visual tasks. However, VL-driven segmentation has been under-explored, and the existing approaches still have the burden of acquiring additional training images or even segmentation annotations to adapt a VL model to downstream segmentation tasks. In this paper, we introduce a novel image-free segmentation task where the goal is to perform semantic segmentation given only a set of the target semantic categories, but without any task-specific images and annotations. To tackle this challenging task, our proposed method, coined IFSeg, generates VL-driven artificial image-segmentation pairs and updates a pretrained VL model to a segmentation task. We construct this artificial training data by creating a 2D map of random semantic categories and another map of their corresponding word tokens. Given that a pretrained VL model projects visual and text tokens into a common space where tokens that share the semantics are located closely, this artificially generated word map can replace the real image inputs for such a VL model. Through an extensive set of experiments, our model not only establishes an effective baseline for this novel task but also demonstrates strong performances compared to existing methods that rely on stronger supervision, such as task-specific images and segmentation masks. Code is available at https://github.com/alinlab/ifseg.

In this study, we investigate the problem of string-based molecular generation via variational autoencoders (VAEs) that have served a popular generative approach for various tasks in artificial intelligence. Our main idea is to maintain multiple decoders while sharing a single encoder, i.e., it is a type of ensemble techniques. Here, we first found that training each decoder independently may not be effective as the bias of the ensemble decoder increases severely under its auto-regressive inference. To alleviate this issue, our proposed technique is two-fold: (a) a different latent variable is sampled for each decoder (from estimated mean and variance offered by the shared encoder) to encourage diverse characteristics of decoders and (b) a collaborative loss is used during training to control the aggregated quality of decoders using different latent variables. In our experiments, the proposed VAE model particularly performs well for generating a sample from out-of-domain distribution.

The development of largely human-annotated benchmarks has driven the success of deep neural networks in various NLP tasks. To enhance the effectiveness of existing benchmarks, collecting new additional input-output pairs is often too costly and challenging, particularly considering their marginal impact on improving the current model accuracy. Instead, additional or complementary annotations on the existing input texts in the benchmarks can be preferable as an efficient way to pay the additional human cost. In this paper, we investigate task-specific preferences between pairs of input texts as a new alternative way for such auxiliary data annotation. From pair-wise comparisons with respect to the task, the auxiliary preference learning enables the model to learn an additional informative training signal that cannot be captured with instance-wise task labels. To this end, we propose a novel multi-task learning framework, called prefer-to-classify (P2C), which can enjoy the cooperative effect of learning both the given classification task and the auxiliary preferences. Here, we provide three different ways to collect preference signals in practice: (a) implicitly extracting from annotation records (for free, but often unavailable), (b) collecting explicitly from crowd workers (high paid), or (c) pre-trained large language models such as GPT-3 (low paid). Given existing classification NLP benchmarks, we demonstrate that the proposed auxiliary preference learning via P2C on them is effective in improving text classifiers. Our codes are publicly available.

We introduce a modality-agnostic neural compression algorithm based on a functional view of data and parameterised as an Implicit Neural Representation (INR). Bridging the gap between latent coding and sparsity, we obtain compact latent representations non-linearly mapped to a soft gating mechanism. This allows the specialisation of a shared INR network to each data item through subnetwork selection. After obtaining a dataset of such latent representations, we directly optimise the rate/distortion trade-off in a modality-agnostic space using neural compression. Variational Compression of Implicit Neural Representations (VC-INR) shows improved performance given the same representational capacity pre quantisation while also outperforming previous quantisation schemes used for other INR techniques.Our experiments demonstrate strong results over a large set of diverse modalities using the same algorithm without any modality-specific inductive biases. We show results on images, climate data, 3D shapes and scenes as well as audio and video, introducing VC-INR as the first INR-based method to outperform codecs as well-known and diverse as JPEG 2000, MP3 and AVC/HEVC on their respective modalities.

Visual robotic manipulation research and applications often use multiple cameras, or views, to better perceive the world. How else can we utilize the richness of multi-view data? In this paper, we investigate how to learn good representations with multi-view data and utilize them for visual robotic manipulation. Specifically, we train a multi-view masked autoencoder which reconstructs pixels of randomly masked viewpoints and then learn a world model operating on the representations from the autoencoder. We demonstrate the effectiveness of our method in a range of scenarios, including multi-view control and single-view control with auxiliary cameras for representation learning. We also show that the multi-view masked autoencoder trained with multiple randomized viewpoints enables training a policy with strong viewpoint randomization and transferring the policy to solve real-robot tasks without camera calibration and an adaptation procedure. Video demonstrations are available at: https://sites.google.com/view/mv-mwm.

Despite the potential of vision-based personal monitoring (e.g., healthcare), private data leakage concerns hinder its wide deployment in personal spaces (e.g., bedrooms). A body of data anonymization designs was proposed throughout image processing and federated learning. They commonly store high-quality images and videos locally, which are anonymized via post-processing before cloud upload. However, the recent IoT camera hacking and local data leakage call for anonymized data at the sensing stage. Also, continuous and pervasive monitoring without blind spots in complicated indoor spaces requires a scalable and economic system. This paper present Mosaic, a vision-based end-to-end action recognition framework that (i) intrinsically achieves data anonymity from the sensing stage and (ii) battery-free operation for blind spot-free continuous monitoring. Mosaic leverages an extremely low resolution (eLR) Near-Infrared (NIR) image sensor with 6 × 10 pixels for video anonymity and RFID-compliant fully-passive tag with four solar cells for real-time eLR video streaming under as low as 50 lux (e.g., deep in the shelf without direct light). This is accompanied by light-weight action recognition neural network for real-time inference (18.4ms on Intel(R) Core i7-8700). Mosaic achieves an average of 98% accuracy on 10 action classes, hitting the balance between data anonymity and high-precision action recognition. By taking advantage of NIR (non-visible) frequency, Mosaic also works in dark without disturbing sleep. Lastly, wildfire detection reaching 20m was demonstrated, showcasing the potential for outdoor monitoring.

Despite its practical importance across a wide range of modalities, recent advances in self-supervised learning (SSL) have been primarily focused on a few well-curated domains, e.g., vision and language, often relying on their domain-specific knowledge. For example, Masked Auto-Encoder (MAE) has become one of the popular architectures in these domains, but less has explored its potential in other modalities. In this paper, we develop MAE as a unified, modality-agnostic SSL framework. In turn, we argue meta-learning as a key to interpreting MAE as a modality-agnostic learner, and propose enhancements to MAE from the motivation to jointly improve its SSL across diverse modalities, coined MetaMAE as a result. Our key idea is to view the mask reconstruction of MAE as a meta-learning task: masked tokens are predicted by adapting the Transformer meta-learner through the amortization of unmasked tokens. Based on this novel interpretation, we propose to integrate two advanced meta-learning techniques. First, we adapt the amortized latent of the Transformer encoder using gradient-based meta-learning to enhance the reconstruction. Then, we maximize the alignment between amortized and adapted latents through task contrastive learning which guides the Transformer encoder to better encode the task-specific knowledge. Our experiment demonstrates the superiority of MetaMAE in the modality-agnostic SSL benchmark (called DABS), significantly outperforming prior baselines.

Along with recent diffusion models, randomized smoothing has become one of a few tangible approaches that offers adversarial robustness to models at scale, e.g., those of large pre-trained models. Specifically, one can perform randomized smoothing on any classifier via a simple "denoise-and-classify" pipeline, so-called denoised smoothing, given that an accurate denoiser is available - such as diffusion model. In this paper, we present scalable methods to address the current trade-off between certified robustness and accuracy in denoised smoothing. Our key idea is to "selectively" apply smoothing among multiple noise scales, coined multi-scale smoothing, which can be efficiently implemented with a single diffusion model. This approach also suggests a new objective to compare the collective robustness of multi-scale smoothed classifiers, and questions which representation of diffusion model would maximize the objective. To address this, we propose to further fine-tune diffusion model (a) to perform consistent denoising whenever the original image is recoverable, but (b) to generate rather diverse outputs otherwise. Our experiments show that the proposed multi-scale smoothing scheme, combined with diffusion fine-tuning, not only allows strong certified robustness at high noise scales but also maintains accuracy close to non-smoothed classifiers. Code is available at https://github.com/jh-jeong/smoothing-multiscale.

Generative priors of large-scale text-to-image diffusion models enable a wide range of new generation and editing applications on diverse visual modalities. However, when adapting these priors to complex visual modalities, often represented as multiple images (e.g., video or 3D scene), achieving consistency across a set of images is challenging. In this paper, we address this challenge with a novel method, Collaborative Score Distillation (CSD). CSD is based on the Stein Variational Gradient Descent (SVGD). Specifically, we propose to consider multiple samples as “particles” in the SVGD update and combine their score functions to distill generative priors over a set of images synchronously. Thus, CSD facilitates the seamless integration of information across 2D images, leading to a consistent visual synthesis across multiple samples. We show the effectiveness of CSD in a variety of editing tasks, encompassing the visual editing of panorama images, videos, and 3D scenes. Our results underline the competency of CSD as a versatile method for enhancing inter-sample consistency, thereby broadening the applicability of text-to-image diffusion models.

A promising technique for exploration is to maximize the entropy of visited state distribution, i.e., state entropy, by encouraging uniform coverage of visited state space. While it has been effective for an unsupervised setup, it tends to struggle in a supervised setup with a task reward, where an agent prefers to visit high-value states to exploit the task reward. Such a preference can cause an imbalance between the distributions of high-value states and low-value states, which biases exploration towards low-value state regions as a result of the state entropy increasing when the distribution becomes more uniform. This issue is exacerbated when high-value states are narrowly distributed within the state space, making it difficult for the agent to complete the tasks. In this paper, we present a novel exploration technique that maximizes the value-conditional state entropy, which separately estimates the state entropies that are conditioned on the value estimates of each state, then maximizes their average. By only considering the visited states with similar value estimates for computing the intrinsic bonus, our method prevents the distribution of low-value states from affecting exploration around high-value states, and vice versa. We demonstrate that the proposed alternative to the state entropy baseline significantly accelerates various reinforcement learning algorithms across a variety of tasks within MiniGrid, DeepMind Control Suite, and Meta-World benchmarks. Source code is available at https://sites.google.com/view/rl-vcse.

Developing an agent capable of adapting to unseen environments remains a difficult challenge in imitation learning. This work presents Adaptive Return-conditioned Policy (ARP), an efficient framework designed to enhance the agent's generalization ability using natural language task descriptions and pre-trained multimodal encoders. Our key idea is to calculate a similarity between visual observations and natural language instructions in the pre-trained multimodal embedding space (such as CLIP) and use it as a reward signal. We then train a return-conditioned policy using expert demonstrations labeled with multimodal rewards. Because the multimodal rewards provide adaptive signals at each timestep, our ARP effectively mitigates the goal misgeneralization. This results in superior generalization performances even when faced with unseen text instructions, compared to existing text-conditioned policies. To improve the quality of rewards, we also introduce a fine-tuning method for pre-trained multimodal encoders, further enhancing the performance. Video demonstrations and source code are available on the project website: \url{https://sites.google.com/view/2023arp}.

Contrastive learning (CL) has been widely investigated with various learning mechanisms and achieves strong capability in learning representations of data in a self-supervised manner using unlabeled data. A common fashion of contrastive learning on this line is employing mega-sized encoders to achieve comparable performance as the supervised learning counterpart. Despite the success of the labelless training, current contrastive learning algorithms *failed* to achieve good performance with lightweight (compact) models, e.g., MobileNet, while the requirements of the heavy encoders impede the energy-efficient computation, especially for resource-constrained AI applications. Motivated by this, we propose a new self-supervised CL scheme, named SACL-XD, consisting of two technical components, **S**limmed **A**symmetrical **C**ontrastive **L**earning (SACL) and **Cross**-**D**istillation (XD), which collectively enable efficient CL with compact models. While relevant prior works employed a strong pre-trained model as the teacher of unsupervised knowledge distillation to a lightweight encoder, our proposed method trains CL models from scratch and outperforms them even without such an expensive requirement. Compared to the SoTA lightweight CL training (distillation) algorithms, SACL-XD achieves 1.79% ImageNet-1K accuracy improvement on MobileNet-V3 with 64$\times$ training FLOPs reduction.

Vision-language models, such as contrastive language-image pre-training (CLIP), have demonstrated impressive results in natural image domains. However, these models often struggle when applied to specialized domains like remote sensing, and adapting to such domains is challenging due to the limited number of image-text pairs available for training. To address this, we propose S-CLIP, a semi-supervised learning method for training CLIP that utilizes additional unpaired images. S-CLIP employs two pseudo-labeling strategies specifically designed for contrastive learning and the language modality. The caption-level pseudo-label is given by a combination of captions of paired images, obtained by solving an optimal transport problem between unpaired and paired images. The keyword-level pseudo-label is given by a keyword in the caption of the nearest paired image, trained through partial label learning that assumes a candidate set of labels for supervision instead of the exact one. By combining these objectives, S-CLIP significantly enhances the training of CLIP using only a few image-text pairs, as demonstrated in various specialist domains, including remote sensing, fashion, scientific figures, and comics. For instance, S-CLIP improves CLIP by 10% for zero-shot classification and 4% for image-text retrieval on the remote sensing benchmark, matching the performance of supervised CLIP while using three times fewer image-text pairs.

We introduce an efficient optimization-based meta-learning technique for large-scale neural field training by realizing significant memory savings through automated online context point selection. This is achieved by focusing each learning step on the subset of data with the highest expected immediate improvement in model quality, resulting in the almost instantaneous modeling of global structure and subsequent refinement of high-frequency details. We further improve the quality of our meta-learned initialization by introducing a bootstrap correction resulting in the minimization of any error introduced by reduced context sets while simultaneously mitigating the well-known myopia of optimization-based meta-learning. Finally, we show how gradient re-scaling at meta-test time allows the learning of extremely high-quality neural fields in significantly shortened optimization procedures. Our framework is model-agnostic, intuitive, straightforward to implement, and shows significant reconstruction improvements for a wide range of signals. We provide an extensive empirical evaluation on nine datasets across multiple multiple modalities, demonstrating state-of-the-art results while providing additional insight through careful analysis of the algorithmic components constituting our method. Code is available at https://github.com/jihoontack/GradNCP

The Mixup scheme suggests mixing a pair of samples to create an augmented training sample and has gained considerable attention recently for improving the generalizability of neural networks. A straightforward and widely used extension of Mixup is to combine with regional dropout-like methods: removing random patches from a sample and replacing it with the features from another sample. Albeit their simplicity and effectiveness, these methods are prone to create harmful samples due to their randomness. To address this issue, 'maximum saliency' strategies were recently proposed: they select only the most informative features to prevent such a phenomenon. However, they now suffer from lack of sample diversification as they always deterministically select regions with maximum saliency, injecting bias into the augmented data. In this paper, we present, a novel, yet simple Mixup-variant that captures the best of both worlds. Our idea is two-fold. By stochastically sampling the features and ‘grafting’ them onto another sample, our method effectively generates diverse yet meaningful samples. Its second ingredient is to produce the label of the grafted sample by mixing the labels in a saliency-calibrated fashion, which rectifies supervision misguidance introduced by the random sampling procedure. Our experiments under CIFAR, Tiny-ImageNet, and ImageNet datasets show that our scheme outperforms the current state-of-the-art augmentation strategies not only in terms of classification accuracy, but is also superior in coping under stress conditions such as data corruption and object occlusion.

Adversarial training (AT) is currently one of the most successful methods to obtain the adversarial robustness of deep neural networks. However, the phenomenon of robust overfitting, i.e., the robustness starts to decrease significantly during AT, has been problematic, not only making practitioners consider a bag of tricks for a successful training, e.g., early stopping, but also incurring a significant generalization gap in the robustness. In this paper, we propose an effective regularization technique that prevents robust overfitting by optimizing an auxiliary `consistency' regularization loss during AT. Specifically, we discover that data augmentation is a quite effective tool to mitigate the overfitting in AT, and develop a regularization that forces the predictive distributions after attacking from two different augmentations of the same instance to be similar with each other. Our experimental results demonstrate that such a simple regularization technique brings significant improvements in the test robust accuracy of a wide range of AT methods. More remarkably, we also show that our method could significantly help the model to generalize its robustness against unseen adversaries, e.g., other types or larger perturbations compared to those used during training. Code is available at https://github.com/alinlab/consistency-adversarial.

Unsupervised image-to-image translation has gained considerable attention due to recent impressive advances in generative adversarial networks (GANs). This paper presents a simple but effective regularization technique for improving GAN-based image-to-image translation. To generate images with realistic local semantics and structures, we propose an auxiliary self-supervision loss that enforces point-wise consistency of the overlapping region between a pair of patches cropped from a single real image during training the discriminator of a GAN. Our experiment shows that the proposed dense consistency regularization improves performance substantially on various image-to-image translation scenarios. It also leads to extra performance gains through the combination with instance-level regularization methods. Furthermore, we verify that the proposed model captures domain-specific characteristics more effectively with only a small fraction of training data.

Contrastive learning (or its variants) has recently become a promising direction in the self-supervised learning domain, achieving similar performance as supervised learning with minimum fine-tuning. Despite the labeling efficiency, wide and large networks are required to achieve high accuracy, which incurs a high amount of computation and hinders the pragmatic merit of self-supervised learning. To effectively reduce the computation of insignificant features or channels, recent dynamic pruning algorithms for supervised learning employed auxiliary salience predictors. However, we found that such salience predictors cannot be easily trained when they are naïvely applied to contrastive learning from scratch. To address this issue, we propose contrastive dual gating (CDG), a novel dynamic pruning algorithm that skips the uninformative features during contrastive learning without hurting the trainability of the networks. We demonstrate the superiority of CDG with ResNet models for CIFAR-10, CIFAR-100, and ImageNet-100 datasets. Compared to our implementations of state-of-the-art dynamic pruning algorithms for self-supervised learning, CDG achieves up to 15% accuracy improvement for CIFAR-10 dataset with higher computation reduction.

Recent self-supervised learning (SSL) methods have shown impressive results in learning visual representations from unlabeled images. This paper aims to improve their performance further by utilizing the architectural advan-tages of the underlying neural network, as the current state-of-the-art visual pretext tasks for SSL do not enjoy the ben-efit, i.e., they are architecture-agnostic. In particular, we fo-cus on Vision Transformers (ViTs), which have gained much attention recently as a better architectural choice, often out-performing convolutional networks for various visual tasks. The unique characteristic of ViT is that it takes a sequence of disjoint patches from an image and processes patch-level representations internally. Inspired by this, we design a simple yet effective visual pretext task, coined Self Patch, for learning better patch-level representations. To be specific, we enforce invariance against each patch and its neigh-bors, i.e., each patch treats similar neighboring patches as positive samples. Consequently, training ViTs with Self-Patch learns more semantically meaningful relations among patches (without using human-annotated labels), which can be beneficial, in particular, to downstream tasks of a dense prediction type. Despite its simplicity, we demonstrate that it can significantly improve the performance of existing SSL methods for various visual tasks, including object detection and semantic segmentation. Specifically, Self Patch signif-icantly improves the recent self-supervised ViT, DINO, by achieving +1.3 AP on COCO object detection, +1.2 AP on COCO instance segmentation, and +2.9 mIoU on ADE20K semantic segmentation.

The teacher-student (TS) framework, training a (student) network by utilizing an auxiliary superior (teacher) network, has been adopted as a popular training paradigm in many machine learning schemes, since the seminal work—Knowledge distillation (KD) for model compression and transfer learning. Many recent self-supervised learning (SSL) schemes also adopt the TS framework, where teacher networks are maintained as the moving average of student networks, called the momentum networks. This paper presents TSPipe, a pipelined approach to accelerate the training process of any TS frameworks including KD and SSL. Under the observation that the teacher network does not need a backward pass, our main idea is to schedule the computation of the teacher and student network separately, and fully utilize the GPU during training by interleaving the computations of the two networks and relaxing their dependencies. In case the teacher network requires a momentum update, we use delayed parameter updates only on the teacher network to attain high model accuracy. Compared to existing pipeline parallelism schemes, which sacrifice either training throughput or model accuracy, TSPipe provides better performance trade-offs, achieving up to 12.15x higher throughput.

In cooperative multi-agent reinforcement learning, the outcomes of agent-wise policies are highly stochastic due to the two sources of risk: (a) random actions taken by teammates and (b) random transition and rewards. Although the two sources have very distinct characteristics, existing frameworks are insufficient to control the risk-sensitivity of agent-wise policies in a disentangled manner. To this end, we propose Disentangled RIsk-sensitive Multi-Agent reinforcement learning (DRIMA) to separately access the risk sources. For example, our framework allows an agent to be optimistic with respect to teammates (who can prosocially adapt) but more risk-neutral with respect to the environment (which does not adapt). Our experiments demonstrate that DRIMA significantly outperforms prior state-of-the-art methods across various scenarios in the StarCraft Multi-agent Challenge environment. Notably, DRIMA shows robust performance where prior methods learn only a highly suboptimal policy, regardless of reward shaping, exploration scheduling, and noisy (random or adversarial) agents.

Understanding temporal dynamics of video is an essential aspect of learning better video representations. Recently, transformer-based architectural designs have been extensively explored for video tasks due to their capability to capture long-term dependency of input sequences. However, we found that these Video Transformers are still biased to learn spatial dynamics rather than temporal ones, and debiasing the spurious correlation is critical for their performance. Based on the observations, we design simple yet effective self-supervised tasks for video models to learn temporal dynamics better. Specifically, for debiasing the spatial bias, our method learns the temporal order of video frames as extra self-supervision and enforces the randomly shuffled frames to have low-confidence outputs. Also, our method learns the temporal flow direction of video tokens among consecutive frames for enhancing the correlation toward temporal dynamics. Under various video action recognition tasks, we demonstrate the effectiveness of our method and its compatibility with state-of-the-art Video Transformers.

Test-time adaptation (TTA) is an emerging paradigm that addresses distributional shifts between training and testing phases without additional data acquisition or labeling cost; only unlabeled test data streams are used for continual model adaptation. Previous TTA schemes assume that the test samples are independent and identically distributed (i.i.d.), even though they are often temporally correlated (non-i.i.d.) in application scenarios, e.g., autonomous driving. We discover that most existing TTA methods fail dramatically under such scenarios. Motivated by this, we present a new test-time adaptation scheme that is robust against non-i.i.d. test data streams. Our novelty is mainly two-fold: (a) Instance-Aware Batch Normalization (IABN) that corrects normalization for out-of-distribution samples, and (b) Prediction-balanced Reservoir Sampling (PBRS) that simulates i.i.d. data stream from non-i.i.d. stream in a class-balanced manner. Our evaluation with various datasets, including real-world non-i.i.d. streams, demonstrates that the proposed robust TTA not only outperforms state-of-the-art TTA algorithms in the non-i.i.d. setting, but also achieves comparable performance to those algorithms under the i.i.d. assumption. Code is available at https://github.com/TaesikGong/NOTE.

Succinct representation of complex signals using coordinate-based neural representations (CNRs) has seen great progress, and several recent efforts focus on extending them for handling videos. Here, the main challenge is how to (a) alleviate a compute-inefficiency in training CNRs to (b) achieve high-quality video encoding while (c) maintaining the parameter-efficiency. To meet all requirements (a), (b), and (c) simultaneously, we propose neural video representations with learnable positional features (NVP), a novel CNR by introducing "learnable positional features" that effectively amortize a video as latent codes. Specifically, we first present a CNR architecture based on designing 2D latent keyframes to learn the common video contents across each spatio-temporal axis, which dramatically improves all of those three requirements. Then, we propose to utilize existing powerful image and video codecs as a compute-/memory-efficient compression procedure of latent codes. We demonstrate the superiority of NVP on the popular UVG benchmark; compared with prior arts, NVP not only trains 2 times faster (less than 5 minutes) but also exceeds their encoding quality as 34.07$\rightarrow$34.57 (measured with the PSNR metric), even using $>$8 times fewer parameters. We also show intriguing properties of NVP, e.g., video inpainting, video frame interpolation, etc.

Contrastive representation learning seeks to acquire useful representations by estimating the shared information between multiple views of data. Here, the choice of data augmentation is sensitive to the quality of learned representations: as harder the data augmentations are applied, the views share more task-relevant information, but also task-irrelevant one that can hinder the generalization capability of representation. Motivated by this, we present a new robust contrastive learning scheme, coined RényiCL, which can effectively manage harder augmentations by utilizing Rényi divergence. Our method is built upon the variational lower bound of a Rényi divergence, but a naive usage of a variational method exhibits unstable training due to the large variance. To tackle this challenge, we propose a novel contrastive objective that conducts variational estimation of a skew Renyi divergence and provides a theoretical guarantee on how variational estimation of skew divergence leads to stable training. We show that Rényi contrastive learning objectives perform innate hard negative sampling and easy positive sampling simultaneously so that it can selectively learn useful features and ignore nuisance features. Through experiments on ImageNet, we show that Rényi contrastive learning with stronger augmentations outperforms other self-supervised methods without extra regularization or computational overhead. Also, we validate our method on various domains such as graph and tabular datasets, showing empirical gain over original contrastive methods.

The idea of using a separately trained target model (or teacher) to improve the performance of the student model has been increasingly popular in various machine learning domains, and meta-learning is no exception; a recent discovery shows that utilizing task-wise target models can significantly boost the generalization performance. However, obtaining a target model for each task can be highly expensive, especially when the number of tasks for meta-learning is large. To tackle this issue, we propose a simple yet effective method, coined Self-improving Momentum Target (SiMT). SiMT generates the target model by adapting from the temporal ensemble of the meta-learner, i.e., the momentum network. This momentum network and its task-specific adaptations enjoy a favorable generalization performance, enabling self-improving of the meta-learner through knowledge distillation. Moreover, we found that perturbing parameters of the meta-learner, e.g., dropout, further stabilize this self-improving process by preventing fast convergence of the distillation loss during meta-training. Our experimental results demonstrate that SiMT brings a significant performance gain when combined with a wide range of meta-learning methods under various applications, including few-shot regression, few-shot classification, and meta-reinforcement learning. Code is available at https://github.com/jihoontack/SiMT.

Many applications utilize sensors on mobile devices and apply deep learning for diverse applications. However, they have rarely enjoyed mainstream adoption due to many differentindividual conditionsusers encounter. Individual conditions are characterized by users’ unique behaviors and different devices they carry, which collectively make sensor inputs different. It is impractical to train countless individual conditions beforehand and we thus argue meta-learning is a great approach in solving this problem. We presentMetaSensethat leverages “seen” conditions in training data to adapt to an “unseen” condition (i.e., the target user). Specifically, we design a meta-learning framework that learns “how to adapt” to the target via iterative training sessions of adaptation. MetaSense requires very few training examples from the target (e.g., one or two) and thus requires minimal user effort. In addition, we propose asimilar condition detector(SCD) that identifies when the unseen condition has similar characteristics to seen conditions and leverages this hint to further improve the accuracy. Our evaluation with 10 different datasets shows that MetaSense improves the accuracy of state-of-the-art transfer learning and meta learning methods by 15 and 11 percent, respectively. Furthermore, our SCD achieves additional accuracy improvement (e.g., 15 percent for human activity recognition).

Pre-trained language models have achieved state-of-the-art accuracies on various text classification tasks, e.g., sentiment analysis, natural language inference, and semantic textual similarity. However, the reliability of the fine-tuned text classifiers is an often underlooked performance criterion. For instance, one may desire a model that can detect out-of-distribution (OOD) samples (drawn far from training distribution) or be robust against domain shifts. We claim that one central obstacle to the reliability is the over-reliance of the model on a limited number of keywords, instead of looking at the whole context. In particular, we find that (a) OOD samples often contain in-distribution keywords, while (b) cross-domain samples may not always contain keywords; over-relying on the keywords can be problematic for both cases. In light of this observation, we propose a simple yet effective fine-tuning method, coined masked keyword regularization (MASKER), that facilitates context-based prediction. MASKER regularizes the model to reconstruct the keywords from the rest of the words and make low-confidence predictions without enough context. When applied to various pre-trained language models (e.g., BERT, RoBERTa, and ALBERT), we demonstrate that MASKER improves OOD detection and cross-domain generalization without degrading classification accuracy. Code is available at https://github.com/alinlab/MASKER.

Retrosynthesis is the task of predicting reactant molecules from a given product molecule and is, important in organic chemistry because the identification of a synthetic path is as demanding as the discovery of new chemical compounds. Recently, the retrosynthesis task has been solved automatically without human expertise using powerful deep learning models. Recent deep models are primarily based on seq2seq or graph neural networks depending on the function of molecular representation, sequence, or graph. Current state-of-the-art models represent a molecule as a graph, but they require joint training with auxiliary prediction tasks, such as the most probable reaction template or reaction center prediction. Furthermore, they require additional labels by experienced chemists, thereby incurring additional cost. Herein, we propose a novel template-free model, i.e., Graph Truncated Attention (GTA), which leverages both sequence and graph representations by inserting graphical information into a seq2seq model. The proposed GTA model masks the self-attention layer using the adjacency matrix of product molecule in the encoder and applies a new loss using atom mapping acquired from an automated algorithm to the cross-attention layer in the decoder. Our model achieves new state-of-the-art records, i.e., exact match top-1 and top-10 accuracies of 51.1% and 81.6% on the USPTO-50k benchmark dataset, respectively, and 46.0% and 70.0% on the USPTO-full dataset, respectively, both without any reaction class information. The GTA model surpasses prior graph-based template-free models by 2% and 7% in terms of the top-1 and top-10 accuracies on the USPTO-50k dataset, respectively, and by over 6% for both the top-1 and top-10 accuracies on the USPTO-full dataset.

It is known that $O(N)$ parameters are sufficient for neural networks to memorize arbitrary $N$ input-label pairs. By exploiting depth, we show that $O(N^{2/3})$ parameters suffice to memorize $N$ pairs, under a mild condition on the separation of input points. In particular, deeper networks (even with width 3) are shown to memorize more pairs than shallow networks, which also agrees with the recent line of works on the benefits of depth for function approximation. We also provide empirical results that support our theoretical findings.

Despite the remarkable performance of deep models on image recognition tasks, they are known to be susceptible to common corruptions such as blur, noise, and low-resolution. Data augmentation is a conventional way to build a robust model by considering these common corruptions during the training. However, a naive data augmentation scheme may result in a non-specialized model for particular corruptions, as the model tends to learn the averaged distribution among corruptions. To mitigate the issue, we propose a new paradigm of training deep image recognition networks that produce clean-like features from any quality image via an invertible neural architecture. The proposed method consists of two stages. In the first stage, we train an invertible network with only clean images under the recognition objective. In the second stage, its inversion, i.e., the invertible decoder, is attached to a new recognition network and we train this encoder-decoder network using both clean and corrupted images by considering recognition and reconstruction objectives. Our two-stage scheme allows the network to produce clean-like and robust features from any quality images, by reconstructing their clean images via the invertible decoder. We demonstrate the effectiveness of our method on image classification and face recognition tasks.

Recent breakthroughs in self-supervised learning show that such algorithms learn visual representations that can be transferred better to unseen tasks than cross-entropy based methods which rely on task-specific supervision. In this paper, we found that the similar holds in the continual learning context: contrastively learned representations are more robust against the catastrophic forgetting than ones trained with the cross-entropy objective. Based on this novel observation, we propose a rehearsal-based continual learning algorithm that focuses on continually learning and maintaining transferable representations. More specifically, the proposed scheme (1) learns representations using the contrastive learning objective, and (2) preserves learned representations using a self-supervised distillation step. We conduct extensive experimental validations under popular benchmark image classification datasets, where our method sets the new state-of-the-art performance. Source code is available at https://github.com/chaht01/Co2L.

Retrosynthetic planning is a fundamental problem in chemistry for finding a pathway of reactions to synthesize a target molecule. Recently, search algorithms have shown promising results for solving this problem by using deep neural networks (DNNs) to expand their candidate solutions, i.e., adding new reactions to reaction pathways. However, the existing works on this line are suboptimal; the retrosynthetic planning problem requires the reaction pathways to be (a) represented by real-world reactions and (b) executable using “building block” molecules, yet the DNNs expand reaction pathways without fully incorporating such requirements. Motivated by this, we propose an end-to-end framework for directly training the DNNs towards generating reaction pathways with the desirable properties. Our main idea is based on a self-improving procedure that trains the model to imitate successful trajectories found by itself. We also propose a novel reaction augmentation scheme based on a forward reaction model. Our experiments demonstrate that our scheme significantly improves the success rate of solving the retrosynthetic problem from 86.84% to 96.32% while maintaining the performance of DNN for predicting valid reactions.

Adversarial learning has emerged as one of the successful techniques to circumvent the susceptibility of existing methods against adversarial perturbations. However, the majority of existing defense methods are tailored to defend against a single category of adversarial perturbation (e.g. $\ell_\infty$-attack). In safety-critical applications, this makes these methods extraneous as the attacker can adopt diverse adversaries to deceive the system. Moreover, training on multiple perturbations simultaneously significantly increases the computational overhead during training. To address these challenges, we propose a novel meta-learning framework that explicitly learns to generate noise to improve the model’s robustness against multiple types of attacks. Its key component is \emph{Meta Noise Generator (MNG)} that outputs optimal noise to stochastically perturb a given sample, such that it helps lower the error on diverse adversarial perturbations. By utilizing samples generated by MNG, we train a model by enforcing the label consistency across multiple perturbations. We validate the robustness of models trained by our scheme on various datasets and against a wide variety of perturbations, demonstrating that it significantly outperforms the baselines across multiple perturbations with a marginal computational cost.

Recent exploration methods have proven to be a recipe for improving sample-efficiency in deep reinforcement learning (RL). However, efficient exploration in high-dimensional observation spaces still remains a challenge. This paper presents Random Encoders for Efficient Exploration (RE3), an exploration method that utilizes state entropy as an intrinsic reward. In order to estimate state entropy in environments with high-dimensional observations, we utilize a k-nearest neighbor entropy estimator in the low-dimensional representation space of a convolutional encoder. In particular, we find that the state entropy can be estimated in a stable and compute-efficient manner by utilizing a randomly initialized encoder, which is fixed throughout training. Our experiments show that RE3 significantly improves the sample-efficiency of both model-free and model-based RL methods on locomotion and navigation tasks from DeepMind Control Suite and MiniGrid benchmarks. We also show that RE3 allows learning diverse behaviors without extrinsic rewards, effectively improving sample-efficiency in downstream tasks.

Retrosynthesis, of which the goal is to find a set of reactants for synthesizing a target product, is an emerging research area of deep learning. While the existing approaches have shown promising results, they currently lack the ability to consider availability (e.g., stability or purchasability) of the reactants or generalize to unseen reaction templates (i.e., chemical reaction rules). In this paper, we propose a new approach that mitigates the issues by reformulating retrosynthesis into a selection problem of reactants from a candidate set of commercially available molecules. To this end, we design an efficient reactant selection framework, named RetCL (retrosynthesis via contrastive learning), for enumerating all of the candidate molecules based on selection scores computed by graph neural networks. For learning the score functions, we also propose a novel contrastive training scheme with hard negative mining. Extensive experiments demonstrate the benefits of the proposed selection-based approach. For example, when all 671k reactants in the USPTO database are given as candidates, our RetCL achieves top-1 exact match accuracy of 71.3% for the USPTO-50k benchmark, while a recent transformer-based approach achieves 59.6%. We also demonstrate that RetCL generalizes well to unseen templates in various settings in contrast to template-based approaches.

Randomized smoothing is currently a state-of-the-art method to construct a certifiably robust classifier from neural networks against $\ell_2$-adversarial perturbations. Under the paradigm, the robustness of a classifier is aligned with the prediction confidence, i.e., the higher confidence from a smoothed classifier implies the better robustness. This motivates us to rethink the fundamental trade-off between accuracy and robustness in terms of calibrating confidences of a smoothed classifier. In this paper, we propose a simple training scheme, coined SmoothMix, to control the robustness of smoothed classifiers via self-mixup: it trains on convex combinations of samples along the direction of adversarial perturbation for each input. The proposed procedure effectively identifies over-confident, near off-class samples as a cause of limited robustness in case of smoothed classifiers, and offers an intuitive way to adaptively set a new decision boundary between these samples for better robustness. Our experimental results demonstrate that the proposed method can significantly improve the certified $\ell_2$-robustness of smoothed classifiers compared to existing state-of-the-art robust training methods.

Goal-conditioned hierarchical reinforcement learning (HRL) has shown promising results for solving complex and long-horizon RL tasks. However, the action space of high-level policy in the goal-conditioned HRL is often large, so it results in poor exploration, leading to inefficiency in training. In this paper, we present HIerarchical reinforcement learning Guided by Landmarks (HIGL), a novel framework for training a high-level policy with a reduced action space guided by landmarks, i.e., promising states to explore. The key component of HIGL is twofold: (a) sampling landmarks that are informative for exploration and (b) encouraging the high level policy to generate a subgoal towards a selected landmark. For (a), we consider two criteria: coverage of the entire visited state space (i.e., dispersion of states) and novelty of states (i.e., prediction error of a state). For (b), we select a landmark as the very first landmark in the shortest path in a graph whose nodes are landmarks. Our experiments demonstrate that our framework outperforms prior-arts across a variety of control tasks, thanks to efficient exploration guided by landmarks.

Recent unsupervised representation learning methods have shown to be effective in a range of vision tasks by learning representations invariant to data augmentations such as random cropping and color jittering. However, such invariance could be harmful to downstream tasks if they rely on the characteristics of the data augmentations, e.g., location- or color-sensitive. This is not an issue just for unsupervised learning; we found that this occurs even in supervised learning because it also learns to predict the same label for all augmented samples of an instance. To avoid such failures and obtain more generalizable representations, we suggest to optimize an auxiliary self-supervised loss, coined AugSelf, that learns the difference of augmentation parameters (e.g., cropping positions, color adjustment intensities) between two randomly augmented samples. Our intuition is that AugSelf encourages to preserve augmentation-aware information in learned representations, which could be beneficial for their transferability. Furthermore, AugSelf can easily be incorporated into recent state-of-the-art representation learning methods with a negligible additional training cost. Extensive experiments demonstrate that our simple idea consistently improves the transferability of representations learned by supervised and unsupervised methods in various transfer learning scenarios. The code is available at https://github.com/hankook/AugSelf.

Implicit neural representations are a promising new avenue of representing general signals by learning a continuous function that, parameterized as a neural network, maps the domain of a signal to its codomain; the mapping from spatial coordinates of an image to its pixel values, for example. Being capable of conveying fine details in a high dimensional signal, unboundedly of its domain, implicit neural representations ensure many advantages over conventional discrete representations. However, the current approach is difficult to scale for a large number of signals or a data set, since learning a neural representation---which is parameter heavy by itself---for each signal individually requires a lot of memory and computations. To address this issue, we propose to leverage a meta-learning approach in combination with network compression under a sparsity constraint, such that it renders a well-initialized sparse parameterization that evolves quickly to represent a set of unseen signals in the subsequent training. We empirically demonstrate that meta-learned sparse neural representations achieve a much smaller loss than dense meta-learned models with the same number of parameters, when trained to fit each signal using the same number of optimization steps.

Contrastive self-supervised learning has shown impressive results in learning visual representations from unlabeled images by enforcing invariance against different data augmentations. However, the learned representations are often contextually biased to the spurious scene correlations of different objects or object and background, which may harm their generalization on the downstream tasks. To tackle the issue, we develop a novel object-aware contrastive learning framework that first (a) localizes objects in a self-supervised manner and then (b) debias scene correlations via appropriate data augmentations considering the inferred object locations. For (a), we propose the contrastive class activation map (ContraCAM), which finds the most discriminative regions (e.g., objects) in the image compared to the other images using the contrastively trained models. We further improve the ContraCAM to detect multiple objects and entire shapes via an iterative refinement procedure. For (b), we introduce two data augmentations based on ContraCAM, object-aware random crop and background mixup, which reduce contextual and background biases during contrastive self-supervised learning, respectively. Our experiments demonstrate the effectiveness of our representation learning framework, particularly when trained under multi-object images or evaluated under the background (and distribution) shifted images. Code is available at https://github.com/alinlab/object-aware-contrastive.

Behavioral cloning has proven to be effective for learning sequential decision-making policies from expert demonstrations. However, behavioral cloning often suffers from the causal confusion problem where a policy relies on the noticeable effect of expert actions due to the strong correlation but not the cause we desire. This paper presents Object-aware REgularizatiOn (OREO), a simple technique that regularizes an imitation policy in an object-aware manner. Our main idea is to encourage a policy to uniformly attend to all semantic objects, in order to prevent the policy from exploiting nuisance variables strongly correlated with expert actions. To this end, we introduce a two-stage approach: (a) we extract semantic objects from images by utilizing discrete codes from a vector-quantized variational autoencoder, and (b) we randomly drop the units that share the same discrete code together, i.e., masking out semantic objects. Our experiments demonstrate that OREO significantly improves the performance of behavioral cloning, outperforming various other regularization and causality-based methods on a variety of Atari environments and a self-driving CARLA environment. We also show that our method even outperforms inverse reinforcement learning methods trained with a considerable amount of environment interaction.

We consider the problem of searching an input maximizing a black-box objective function given a static dataset of input-output queries. A popular approach to solving this problem is maintaining a proxy model, e.g., a deep neural network (DNN), that approximates the true objective function. Here, the main challenge is how to avoid adversarially optimized inputs during the search, i.e., the inputs where the DNN highly overestimates the true objective function. To handle the issue, we propose a new framework, coined robust model adaptation (RoMA), based on gradient-based optimization of inputs over the DNN. Specifically, it consists of two steps: (a) a pre-training strategy to robustly train the proxy model and (b) a novel adaptation procedure of the proxy model to have robust estimates for a specific set of candidate solutions. At a high level, our scheme utilizes the local smoothness prior to overcome the brittleness of the DNN. Experiments under various tasks show the effectiveness of RoMA compared with previous methods, obtaining state-of-the-art results, e.g., RoMA outperforms all at 4 out of 6 tasks and achieves runner-up results at the remaining tasks.

The Neural Tangent Kernel (NTK) characterizes the behavior of infinitely-wide neural networks trained under least squares loss by gradient descent. Recent works also report that NTK regression can outperform finitely-wide neural networks trained on small-scale datasets. However, the computational complexity of kernel methods has limited its use in large-scale learning tasks. To accelerate learning with NTK, we design a near input-sparsity time approximation algorithm for NTK, by sketching the polynomial expansions of arc-cosine kernels: our sketch for the convolutional counterpart of NTK (CNTK) can transform any image using a linear runtime in the number of pixels. Furthermore, we prove a spectral approximation guarantee for the NTK matrix, by combining random features (based on leverage score sampling) of the arc-cosine kernels with a sketching algorithm. We benchmark our methods on various large-scale regression and classification tasks and show that a linear regressor trained on our CNTK features matches the accuracy of exact CNTK on CIFAR-10 dataset while achieving 150x speedup.

In most real-world scenarios, labeled training datasets are highly class-imbalanced, where deep neural networks suffer from generalizing to a balanced testing criterion. In this paper, we explore a novel yet simple way to alleviate this issue by augmenting less-frequent classes via translating samples (e.g., images) from more-frequent classes. This simple approach enables a classifier to learn more generalizable features of minority classes, by transferring and leveraging the diversity of the majority information. Our experimental results on a variety of class-imbalanced datasets show that the proposed method improves the generalization on minority classes significantly compared to other existing re-sampling or re-weighting methods. The performance of our method even surpasses those of previous state-of-the-art methods for the imbalanced classification.

Deep neural networks with millions of parameters may suffer from poor generalization due to overfitting. To mitigate the issue, we propose a new regularization method that penalizes the predictive distribution between similar samples. In particular, we distill the predictive distribution between different samples of the same label during training. This results in regularizing the dark knowledge (i.e., the knowledge on wrong predictions) of a single network (i.e., a self-knowledge distillation) by forcing it to produce more meaningful and consistent predictions in a class-wise manner. Consequently, it mitigates overconfident predictions and reduces intra-class variations. Our experimental results on various image classification tasks demonstrate that the simple yet powerful method can significantly improve not only the generalization ability but also the calibration performance of modern convolutional neural networks.

Designing efficient algorithms for combinatorial optimization appears ubiquitously in various scientific fields. Recently, deep reinforcement learning (DRL) frameworks have gained considerable attention as a new approach: they can automate the design of a solver while relying less on sophisticated domain knowledge of the target problem. However, the existing DRL solvers determine the solution using a number of stages proportional to the number of elements in the solution, which severely limits their applicability to large-scale graphs. In this paper, we seek to resolve this issue by proposing a novel DRL scheme, coined learning what to defer (LwD), where the agent adaptively shrinks or stretch the number of stages by learning to distribute the element-wise decisions of the solution at each stage. We apply the proposed framework to the maximum independent set (MIS) problem, and demonstrate its significant improvement over the current state-of-the-art DRL scheme. We also show that LwD can outperform the conventional MIS solvers on large-scale graphs having millions of vertices, under a limited time budget.

This paper studies how to sketch element-wise functions of low-rank matrices. Formally, given low-rank matrix A = [Aij] and scalar non-linear function f, we aim for finding an approximated low-rank representation of the (possibly high-rank) matrix [f(Aij)]. To this end, we propose an efficient sketching-based algorithm whose complexity is significantly lower than the number of entries of A, i.e., it runs without accessing all entries of [f(Aij)] explicitly. The main idea underlying our method is to combine a polynomial approximation of f with the existing tensor sketch scheme for approximating monomials of entries of A. To balance the errors of the two approximation components in an optimal manner, we propose a novel regression formula to find polynomial coefficients given A and f. In particular, we utilize a coreset-based regression with a rigorous approximation guarantee. Finally, we demonstrate the applicability and superiority of the proposed scheme under various machine learning tasks.

Self-supervised learning, which learns by constructing artificial labels given only the input signals, has recently gained considerable attention for learning representations with unlabeled datasets, i.e., learning without any human-annotated supervision. In this paper, we show that such a technique can be used to significantly improve the model accuracy even under fully-labeled datasets. Our scheme trains the model to learn both original and self-supervised tasks, but is different from conventional multi-task learning frameworks that optimize the summation of their corresponding losses. Our main idea is to learn a single unified task with respect to the joint distribution of the original and self-supervised labels, i.e., we augment original labels via self-supervision. This simple, yet effective approach allows to train models easier by relaxing a certain invariant constraint during learning the original and self-supervised tasks simultaneously. It also enables an aggregated inference which combines the predictions from different augmentations to improve the prediction accuracy. Furthermore, we propose a novel knowledge transfer technique, which we refer to as self-distillation, that has the effect of the aggregated inference in a single (faster) inference. We demonstrate the large accuracy improvement and wide applicability of our framework on various fully-supervised settings, e.g., the few-shot and imbalanced classification scenarios.

Model-based reinforcement learning (RL) enjoys several benefits, such as data-efficiency and planning, by learning a model of the environment’s dynamics. However, learning a global model that can generalize across different dynamics remains a challenge. To tackle this problem, we decompose the task of learning a global dynamics model into two stages: (a) learning a context latent vector that captures the local dynamics, then (b) predicting the next state conditioned on it. In order to encode dynamics-specific information into the context latent vector, we introduce a novel loss function that encourages the context latent vector to be useful for predicting both forward and backward dynamics. The proposed method achieves superior generalization ability across various simulated robotics and control tasks, compared to existing RL schemes.

Despite the remarkable performance of deep neural networks on various computer vision tasks, they are known to be susceptible to adversarial perturbations, which makes it challenging to deploy them in real-world safety-critical applications. In this paper, we conjecture that the leading cause of adversarial vulnerability is the distortion in the latent feature space, and provide methods to suppress them effectively. Explicitly, we define \emph{vulnerability} for each latent feature and then propose a new loss for adversarial learning, \emph{Vulnerability Suppression (VS)} loss, that aims to minimize the feature-level vulnerability during training. We further propose a Bayesian framework to prune features with high vulnerability to reduce both vulnerability and loss on adversarial samples. We validate our \emph{Adversarial Neural Pruning with Vulnerability Suppression (ANP-VS)} method on multiple benchmark datasets, on which it not only obtains state-of-the-art adversarial robustness but also improves the performance on clean examples, using only a fraction of the parameters used by the full network. Further qualitative analysis suggests that the improvements come from the suppression of feature-level vulnerability.

De novo molecular design attempts to search over the chemical space for molecules with the desired property. Recently, deep learning has gained considerable attention as a promising approach to solve the problem. In this paper, we propose genetic expert-guided learning (GEGL), a simple yet novel framework for training a deep neural network (DNN) to generate highly-rewarding molecules. Our main idea is to design a "genetic expert improvement" procedure, which generates high-quality targets for imitation learning of the DNN. Extensive experiments show that GEGL significantly improves over state-of-the-art methods. For example, GEGL manages to solve the penalized octanol-water partition coefficient optimization with a score of 31.40, while the best-known score in the literature is 27.22. Besides, for the GuacaMol benchmark with 20 tasks, our method achieves the highest score for 19 tasks, in comparison with state-of-the-art methods, and newly obtains the perfect score for three tasks. Our training code is available at https://github.com/sungsoo-ahn/genetic-expert-guided-learning.

Time-reversal symmetry, which requires that the dynamics of a system should not change with the reversal of time axis, is a fundamental property that frequently holds in classical and quantum mechanics. In this paper, we propose a novel loss function that measures how well our ordinary differential equation (ODE) networks comply with this time-reversal symmetry; it is formally defined by the discrepancy in the time evolutions of ODE networks between forward and backward dynamics. Then, we design a new framework, which we name as Time-Reversal Symmetric ODE Networks (TRS-ODENs), that can learn the dynamics of physical systems more sample-efficiently by learning with the proposed loss function. We evaluate TRS-ODENs on several classical dynamics, and find they can learn the desired time evolution from observed noisy and complex trajectories. We also show that, even for systems that do not possess the full time-reversal symmetry, TRS-ODENs can achieve better predictive performances over baselines.

A recent technique of randomized smoothing has shown that the worst-case (adversarial) l2-robustness can be transformed into the average-case Gaussian-robustness by "smoothing" a classifier, i.e., by considering the averaged prediction over Gaussian noise. In this paradigm, one should rethink the notion of adversarial robustness in terms of generalization ability of a classifier under noisy observations. We found that the trade-off between accuracy and certified robustness of smoothed classifiers can be greatly controlled by simply regularizing the prediction consistency over noise. This relationship allows us to design a robust training objective without approximating a non-existing smoothed classifier, e.g., via soft smoothing. Our experiments under various deep neural network architectures and datasets show that the "certified" l2-robustness can be dramatically improved with the proposed regularization, even achieving better or comparable results to the state-of-the-art approaches with significantly less training costs and hyperparameters.

While semi-supervised learning (SSL) has proven to be a promising way for leveraging unlabeled data when labeled data is scarce, the existing SSL algorithms typically assume that training class distributions are balanced. However, these SSL algorithms trained under imbalanced class distributions can severely suffer when generalizing to a balanced testing criterion, since they utilize biased pseudo-labels of unlabeled data toward majority classes. To alleviate this issue, we formulate a convex optimization problem to softly refine the pseudo-labels generated from the biased model, and develop a simple algorithm, named Distribution Aligning Refinery of Pseudo-label (DARP) that solves it provably and efficiently. Under various class imbalanced semi-supervised scenarios, we demonstrate the effectiveness of DARP and its compatibility with state-of-the-art SSL schemes.

While humans can solve a visual puzzle that requires logical reasoning by observing only few samples, it would require training over a large number of samples for state-of-the-art deep reasoning models to obtain similar performance on the same task. In this work, we propose to solve such a few-shot (or low-shot) abstract visual reasoning problem by resorting to \emph{analogical reasoning}, which is a unique human ability to identify structural or relational similarity between two sets. Specifically, we construct analogical and non-analogical training pairs of two different problem instances, e.g., the latter is created by perturbing or shuffling the original (former) problem. Then, we extract the structural relations among elements in both domains in a pair by enforcing analogical ones to be as similar as possible, while minimizing similarities between non-analogical ones. This analogical contrastive learning allows to effectively learn the relational representations of given abstract reasoning tasks. We validate our method on RAVEN dataset, on which it outperforms state-of-the-art method, with larger gains when the training data is scarce. We further meta-learn our analogical contrastive learning model over the same tasks with diverse attributes, and show that it generalizes to the same visual reasoning problem with unseen attributes.

In risk-sensitive learning, one aims to find a hypothesis that minimizes a risk-averse (or risk-seeking) measure of loss, instead of the standard expected loss. In this paper, we propose to study the generalization properties of risk-sensitive learning schemes whose optimand is described via optimized certainty equivalents (OCE): our general scheme can handle various known risks, e.g., the entropic risk, mean-variance, and conditional value-at-risk, as special cases. We provide two learning bounds on the performance of empirical OCE minimizer. The first result gives an OCE guarantee based on the Rademacher average of the hypothesis space, which generalizes and improves existing results on the expected loss and the conditional value-at-risk. The second result, based on a novel variance-based characterization of OCE, gives an expected loss guarantee with a suppressed dependence on the smoothness of the selected OCE. Finally, we demonstrate the practical implications of the proposed bounds via exploratory experiments on neural networks.

Neural networks often learn to make predictions that overly rely on spurious corre- lation existing in the dataset, which causes the model to be biased. While previous work tackles this issue by using explicit labeling on the spuriously correlated attributes or presuming a particular bias type, we instead utilize a cheaper, yet generic form of human knowledge, which can be widely applicable to various types of bias. We first observe that neural networks learn to rely on the spurious correlation only when it is “easier” to learn than the desired knowledge, and such reliance is most prominent during the early phase of training. Based on the obser- vations, we propose a failure-based debiasing scheme by training a pair of neural networks simultaneously. Our main idea is twofold; (a) we intentionally train the first network to be biased by repeatedly amplifying its “prejudice”, and (b) we debias the training of the second network by focusing on samples that go against the prejudice of the biased network in (a). Extensive experiments demonstrate that our method significantly improves the training of network against various types of biases in both synthetic and real-world datasets. Surprisingly, our framework even occasionally outperforms the debiasing methods requiring explicit supervision of the spuriously correlated attributes.

Model-based reinforcement learning (RL) has shown great potential in various control tasks in terms of both sample-efficiency and final performance. However, learning a generalizable dynamics model robust to changes in dynamics remains a challenge since the target transition dynamics follow a multi-modal distribution. In this paper, we present a new model-based RL algorithm, coined trajectory-wise multiple choice learning, that learns a multi-headed dynamics model for dynamics generalization. The main idea is updating the most accurate prediction head to specialize each head in certain environments with similar dynamics, i.e., clustering environments. Moreover, we incorporate context learning, which encodes dynamics-specific information from past experiences into the context latent vector, enabling the model to perform online adaptation to unseen environments. Finally, to utilize the specialized prediction heads more effectively, we propose an adaptive planning method, which selects the most accurate prediction head over a recent experience. Our method exhibits superior zero-shot generalization performance across a variety of control tasks, compared to state-of-the-art RL methods. Source code and videos are available at https://sites.google.com/view/trajectory-mcl.

Novelty detection, i.e., identifying whether a given sample is drawn from outside the training distribution, is essential for reliable machine learning. To this end, there have been many attempts at learning a representation well-suited for novelty detection and designing a score based on such representation. In this paper, we propose a simple, yet effective method named contrasting shifted instances (CSI), inspired by the recent success on contrastive learning of visual representations. Specifically, in addition to contrasting a given sample with other instances as in conventional contrastive learning methods, our training scheme contrasts the sample with distributionally-shifted augmentations of itself. Based on this, we propose a new detection score that is specific to the proposed training scheme. Our experiments demonstrate the superiority of our method under various novelty detection scenarios, including unlabeled one-class, unlabeled multi-class and labeled multi-class settings, with various image benchmark datasets. Code and pre-trained models are available at https://github.com/alinlab/CSI.

In this paper, we study a problem of detecting the source of diffused information by querying individuals, given a sample snapshot of the information diffusion graph, where two queries are asked: (i) whether the respondent is the source or not, and (ii) if not, which neighbor spreads the information to the respondent. We consider the case when respondents may not always be truthful and some cost is taken for each query. Our goal is to quantify the necessary and sufficient budgets to achieve the detection probability 1- δ for any given 0 <; δ <; 1. To this end, we study two types of algorithms: adaptive and non-adaptive ones, each of which corresponds to whether we adaptively select the next respondents based on the answers of the previous respondents or not. We first provide the information theoretic lower bounds for the necessary budgets in both algorithm types. In terms of the sufficient budgets, we propose two practical estimation algorithms, each of non-adaptive and adaptive types, and for each algorithm, we quantitatively analyze the budget which ensures 1- δ detection accuracy. This theoretical analysis not only quantifies the budgets needed by practical estimation algorithms achieving a given target detection accuracy in finding the diffusion source, but also enables us to quantitatively characterize the amount of extra budget required in non-adaptive type of estimation, referred to as adaptivity gap. We validate our theoretical findings over synthetic and real-world social network topologies.

In order to handle a high traffic demand, dense wireless local area networks (WLANs) have been deployed rapidly in the past years. However, dense WLANs cause two critical issues: wastage of energy and severe interference. To address these issues, the centralized management of dense WLANs has been emerged as a powerful paradigm for improving energy efficiency as well as avoiding severe interference. In this paper, we study the joint optimization problem of power-operation modes in access points (APs), channel selections and user-AP associations for improving energy efficiency and avoiding interference without sacrificing users' demands. To this end, we first formulate it as a mixed-integer programming using the popular Lyapunov approach, but it turns out to be computationally intractable, i.e., NP-hard. To address the issue, we propose a polynomial-time approximation algorithm and prove that it achieves a constant-factor approximation guarantee under mild assumptions. The main novelty underlying our algorithm design is based on a linear programming relaxation combining with two different greedy rounding schemes, where each achieves a constant-factor approximation in different regimes of parameters. We verify the performance of the proposed algorithm via extensive simulations and also demonstrate its practicability by implementing it at commercial APs using a Software-defined Networking framework. Results from our experiments show that it reduces the wasted energy significantly while maintaining even higher throughput.

Crowdsourcing platforms emerged as popular venues for purchasing human intelligence at low cost for large volume of tasks. As many low-paid workers are prone to give noisy answers, a common practice is to add redundancy by assigning multiple workers to each task and then simply average out these answers. However, to fully harness the wisdom of the crowd, one needs to learn the heterogeneous quality of each worker. We resolve this fundamental challenge in crowdsourced regression tasks, i.e., the answer takes continuous labels, where identifying good or bad workers becomes much more non-trivial compared to a classification setting of discrete labels. In particular, we introduce a Bayesian iterative scheme and show that it provably achieves the optimal mean squared error. Our evaluations on synthetic and real-world datasets support our theoretical results and show the superiority of the proposed scheme.

Lifelong learning with deep neural networks is well-known to suffer from catastrophic forgetting: the performance on previous tasks drastically degrades when learning a new task. To alleviate this effect, we propose to leverage a large stream of unlabeled data easily obtainable in the wild. In particular, we design a novel class-incremental learning scheme with (a) a new distillation loss, termed global distillation, (b) a learning strategy to avoid overfitting to the most recent task, and (c) a confidence-based sampling method to effectively leverage unlabeled external data. Our experimental results on various datasets, including CIFAR and ImageNet, demonstrate the superiority of the proposed methods over prior methods, particularly when a stream of unlabeled data is accessible: our method shows up to 15.8% higher accuracy and 46.5% less forgetting compared to the state-of-the-art method. The code is available at https://github.com/kibok90/iccv2019-inc.

This paper aims to classify the phase of gait for passive transfemoral prostheses as a preliminary study for the development of a knee flexion angle control device in prosthetics by attaching it to the knee joint in order to produce a walk trajectory like a normal person, while walking on a flat. However, it is not possible to determine a gait stage according to the inflection point of a knee, since there are few angular changes in the knee joint in the form of a seat that will support the body. Thus, in previous studies, algorithms were developed to distinguish between three stages of the stance in the swing phase using a decision tree learning method. However, the decision-making tree is prone to overfitting. This can be a high level of accuracy for training data, but it is difficult to generalize when verification data or new data are entered. Therefore, in this paper, we want to develop an algorithm for preventing the overfitting step-by-step using two different methods.

As the application of deep learning has expanded to real-world problems with insufficient volume of training data, transfer learning recently has gained much attention as means of improving the performance in such small-data regime. However, when existing methods are applied between heterogeneous architectures and tasks, it becomes more important to manage their detailed configurations and often requires exhaustive tuning on them for the desired performance. To address the issue, we propose a novel transfer learning approach based on meta-learning that can automatically learn what knowledge to transfer from the source network to where in the target network. Given source and target networks, we propose an efficient training scheme to learn meta-networks that decide (a) which pairs of layers between the source and target networks should be matched for knowledge transfer and (b) which features and how much knowledge from each feature should be transferred. We validate our meta-transfer approach against recent transfer learning methods on various datasets and network architectures, on which our automated scheme significantly outperforms the prior baselines that find “what and where to transfer” in a hand-crafted manner.

Recent progress in deep convolutional neural networks (CNNs) have enabled a simple paradigm of architecture design: larger models typically achieve better accuracy. Due to this, in modern CNN architectures, it becomes more important to design models that generalize well under certain resource constraints, e.g. the number of parameters. In this paper, we propose a simple way to improve the capacity of any CNN model having large-scale features, without adding more parameters. In particular, we modify a standard convolutional layer to have a new functionality of channel-selectivity, so that the layer is trained to select important channels to re-distribute their parameters. Our experimental results under various CNN architectures and datasets demonstrate that the proposed new convolutional layer allows new optima that generalize better via efficient resource utilization, compared to the baseline.

Large-scale datasets may contain significant proportions of noisy (incorrect) class labels, and it is well-known that modern deep neural networks (DNNs) poorly generalize from such noisy training datasets. To mitigate the issue, we propose a novel inference method, termed Robust Generative classifier (RoG), applicable to any discriminative (e.g., softmax) neural classifier pre-trained on noisy datasets. In particular, we induce a generative classifier on top of hidden feature spaces of the pre-trained DNNs, for obtaining a more robust decision boundary. By estimating the parameters of generative classifier using the minimum covariance determinant estimator, we significantly improve the classification accuracy with neither re-training of the deep model nor changing its architectures. With the assumption of Gaussian distribution for features, we prove that RoG generalizes better than baselines under noisy labels. Finally, we propose the ensemble version of RoG to improve its performance by investigating the layer-wise characteristics of DNNs. Our extensive experimental results demonstrate the superiority of RoG given different learning models optimized by several training techniques to handle diverse scenarios of noisy labels.

Given a graphical model (GM), computing its partition function is the most essential inference task, but it is computationally intractable in general. To address the issue, iterative approximation algorithms exploring certain local structure/consistency of GM have been investigated as popular choices in practice. However, due to their local/iterative nature, they often output poor approximations or even do not converge, e.g., in low-temperature regimes (hard instances of large parameters). To overcome the limitation, we propose a novel approach utilizing the global spectral feature of GM. Our contribution is two-fold: (a) we first propose a fully polynomial-time approximation scheme (FPTAS) for approximating the partition function of GM associating with a low-rank coupling matrix; (b) for general high-rank GMs, we design a spectral mean-field scheme utilizing (a) as a subroutine, where it approximates a high-rank GM into a product of rank-1 GMs for an efficient approximation of the partition function. The proposed algorithm is more robust in its running time and accuracy than prior methods, i.e., neither suffers from the convergence issue nor depends on hard local structures, as demonstrated in our experiments.

Deep mobile sensing applications are suffering from various individual conditions in the wild. We propose a meta-learned adaptation technique to adapt to a target condition with a few labeled data. We evaluate our system on a public dataset and it outperforms baselines.

Conditional generative adversarial networks (cGANs) have gained a considerable attention in recent years due to its class-wise controllability and superior quality for complex generation tasks. We introduce a simple yet effective approach to improving cGANs by measuring the discrepancy between the data distribution and the model distribution on given samples. The proposed measure, coined the gap of log-densities (GOLD), provides an effective self-diagnosis for cGANs while being efficiently, computed from the discriminator. We propose three applications of the GOLD: example re-weighting, rejection sampling, and active learning, which improve the training, inference, and data selection of cGANs, respectively. Our experimental results demonstrate that the proposed methods outperform corresponding baselines for all three applications on different image datasets.

Recent improvements in deep learning and hardware support offer a new breakthrough in mobile sensing; we could enjoy context-aware services and mobile healthcare on a mobile device powered by artificial intelligence. However, most related studies perform well only with a certain level of similarity between trained and target data distribution, while in practice, a specific user's behaviors and device make sensor inputs different. Consequently, the performance of such applications might suffer in diverse user and device conditions as training deep models in such countless conditions is infeasible. To mitigate the issue, we propose MetaSense, an adaptive deep mobile sensing system utilizing only a few (e.g., one or two) data instances from the target user. MetaSense employs meta learning that learns how to adapt to the target user's condition, by rehearsing multiple similar tasks generated from our unique task generation strategies in offline training. The trained model has the ability to rapidly adapt to the target user's condition when a few data are available. Our evaluation with real-world traces of motion and audio sensors shows that MetaSense not only outperforms the state-of-the-art transfer learning by 18% and meta learning based approaches by 15% in terms of accuracy, but also requires significantly less adaptation time for the target user.

Bitcoin has become the most popular cryptocurrency based on a peer-to-peer network. In Aug. 2017, Bitcoin was split into the original Bitcoin (BTC) and Bitcoin Cash (BCH). Since then, miners have had a choice between BTC and BCH mining because they have compatible proof-of-work algorithms. Therefore, they can freely choose which coin to mine for higher profit, where the profitability depends on both the coin price and mining difficulty. Some miners can immediately switch the coin to mine only when mining difficulty changes because the difficulty changes are more predictable than that for the coin price, and we call this behavior fickle mining. In this paper, we study the effects of fickle mining by modeling a game between two coins. To do this, we consider both fickle miners and some factions (e.g., BITMAIN for BCH mining) that stick to mining one coin to maintain that chain. In this model, we show that fickle mining leads to a Nash equilibrium in which only a faction sticking to its coin mining remains as a loyal miner to the less valued coin (e.g., BCH), where loyal miners refer to those who conduct mining even after coin mining difficulty increases. This situation would cause severe centralization, weakening the security of the coin system. To determine which equilibrium the competing coin systems (e.g., BTC vs. BCH) are moving toward, we traced the historical changes of mining power for BTC and BCH and found that BCH often lacked loyal miners until Nov. 13, 2017, when the difficulty adjustment algorithm of BCH mining was changed. However, the change in difficulty adjustment algorithm of BCH mining led to a state close to the stable coexistence of BTC and BCH. We also demonstrate that the lack of BCH loyal miners may still be reached when a fraction of miners automatically and repeatedly switches to the most profitable coin to mine (i.e., automatic mining). According to our analysis, as of Dec. 2018, loyal miners to BCH would leave if more than about 5% of the total mining capacity for BTC and BCH has engaged in the automatic mining. In addition, we analyze the recent “hash war” between Bitcoin ABC and SV, which confirms our theoretical analysis. Finally, we note that our results can be applied to any competing cryptocurrency systems in which the same hardware (e.g., ASICs or GPUs) can be used for mining. Therefore, our study brings new and important angles in competitive coin markets: a coin can intentionally weaken the security and decentralization level of the other rival coin when mining hardware is shared between them, allowing for automatic mining.

Computing the partition function Z of a discrete graphical model is a fundamental inference challenge. Since this is computationally intractable, variational approximations are often used in practice. Recently, so-called gauge transformations were used to improve variational lower bounds on Z. In this paper, we propose a new gauge-variational approach, termed WMBE-G, which combines gauge transformations with the weighted mini-bucket elimination (WMBE) method. WMBE-G can provide both upper and lower bounds on Z, and is easier to optimize than the prior gauge-variational algorithm. We show that WMBE-G strictly improves the earlier WMBE approximation for symmetric models including Ising models with no magnetic field. Our experimental results demonstrate the effectiveness of WMBE-G even for generic, nonsymmetric models.

Deep neural networks have achieved impressive success in large-scale visual object recognition tasks with a predefined set of classes. However, recognizing objects of novel classes unseen during training still remains challenging. The problem of detecting such novel classes has been addressed in the literature, but most prior works have focused on providing simple binary or regressive decisions, e.g., the output would be "known," "novel," or corresponding confidence intervals. In this paper, we study more informative novelty detection schemes based on a hierarchical classification framework. For an object of a novel class, we aim for finding its closest super class in the hierarchical taxonomy of known classes. To this end, we propose two different approaches termed top-down and flatten methods, and their combination as well. The essential ingredients of our methods are confidence-calibrated classifiers, data relabeling, and the leave-one-out strategy for modeling novel classes under the hierarchical taxonomy. Furthermore, our method can generate a hierarchical embedding that leads to improved generalized zero-shot learning performance in combination with other commonly-used semantic embeddings.

Probabilistic graphical models are a key tool in machine learning applications. Computing the partition function, i.e., normalizing constant, is a fundamental task of statistical inference but is generally computationally intractable, leading to extensive study of approximation methods. Iterative variational methods are a popular and successful family of approaches. However, even state of the art variational methods can return poor results or fail to converge on difficult instances. In this paper, we instead consider computing the partition function via sequential summation over variables. We develop robust approximate algorithms by combining ideas from mini-bucket elimination with tensor network and renormalization group methods from statistical physics. The resulting “convergence-free” methods show good empirical performance on both synthetic and real-world benchmark models, even for difficult instances.

A large class of machine learning techniques requires the solution of optimization problems involving spectral functions of parametric matrices, e.g. log-determinant and nuclear norm. Unfortunately, computing the gradient of a spectral function is generally of cubic complexity, as such gradient descent methods are rather expensive for optimizing objectives involving the spectral function. Thus, one naturally turns to stochastic gradient methods in hope that they will provide a way to reduce or altogether avoid the computation of full gradients. However, here a new challenge appears: there is no straightforward way to compute unbiased stochastic gradients for spectral functions. In this paper, we develop unbiased stochastic gradients for spectral-sums, an important subclass of spectral functions. Our unbiased stochastic gradients are based on combining randomized trace estimators with stochastic truncation of the Chebyshev expansions. A careful design of the truncation distribution allows us to offer distributions that are variance-optimal, which is crucial for fast and stable convergence of stochastic gradient methods. We further leverage our proposed stochastic gradients to devise stochastic methods for objective functions involving spectral-sums, and rigorously analyze their convergence rate. The utility of our methods is demonstrated in numerical experiments.

Detecting test samples drawn sufficiently far away from the training distribution statistically or adversarially is a fundamental requirement for deploying a good classifier in many real-world machine learning applications. However, deep neural networks with the softmax classifier are known to produce highly overconfident posterior distributions even for such abnormal samples. In this paper, we propose a simple yet effective method for detecting any abnormal samples, which is applicable to any pre-trained softmax neural classifier. We obtain the class conditional Gaussian distributions with respect to (low- and upper-level) features of the deep models under Gaussian discriminant analysis, which result in a confidence score based on the Mahalanobis distance. While most prior methods have been evaluated for detecting either out-of-distribution or adversarial samples, but not both, the proposed method achieves the state-of-the-art performances for both cases in our experiments. Moreover, we found that our proposed method is more robust in harsh cases, e.g., when the training dataset has noisy labels or small number of samples. Finally, we show that the proposed method enjoys broader usage by applying it to class-incremental learning: whenever out-of-distribution samples are detected, our classification rule can incorporate new classes well without further training deep models.

Visual Question Answering (VQA) is a notoriously challenging problem because it involves various heterogeneous tasks defined by questions within a unified framework. Learning specialized models for individual types of tasks is intuitively attracting but surprisingly difficult; it is not straightforward to outperform naive independent ensemble approach. We present a principled algorithm to learn specialized models with knowledge distillation under a multiple choice learning (MCL) framework, where training examples are assigned dynamically to a subset of models for updating network parameters. The assigned and non-assigned models are learned to predict ground-truth answers and imitate their own base models before specialization, respectively. Our approach alleviates the limitation of data deficiency in existing MCL frameworks, and allows each model to learn its own specialized expertise without forgetting general knowledge. The proposed framework is model-agnostic and applicable to any tasks other than VQA, e.g., image classification with a large number of labels but few per-class examples, which is known to be difficult under existing MCL schemes. Our experimental results indeed demonstrate that our method outperforms other baselines for VQA and image classification.

We study the maximum weight matching (MWM) problem for general graphs through the max-product belief propagation (BP) and related Linear Programming (LP). The BP approach provides distributed heuristics for finding the maximum a posteriori (MAP) assignment in a joint probability distribution represented by a graphical model (GM), and respective LPs can be considered as continuous relaxations of the discrete MAP problem. It was recently shown that a BP algorithm converges to the correct MAP/MWM assignment under a simple GM formulation of MWM, as long as the corresponding LP relaxation is tight. First, under the motivation for forcing the tightness condition, we consider a new GM formulation of MWM, say C-GM, using non-intersecting odd-sized cycles in the graph; the new corresponding LP relaxation, say C-LP, becomes tight for more MWM instances. However, the tightness of C-LP now does not guarantee such convergence and correctness of the new BP on C-GM. To address the issue, we introduce a novel graph transformation applied to C-GM, which results in another GM formulation of MWM, and prove that the respective BP on it converges to the correct MAP/MWM assignment, as long as C-LP is tight. Finally, we also show that C-LP always has half-integral solutions, which leads to an efficient BP-based MWM heuristic consisting of making sequential, “cutting plane”, modifications to the underlying GM. Our experiments show that this BP-based cutting plane heuristic performs, as well as that based on traditional LP solvers.

Crowdsourcing systems are popular for solving large-scale labeling tasks with low-paid workers. We study the problem of recovering the true labels from the possibly erroneous crowdsourced labels under the popular Dawid-Skene model. To address this inference problem, several algorithms have recently been proposed, but the best known guarantee is still significantly larger than the fundamental limit. We close this gap by introducing a tighter lower bound on the fundamental limit and proving that the belief propagation (BP) exactly matches the lower bound. The guaranteed optimality of BP is the strongest in the sense that it is information-theoretically impossible for any other algorithm to correctly label a larger fraction of the tasks. Experimental results suggest that the BP is close to optimal for all regimes considered and improves upon competing the state-of-the-art algorithms.

Game-theoretic approaches have provided valuable insights into the design of robust local control rules for the individuals in multi-agent systems, e.g., Internet congestion control, road transportation networks, and so on. In this paper, we introduce a non-cooperative medium access control game for wireless networks and propose new fully distributed carrier sense multiple access (CSMA) algorithms that are provably optimal in the sense that their long-term throughputs converge to the optimal solution of a utility maximization problem over the maximum throughput region. The most significant part of our approach lies in introducing novel price functions in agents' utilities so that the proposed game admits an ordinal potential function with no price-of-anarchy. The game formulation naturally leads to game-based dynamics finding a Nash equilibrium, but they often require global information. Toward our goal of designing fully distributed operations, we propose new game-inspired dynamics by utilizing a certain property of CSMA that enables links to estimate their temporary throughputs without message passing. They can be thought of as stochastic approximations to the standard dynamics, which is a new feature in our work, not prevalent in other traditional game-theoretic approaches. We show that they converge to a Nash equilibrium, and numerically evaluate their performance to support our theoretical findings.

The Gibbs sampler is the most popular Markov chain used for learning and inference problems in Graphical Models (GM). These tasks are computationally intractable in general, and the Gibbs sampler often suffers from slow mixing. In this paper, we study the Swendsen-Wang dynamics which is a more sophisticated Markov chain designed to overcome bottlenecks that impede Gibbs sampler. We prove O(log n) mixing time for attractive binary pairwise GMs (i.e., ferromagnetic Ising models) on stochastic partitioned graphs having n vertices, under some mild conditions including low temperature regions where the Gibbs sampler provably mixes exponentially slow. Our experiments also confirm that the Swendsen-Wang sampler significantly outperforms the Gibbs sampler for learning parameters of attractive GMs.

Determinantal point processes (DPPs) are popular probabilistic models that arise in many machine learning tasks, where distributions of diverse sets are characterized by determinants of their features. In this paper, we develop fast algorithms to find the most likely configuration (MAP) of large-scale DPPs, which is NP-hard in general. Due to the submodular nature of the MAP objective, greedy algorithms have been used with empirical success. Greedy implementations require computation of log-determinants, matrix inverses or solving linear systems at each iteration. We present faster implementations of the greedy algorithms by utilizing the orthogonal benefits of two log-determinant approximation schemes: (a) first-order expansions to the matrix log-determinant function and (b) high-order expansions to the scalar log function with stochastic trace estimators. In our experiments, our algorithms are orders of magnitude faster than their competitors, while sacrificing marginal accuracy.

Ensemble methods are arguably the most trustworthy techniques for boosting the performance of machine learning models. Popular independent ensembles (IE) relying on naive averaging/voting scheme have been of typical choice for most applications involving deep neural networks, but they do not consider advanced collaboration among ensemble models. In this paper, we propose new ensemble methods specialized for deep neural networks, called confident multiple choice learning (CMCL): it is a variant of multiple choice learning (MCL) via addressing its overconfidence issue.In particular, the proposed major components of CMCL beyond the original MCL scheme are (i) new loss, i.e., confident oracle loss, (ii) new architecture, i.e., feature sharing and (iii) new training method, i.e., stochastic labeling. We demonstrate the effect of CMCL via experiments on the image classification on CIFAR and SVHN, and the foreground-background segmentation on the iCoseg. In particular, CMCL using 5 residual networks provides 14.05\% and 6.60\% relative reductions in the top-1 error rates from the corresponding IE scheme for the classification task on CIFAR and SVHN, respectively.

Social networks are the major routes for most individuals to exchange their opinions about new products, social trends and political issues via their interactions. It is often of significant importance to figure out who initially diffuses the information, i.e., finding a rumor source or a trend setter. It is known that such a task is highly challenging and the source detection probability cannot be beyond 31% for regular trees, if we just estimate the source from a given diffusion snapshot. In practice, finding the source often entails the process of querying that asks “Are you the rumor source?” or “Who tells you the rumor?” that would increase the chance of detecting the source. In this paper, we consider two kinds of querying: (a) simple batch querying and (b) interactive querying with direction under the assumption that queriees can be untruthful with some probability. We propose estimation algorithms for those queries, and quantify their detection performance and the amount of extra budget due to untruthfulness, analytically showing that querying significantly improves the detection performance. We perform extensive simulations to validate our theoretical findings over synthetic and real-world social network topologies.

We consider a problem of how to effectively diffuse a new product over social networks by incentivizing selfish users. Traditionally, this problem has been studied in the form of influence maximization via seeding, where most prior work assumes that seeded users unconditionally and immediately start by adopting the new product and they stay at the new product throughout their lifetime. However, in practice, seeded users often adjust the degree of their willingness to diffuse, depending on how much incentive is given. To address such diffusion willingness, we propose a new incentive model and characterize the speed of diffusion as the value of a combinatorial optimization. Then, we apply the characterization to popular network graph topologies (Erdos-Renyi, planted partition and power law graphs) as well as general ones, for asymptotically computing the diffusion time for those graphs. Our analysis shows that the diffusion time undergoes two levels of order-wise reduction, where the first and second one are solely contributed by the number of seeded users, i.e., quantity, and the amount of incentives, i.e., quality, respectively. In other words, it implies that the best strategy given budget is (a) first identify the minimum seed set depending on the underlying graph topology, and (b) then assign largest possible incentives to users in the set. We believe that our theoretical results provide useful implications and guidelines for designing successful advertising strategies in various practical applications.

This paper studies the problem of parameter learning in graphical models having latent variables, where the standard approach is the expectation maximization algorithm alternating expectation (E) and maximization (M) steps. However, both E and M steps are computationally intractable for high dimensional data, while the substitution of one step to a faster surrogate for combating against intractability can often cause failure in convergence. To tackle the issue, the Contrastive Divergence (CD) learning scheme has been popularly used in the deep learning community, where it runs the mean-field approximation in E step and a few cycles of Markov Chains (MC) in M step. In this paper, we analyze a variant of CD, called Adiabatic Persistent Contrastive Divergence (APCD), which runs a few cycles of MCs in both E and M steps. Using multi-time-scale stochastic approximation theory, we prove that APCD converges to a correct optimum, where the standard CD is impossible to have such a guarantee due to the mean-field approximation gap in E step. Despite of such stronger theoretical guarantee of APCD, its possible drawback is on slow mixing on E step for practical purposes. To address the issue, we also design a hybrid approach applying both mean-field and MC approximations in E step, where it outperforms the standard mean-field-based CD in our experiments on real-world datasets.

Computing partition function is the most important statistical inference task arising in applications of Graphical Models (GM). Since it is computationally intractable, approximate methods have been used in practice, where mean-field (MF) and belief propagation (BP) are arguably the most popular and successful approaches of a variational type. In this paper, we propose two new variational schemes, coined Gauged-MF (G-MF) and Gauged-BP (G-BP), improving MF and BP, respectively. Both provide lower bounds for the partition function by utilizing the so-called gauge transformation which modifies factors of GM while keeping the partition function invariant. Moreover, we prove that both G-MF and G-BP are exact for GMs with a single loop of a special structure, even though the bare MF and BP perform badly in this case. Our extensive experiments indeed confirm that the proposed algorithms outperform and generalize MF and BP.

The max-product belief propagation (BP) is a popular message-passing heuristic for approximating a maximum-a-posteriori assignment in a joint distribution represented by a graphical model. In the past years, it has been shown that BP can solve a few classes of linear programming (LP) formulations to combinatorial optimization problems including maximum weight matching, shortest path, and network flow, i.e., BP can be used as a message-passing solver for certain combinatorial optimizations. However, those LPs and corresponding BP analysis are very sensitive to underlying problem setups, and it has been not clear what extent these results can be generalized to. In this paper, we obtain a generic criteria that BP converges to the optimal solution of given LP and show that it is satisfied in LP formulations associated to many classical combinatorial optimization problems including maximum weight perfect matching, shortest path, traveling salesman, cycle packing, vertex/edge cover, and network flow.

Crowdsourcing systems are popular for solving large-scale labelling tasks with low-paid (or even non-paid) workers. We study the problem of recovering the true labels from noisy crowdsourced labels under the popular Dawid-Skene model. To address this inference problem, several algorithms have recently been proposed, but the best known guarantee is still significantly larger than the fundamental limit. We close this gap under a simple but canonical scenario where each worker is assigned at most two tasks. In particular, we introduce a tighter lower bound on the fundamental limit and prove that Belief Propagation (BP) exactly matches this lower bound. The guaranteed optimality of BP is the strongest in the sense that it is information-theoretically impossible for any other algorithm to correctly la- bel a larger fraction of the tasks. In the general setting, when more than two tasks are assigned to each worker, we establish the dominance result on BP that it outperforms other existing algorithms with known provable guarantees. Experimental results suggest that BP is close to optimal for all regimes considered, while existing state-of-the-art algorithms exhibit suboptimal performances.

Recently, the problem of detecting the rumor source in a social network has been much studied, where it has been shown that the detection probability cannot be beyond 31% even for regular trees. In this paper, we study the impact of an anti-rumor on the rumor source detection. We first show a negative result: the anti-rumor's diffusion does not increase the detection probability under Maximum-Likelihood-Estimator (MLE) when the number of infected nodes are sufficiently large by passive diffusion that the anti-rumor starts to be spread by a special node, called the protector, after is reached by the rumor. We next consider the case when the distance between the rumor source and the protector follows a certain type of distribution, but its parameter is hidden. Then, we propose the following learning algorithm: a) learn the distance distribution parameters under MLE, and b) detect the rumor source under Maximum-A-Posterior-Estimator (MAPE) based on the learnt parameters. We provide an analytic characterization of the rumor source detection probability for regular trees under the proposed algorithm, where MAPE outperforms MLE by up to 50% for 3-regular trees and by up to 63% when the degree of the regular tree becomes large. We demonstrate our theoretical findings through numerical results, and further present the simulation results for general topologies (e.g., Facebook and US power grid networks) even without knowledge of the distance distribution, showing that under a simple protector placement algorithm, MAPE produces the detection probability much larger than that by MLE.

In the past years, the centralized management of dense wireless local area networks has been emerged as a powerful paradigm for improving energy efficiency as well as avoiding severe interference. In this paper, we study the joint optimization on power-operation modes in access points (APs), channel selections and user-AP associations for improving energy efficiency and avoiding interference without sacrificing users' demands. To this end, we first formulate it as a mixed-integer programming using the popular Lyapunov approach, but it turns out to be computationally intractable, i.e., NP-hard. To address the issue, we propose a polynomial-time approximation algorithm and prove that it achieves a constant-factor approximation guarantee under mild assumptions. The main novelty underlying our algorithm design is based on a linear programming relaxation combining with two different greedy rounding schemes, where each achieves a constant-factor approximation in different regimes of parameters. We verify the performance of the proposed algorithm via extensive simulations and also demonstrate its practicability by implementing it at commercial APs using a Software-defined Networking framework, which shows that it reduces the wasted energy significantly while maintaining even higher throughput.

In various online/offline networked environments, it is very popular that the system can benefit from coordinating actions of two interacting nodes, but incur some cost due to such coordination. Examples include a wireless sensor networks with duty cycling, where a sensor node consumes a certain amount of energy when it is awake, but a coordinated operation of sensors enables some meaningful tasks, e.g., sensed data forwarding, collaborative sensing of a phenomenon, or efficient decision of further sensing actions. In this paper, we formulate an optimization problem that captures the amount of coordination gain at the cost of node activation over networks. This problem is challenging since the target utility is a function of the long-term time portion of the inter-coupled activations of two adjacent nodes, and thus a standard Lagrange duality theory is hard to apply to obtain a distributed decomposition as in the standard NUM (Network Utility Maximization). We propose a fully-distributed algorithm that requires only one-hop message passing. Our approach is inspired by a control of Ising model in statistical physics, and the proposed algorithm is motivated by a stochastic approximation method that runs a Markov chain incompletely over time, but provably guarantees its convergence to the optimal solution. We validate our theoretical findings on convergence and optimality through extensive simulations under various scenarios.

Markov Chain Monte Carlo (MCMC) and Belief Propagation (BP) are the most popular algorithms for computational inference in Graphical Models (GM). In principle, MCMC is an exact probabilistic method which, however, often suffers from exponentially slow mixing. In contrast, BP is a deterministic method, which is typically fast, empirically very successful, however in general lacking control of accuracy over loopy graphs. In this paper, we introduce MCMC algorithms correcting the approximation error of BP, i.e., we provide a way to compensate for BP errors via a consecutive BP-aware MCMC. Our framework is based on the Loop Calculus (LC) approach which allows to express the BP error as a sum of weighted generalized loops. Although the full series is computationally intractable, it is known that a truncated series, summing up all 2-regular loops, is computable in polynomial-time for planar pair-wise binary GMs and it also provides a highly accurate approximation empirically. Motivated by this, we, first, propose a polynomial-time approximation MCMC scheme for the truncated series of general (non-planar) pair-wise binary models. Our main idea here is to use the Worm algorithm, known to provide fast mixing in other (related) problems, and then design an appropriate rejection scheme to sample 2-regular loops. Furthermore, we also design an efficient rejection-free MCMC scheme for approximating the full series. The main novelty underlying our design is in utilizing the concept of cycle basis, which provides an efficient decomposition of the generalized loops. In essence, the proposed MCMC schemes run on transformed GM built upon the non-trivial BP solution, and our experiments show that this synthesis of BP and MCMC outperforms both direct MCMC and bare BP schemes.

Trapping sets strongly degrade performance of low-density parity check (LDPC) codes in the low-error-rate region. This creates significant difficulties for the deployment of LDPC codes to low-error-rate applications such as storage and wireless systems with no or limited retransmission options. We propose a novel technique for breaking trapping sets based on collaborative decoding that utilizes two different decoding modes. While the main decoding mode executes message passing based on the original parity check matrix of the corresponding LDPC code, the sub-decoding mode operates on a modified parity check matrix formed by removing a portion of check nodes in the factor graph representation of the given code. The modified parity check matrix is designed to promote a passing of correct information into erroneous variable nodes in the trapping set. Theoretical properties of the proposed trapping-set-breaking technique have been established based on the notion of the improved separation for the trapped variable nodes. Simulation results show that the proposed collaborative LDPC decoding scheme switching between the two decoding modes back and forth effectively breaks dominant trapping sets of various known types of regular and irregular LDPC codes.

A variety of models have been proposed and analyzed to understand how a new innovation (e.g., a technology, a product, or even a behavior) diffuses over a social network, broadly classified into either of epidemic-based or game-based ones. In this paper, we consider a game-based model, where each individual makes a selfish, rational choice in terms of its payoff in adopting the new innovation, but with some noise. We address the following two questions on the diffusion speed of a new innovation under the game-based model: (1) what is a good subset of individuals to seed for reducing the diffusion time significantly, i.e., convincing them to preadopt a new innovation and (2) how much diffusion time can be reduced by such a good seeding. For (1), we design near-optimal polynomial-time seeding algorithms for three representative classes of social network models, Erdös-Rényi, planted partition and geometrically structured graphs, and provide their performance guarantees in terms of approximation and complexity. For (2), we asymptotically quantify the diffusion time for these graph topologies; further derive the seed budget threshold above which the diffusion time is dramatically reduced, i.e., phase transition of diffusion time. Furthermore, based on our theoretical findings, we propose a practical seeding algorithm, called Practical Partitioning and Seeding (PrPaS) and demonstrate that PrPaS outperforms other baseline algorithms in terms of the diffusion speed over a real social network topology. We believe that our results provide new insights on how to seed over a social network depending on its connectivity structure, where individuals rationally adopt a new innovation.

In the past few years, an exciting progress has been made on CSMA (Carrier Sense Multiple Access) algorithms that achieve throughput and utility optimality for wireless networks. However, most of these algorithms are known to exhibit poor delay performance making them impractical for implementation. Recently, several papers have addressed the delay issue of CSMA and yet, most of them are limited, in the sense that they focus merely on specific network scenarios with certain conditions rather than general network topology, achieve low delay at the cost of throughput reduction, or lack rigorous provable guarantees. In this paper, we focus on the recent idea of exploiting multiple channels (actually or virtually) for delay reduction in CSMA, and prove that it is per-link delay order-optimal, i.e., O(1)-asymptotic-delay per link, if the number of virtual channels is logarithmic with respect to mixing time of the underlying CSMA Markov chain. The logarithmic number is typically small, i.e., at most linear with respect to the network size. In other words, our contribution provides not only a provable framework for the multiple-channel based CSMA, but also the required explicit number of virtual-multi-channels, which is of great importance for actual implementation. The key step of our analytic framework lies in using quadratic Lyapunov functions in conjunction with (recursively applying) Lindley equation and Azuma's inequality for obtaining an exponential decaying property in certain queueing dynamics. We believe that our technique is of broader interest in analyzing the delay performance of queueing systems with multiple periodic schedulers.

Recent studies on MAC scheduling have shown that carrier sense multiple access (CSMA) algorithms can be throughput optimal for arbitrary wireless network topology. However, these results are highly sensitive to the underlying assumption on `static' or `fixed' system conditions. For example, if channel conditions are time-varying, it is unclear how each node can adjust its CSMA parameters, so-called backoff and channel holding times, using its local channel information for the desired high performance. In this paper, we study `channel-aware' CSMA (A-CSMA) algorithms in time-varying channels, where they adjust their parameters as some function of the current channel capacity. First, we assume that backoff rates can be arbitrary large and show that the achievable rate region of A-CSMA equals to the maximum rate region if and only if the function is exponential. Furthermore, given an exponential function in A-CSMA, we design updating rules for their parameters, which achieve throughput optimality for an arbitrary wireless network topology. They are the first CSMA algorithms in the literature which are proved to be throughput optimal under time-varying channels. Moreover, we also consider the case when back-off rates of A-CSMA are restricted compared to the speed of channel variations, and characterize the throughput performance of A-CSMA in terms of the underlying wireless network topology. Our results not only guide a high-performance design on MAC scheduling under highly time-varying scenarios, but also provide new insights on the performance of CSMA algorithms in relation to their backoff rates and underlying network topologies.

Graphical Model (GM) has provided a popular framework for big data analytics because it often lends itself to distributed and parallel processing by utilizing graph-based ‘local’ structures. It models correlated random variables where in particular, the max-product Belief Propagation (BP) is the most popular heuristic to compute the most-likely assignment in GMs. In the past years, it has been proven that BP can solve a few classes of combinatorial optimization problems under certain conditions. Motivated by this, we explore the prospect of using BP to solve generic combinatorial optimization problems. The challenge is that, in practice, BP may converge very slowly and even if it does converge, the BP decision often violates the constraints of the original problem. This paper proposes a generic framework that enables us to apply BP-based algorithms to compute an approximate feasible solution for an arbitrary combinatorial optimization task. The main novel ingredients include (a) careful initialization of BP messages, (b) hybrid damping on BP updates, and (c) post-processing using BP beliefs. Utilizing the framework, we develop parallel algorithms for several large-scale combinatorial optimization problems including maximum weight matching, vertex cover and independent set. We demonstrate that our framework delivers high approximation ratio, speeds up the process by parallelization, and allows large-scale processing involving billions of variables.

Logarithms of determinants of large positive definite matrices appear ubiquitously in machine learning applications including Gaussian graphical and Gaussian process models, partition functions of discrete graphical models, minimum-volume ellipsoids and metric and kernel learning. Log-determinant computation involves the Cholesky decomposition at the cost cubic in the number of variables (i.e., the matrix dimension), which makes it prohibitive for large-scale applications. We propose a linear-time randomized algorithm to approximate log-determinants for very large-scale positive definite and general non-singular matrices using a stochastic trace approximation, called the Hutchinson method, coupled with Chebyshev polynomial expansions that both rely on efficient matrix-vector multiplications. We establish rigorous additive and multiplicative approximation error bounds depending on the condition number of the input matrix. In our experiments, the proposed algorithm can provide very high accuracy solutions at orders of magnitude faster time than the Cholesky decomposition and Shur completion, and enables us to compute log-determinants of matrices involving tens of millions of variables.

We study how an innovation (e.g., product or technology) diffuses over a social network when individuals strategically make selfish, rational choices in adopting the new innovation. This diffusion has been studied by modeling individuals' interactions with a noisy best response dynamic over a networked coordination game, but mainly in the nonprogressive setup. In this paper, we study the case when people are progressive, i.e., never going back to the old technology once the new technology is chosen, where such a progressive behavior is explained using the notion of sunk cost fallacy in social psychology. Our main focus is on the diffusion time, i.e., time till all choose the new innovation. To this end, we first provide a combinatorial characterization of the diffusion time that corresponds to the time reaching the absorbing state in a Markov chain. Based on this, we propose a polynomial-time algorithm that computes the diffusion time, where such a task is known to be computationally intractable in the non-progressive diffusion. Second, we asymptotically quantify the diffusion times for a class of well-known social graph topologies, and compare them to those under the non-progressive diffusion. Finally, we study the impact of seeding to speed up the diffusion in the progressive setup, and show that the diffusion speed is impossible to significantly accelerate with just a small-budget seeding, which is in part in stark contrast to that in the non-progressive diffusion. Our results provide not only understandings on the progressive strategic diffusion in a social network, but also computational tractability on other related problems, e.g., seeding, which we believe should be of broader interest in the future.

Max-product Belief Propagation (BP) is a popular message-passing algorithm for computing a Maximum-A-Posteriori (MAP) assignment over a distribution represented by a Graphical Model (GM). It has been shown that BP can solve a number of combinatorial optimization problems including minimum weight matching, shortest path, network flow and vertex cover under the following common assumption: the respective Linear Programming (LP) relaxation is tight, i.e., no integrality gap is present. However, when LP shows an integrality gap, no model has been known which can be solved systematically via sequential applications of BP. In this paper, we develop the first such algorithm, coined Blossom-BP, for solving the minimum weight matching problem over arbitrary graphs. Each step of the sequential algorithm requires applying BP over a modified graph constructed by contractions and expansions of blossoms, i.e., odd sets of vertices. Our scheme guarantees termination in O(n^2) of BP runs, where n is the number of vertices in the original graph. In essence, the Blossom-BP offers a distributed version of the celebrated Edmonds' Blossom algorithm by jumping at once over many sub-steps with a single BP. Moreover, our result provides an interpretation of the Edmonds' algorithm as a sequence of LPs.

Carrier sense multiple access (CSMA), which resolves contentions over wireless networks in a fully distributed fashion, has recently gained a lot of attentions since it has been proved that appropriate control of CSMA parameters guarantees optimality in terms of stability (i.e., scheduling) and system-wide utility (i.e., scheduling and congestion control). Most CSMA-based algorithms rely on the popular Markov chain Monte Carlo technique, which enables one to find optimal CSMA parameters through iterative loops of simulation-and-update. However, such a simulation-based approach often becomes a major cause of exponentially slow convergence, being poorly adaptive to flow/topology changes. In this paper, we develop distributed iterative algorithms which produce approximate solutions with convergence in polynomial time for both stability and utility maximization problems. In particular, for the stability problem, the proposed distributed algorithm requires, somewhat surprisingly, only one iteration among links. Our approach is motivated by the Bethe approximation (introduced by Yedidia, Freeman, and Weiss) allowing us to express approximate solutions via a certain nonlinear system with polynomial size. Our polynomial convergence guarantee comes from directly solving the nonlinear system in a distributed manner, rather than multiple simulation-and-update loops in existing algorithms. We provide numerical results to show that the algorithm produces highly accurate solutions and converges much faster than the prior ones.

This paper considers a hybrid peer-to-peer (p2p) system, a dynamic distributed caching system with an authoritative server dispensing contents only if the contents fail to be found by searching an unstructured p2p system. We study the case when some peers may not be fully cooperative in the search process and examine the impact of various noncooperative behaviors in the aspect of scalability, more specifically average server load and average peer load as the peer population size increases. We categorize selfish peers into three classes: impatient peers that directly query the server without searching the p2p system, non-forwarders that refuse to forward query requests, and non-resolvers that refuse to share contents. It is shown that in the hybrid p2p system, impatient and/or non-forwarding behaviors prevent the system from scaling well because of the high server load, while the system scales well under the non-resolving selfish peers. Our study implies that the hybrid p2p system does not mandate an incentive mechanism for content sharing, which is in stark contrast to unstructured p2p systems, where incentivizing peers to share contents is known to be a key factor for the system's scalability.

Game-theoretic modeling and equilibrium analysis have provided valuable insights into the design of robust local control rules for the individual agents in multi-agent systems, e.g., Internet congestion control, road transportation networks, etc. In this paper, we introduce a non-cooperative MAC (Medium Access Control) game for wireless networks and propose new fully-distributed CSMA (Carrier Sense Multiple Access) learning algorithms that are probably optimal in the sense that their long-term throughputs converge to the optimal solution of a utility maximization problem over the maximum throughput region. The most significant part of our approach lies in introducing a novel cost function in agents' utilities so that the proposed game admits an ordinal potential function with (asymptotically) no price-of-anarchy. The game formulation naturally leads to known game-based learning rules to find a Nash equilibrium, but they are computationally inefficient and often require global information. Towards our goal of fully-distributed operation, we propose new fully-distributed learning algorithms by utilizing a unique property of CSMA that enables each link to estimate its temporary link throughput without message passing for the applied CSMA parameters. The proposed algorithms can be thought as `stochastic approximations' to the standard learning rules, which is a new feature in our work, not prevalent in other traditional game-theoretic approaches. We show that they converge to a Nash equilibrium, which is a utility-optimal point, numerically evaluate their performance to support our theoretical findings and further examine various features such as convergence speed and its tradeoff with efficiency.

In the past few years, an exciting progress has been made on CSMA (Carrier Sense Multiple Access) algorithms that achieve throughput and utility optimality for wireless networks. However, most of these algorithms are known to exhibit poor delay performance making them impractical for implementation. Recently, several papers have addressed the delay issue of CSMA and yet, most of them are limited, in the sense that they focus merely on specific network scenarios with certain conditions rather than general network topology, achieve low delay at the cost of throughput reduction, or lack rigorous provable guarantees. In this paper, we focus on the recent idea of exploiting multiple channels (actually or virtually) for delay reduction in CSMA, and prove that it isper-link delay orderoptimal, i.e.,O(1)-asymptotic-delay per link, if the number of virtual channels is logarithmic with respect to mixing time of the underlying CSMA Markov chain. The logarithmic number is typically small, i.e., at most linear with respect to the network size. In other words, our contribution provides not only a provable framework for the multiple-channel based CSMA, but also the required explicit number of virtual-multi-channels, which is of great importance for actual implementation. The key step of our analytic framework lies in using quadratic Lyapunov functions in conjunction with (recursively applying) Lindley equation and Azuma's inequality for obtaining an exponential decaying property in certain queueing dynamics. We believe that our technique is of broad interest in analyzing the delay performances of other general queueing systems.

A variety of models have been proposed and analyzed to understand how a new innovation (e.g., a technology, a product, or even a behavior) diffuses over a social network, broadly classified into either of epidemic-based or game-based ones. In this paper, we consider a game-based model, where each individual makes a selfish, rational choice in terms of its payoff in adopting the new innovation, but with some noise. We study how diffusion effect can be maximized by seeding a subset of individuals (within a given budget), i.e., convincing them to pre-adopt a new innovation. In particular, we aim at finding `good' seeds for minimizing the time to infect all others, i.e., diffusion speed maximization. To this end, we design polynomial-time approximation algorithms for three representative classes, Erdőos-Réenyi, planted partition and geometrically structured graph models, which correspond to globally well-connected, locally well-connected with large clusters and locally well-connected with small clusters, respectively, provide their performance guarantee in terms of approximation and complexity. First, for the dense Erdős-Rényi and planted partition graphs, we show that an arbitrary seeding and a simple seeding proportional to the size of clusters are almost optimal with high probability. Second, for geometrically structured sparse graphs, including planar and d-dimensional graphs, our algorithm that (a) constructs clusters, (b) seeds the border individuals among clusters, and (c) greedily seeds inside each cluster always outputs an almost optimal solution. We validate our theoretical findings with extensive simulations under a real social graph. We believe that our results provide new practical insights on how to seed over a social network depending on its connection structure, where individuals rationally adopt a new innovation. To our best knowledge, we are the first to study such diffusion speed maximization on the game-based diffusion, while the extensive research efforts have been made in epidemic-based models, often referred to as influence maximization.

Since Tassiulas and Ephremides proposed the maximum weight scheduling algorithm of throughput-optimality for constrained queueing networks in 1992, extensive research efforts have been made for resolving its high complexity issue under various directions. In this paper, we resolve this issue by developing a generic framework for designing throughput-optimal and low-complexity scheduling algorithms. Under the framework, an algorithm updates current schedules via an interaction with a given oracle system that generates a solution of a certain discrete optimization problem in a finite number of interactive queries. The complexity of the resulting algorithm is decided by the number of operations required for an oracle processing a single query, which is typically very small. Somewhat surprisingly, we prove that an algorithm using any such oracle is throughput-optimal for general constrained queueing network models that arise in the context of emerging large-scale communication networks. To our best knowledge, our result is the first that establishes a rigorous connection between iterative optimization methods and low-complexity scheduling algorithms, which we believe provides various future directions and new insights in both areas.

This paper resolves a common complexity issue in the Bethe approximation of statistical physics and the belief propagation (BP) algorithm of artificial intelligence. The Bethe approximation and the BP algorithm are heuristic methods for estimating the partition function and marginal probabilities in graphical models, respectively. The computational complexity of the Bethe approximation is decided by the number of operations required to solve a set of nonlinear equations, the so-called Bethe equation. Although the BP algorithm was inspired and developed independently, Yedidia, Freeman, and Weiss showed that the BP algorithm solves the Bethe equation if it converges (however, it often does not). This naturally motivates the following question to understand limitations and empirical successes of the Bethe and BP methods: is the Bethe equation computationally easy to solve? We present a message-passing algorithm solving the Bethe equation in a polynomial number of operations for general binary graphical models of n variables, where the maximum degree in the underlying graph is O(logn). Equivalently, it finds a stationary point of the Bethe free energy function. Our algorithm can be used as an alternative to BP fixing its convergence issue and is the first fully polynomial-time approximation scheme for the BP fixed-point computation in such a large class of graphical models, whereas the approximate fixed-point computation is known to be polynomial parity arguments on directed graphs (PPAD-)hard in general. We believe that our technique is of broader interest to understand the computational complexity of the cavity method in statistical physics.

This paper studies how global information affects the diffusion of innovations on a network. The diffusion of innovation is modeled by the logit dynamics of a weighted N-person coordination game among (bounded) rational users where innovations spread through users' strategic choices. We find a critical asymptotic threshold for the weight on global information where the diffusion of innovations undergoes a transition in the rate of convergence regardless of any network structure. In particular, it is found that the convergence to the pervasive adoption is slowed down by global information.

This paper considers a hybrid peer-to-peer (p2p) system, a dynamic distributed caching system with an authoritative server dispensing contents only if the contents fail to be found by searching an unstructured peer-to-peer (p2p) system. We study the case when some peers may not be fully cooperative in the search process and examine the impact of various noncooperative behaviors on the querying load on the server as the peer population size increases. We categorize selfish peers into three classes: impatient peers that directly query the server without searching the p2p system, non-forwarders that refuse to forward query requests, and non-resolvers that refuse to share contents. It is shown that in the hybrid p2p system, impatient and/or nonforwarding behaviors prevent the system from scaling well because of the high server load, while the system scales well under the non-resolving selfish peers. Our study implies that the hybrid p2p system does not mandate an incentive mechanism for content sharing, which is in stark contrast to unstructured p2p systems, where incentivizing peers to share contents is known to be a key factor for the system's scalability.

Belief Propagation (BP) is a popular, distributed heuristic for performing MAP computations in Graphical Models. BP can be interpreted, from a variational perspective, as minimizing the Bethe Free Energy (BFE). BP can also be used to solve a special class of Linear Programming (LP) problems. For this class of problems, MAP inference can be stated as an integer LP with an LP relaxation that coincides with minimization of the BFE at “zero temperature”. We generalize these prior results and establish a tight characterization of the LP problems that can be formulated as an equivalent LP relaxation of MAP inference. Moreover, we suggest an efficient, iterative annealing BP algorithm for solving this broader class of LP problems. We demonstrate the algorithm's performance on a set of weighted matching problems by using it as a cutting plane method to solve a sequence of LPs tightened by adding “blossom” inequalities.

CSMA (Carrier Sense Multiple Access), which resolves contentions over wireless networks in a fully distributed fashion, has recently gained a lot of attentions since it has been proved that appropriate control of CSMA parameters guarantees optimality in terms of system-wide utility. Most algorithms rely on the popular MCMC (Markov Chain Monte Carlo) technique, which enables one to find optimal CSMA parameters through iterative loops of simulation-and-update. However, such a simulation-based approach often becomes a major cause of exponentially slow convergence, being poorly adaptive to flow/topology changes. In this paper, we develop a distributed iterative algorithm which produces approximate solutions with convergence in polynomial time. Our approach is motivated by a scheme in statistical physics, referred to as the Bethe approximation, allowing us to express approximate solutions via a certain non-linear system with polynomial size. We provide numerical results to show that the algorithm produces highly accurate solutions and converges much faster than prior ones.

Recent studies on MAC scheduling have shown that carrier sense multiple access (CSMA) algorithms can be throughput optimal for arbitrary wireless network topology. However, these results are highly sensitive to the underlying assumption on `static' or `fixed' system conditions. For example, if channel conditions are time-varying, it is unclear how each node can adjust its CSMA parameters, so-called backoff and channel holding times, using its local channel information for the desired high performance. In this paper, we study `channel-aware' CSMA (A-CSMA) algorithms in time-varying channels, where they adjust their parameters as some function of the current channel capacity. First, we show that the achievable rate region of A-CSMA equals to the maximum rate region if and only if the function is exponential. Furthermore, given an exponential function in A-CSMA, we design updating rules for their parameters, which achieve throughput optimality for an arbitrary wireless network topology. They are the first CSMA algorithms in the literature which are proved to be throughput optimal under time-varying channels. Moreover, we also consider the case when back-off rates of A-CSMA are highly restricted compared to the speed of channel variations, and characterize the throughput performance of A-CSMA in terms of the underlying wireless network topology. Our results not only guide a high-performance design on MAC scheduling under highly time-varying scenarios, but also provide new insights on the performance of CSMA algorithms in relation to their backoff rates and underlying network topologies.

Max-product ‘belief propagation’ (BP) is a popular distributed heuristic for finding the Maximum A Posteriori (MAP) assignment in a joint probability distribution represented by a Graphical Model (GM). It was recently shown that BP converges to the correct MAP assignment for a class of loopy GMs with the following common feature: the Linear Programming (LP) relaxation to the MAP problem is tight (has no integrality gap). Unfortunately, tightness of the LP relaxation does not, in general, guarantee convergence and correctness of the BP algorithm. The failure of BP in such cases motivates reverse engineering a solution – namely, given a tight LP, can we design a ‘good’ BP algorithm. In this paper, we design a BP algorithm for the Maximum Weight Matching (MWM) problem over general graphs. We prove that the algorithm converges to the correct optimum if the respective LP relaxation, which may include inequalities associated with non-intersecting odd-sized cycles, is tight. The most significant part of our approach is the introduction of a novel graph transformation designed to force convergence of BP. Our theoretical result suggests an efficient BP-based heuristic for the MWM problem, which consists of making sequential, “cutting plane”, modifications to the underlying GM. Our experiments show that this heuristic performs as well as traditional cutting-plane algorithms using LP solvers on MWM problems.

The hard-core model has received much attention in the past couple of decades as a lattice gas model with hard constraints in statistical physics, a multicast model of calls in communication networks, and as a weighted independent set problem in combinatorics, probability and theoretical computer science. In this model, each independent set I in a graph G is weighted proportionally to λ|I|, for a positive real parameter λ. For large λ, computing the partition function (namely, the normalizing constant which makes the weighting a probability distribution on a finite graph) on graphs of maximum degree Δ ≥ 3, is a well known computationally challenging problem. More concretely, let λc(TΔ) denote the critical value for the so-called uniqueness threshold of the hard-core model on the infinite Δ-regular tree; recent breakthrough results of Dror Weitz (2006) and Allan Sly (2010) have identified λc(TΔ) as a threshold where the hardness of estimating the above partition function undergoes a computational transition. We focus on the well-studied particular case of the square lattice Z2, and provide a new lower bound for the uniqueness threshold, in particular taking it well above λc(T4). Our technique refines and builds on the tree of self-avoiding walks approach of Weitz, resulting in a new technical sufficient criterion (of wider applicability) for establishing strong spatial mixing (and hence uniqueness) for the hard-core model. Our new criterion achieves better bounds on strong spatial mixing when the graph has extra structure, improving upon what can be achieved by just using the maximum degree. Applying our technique to Z2we prove that strong spatial mixing holds for all λ <; 2.3882, improving upon the work of Weitz that held for λ <; 27/16 = 1.6875. Our results imply a fully-polynomial deterministic approximation algorithm for estimating the partition function, as well as rapid mixing of the associated Glauber dynamics to sample from the hard-core distribution.

Consider a wireless network of n nodes represented by a (undirected) graph G where an edge (i,j) models the fact that transmissions of i and j interfere with each other, i.e. simultaneous transmissions of i and j become unsuccessful. Hence it is required that at each time instance a set of non-interfering nodes (corresponding to an independent set in G) access the wireless medium. To utilize wireless resources efficiently, it is required to arbitrate the access of medium among interfering nodes properly. Moreover, to be of practical use, such a mechanism is required to be totally distributed as well as simple. As the main result of this paper, we provide such a medium access algorithm. It is randomized, totally distributed and simple: each node attempts to access medium at each time with probability that is a function of its local information. We establish efficiency of the algorithm by showing that the corresponding network Markov chain is positive recurrent as long as the demand imposed on the network can be supported by the wireless network (using any algorithm). In that sense, the proposed algorithm is optimal in terms of utilizing wireless resources. The algorithm is oblivious to the network graph structure, in contrast with the so-called polynomial back-off algorithm by Hastad-Leighton-Rogoff (STOC '87, SICOMP '96) that is established to be optimal for the complete graph and bipartite graphs (by Goldberg-MacKenzie (SODA '96, JCSS '99)).

We consider the #P-complete problem of counting the number of independent sets in a given graph. Our interest is in understanding the effectiveness of the popular belief propagation (BP) heuristic. BP is a simple iterative algorithm that is known to have at least one fixed point, where each fixed point corresponds to a stationary point of the Bethe free energy (introduced by Yedidia, Freeman, and Weiss [IEEE Trans. Inform. Theory, 51 (2004), pp. 2282–2312] in recognition of Bethe’s earlier work in 1935). The evaluation of the Bethe free energy at such a stationary point (or BP fixed point) leads to the Bethe approximation for the number of independent sets of the given graph. BP is not known to converge in general, nor is an efficient, convergent procedure for finding stationary points of the Bethe free energy known. Furthermore, the effectiveness of the Bethe approximation is not well understood. As the first result of this paper we propose a BP-like algorithm that always converges to a stationary point of the Bethe free energy for any graph for the independent set problem. This procedure finds an [Formula: see text]-approximate stationary point in [Formula: see text] iterations for a graph of [Formula: see text] nodes with max-degree [Formula: see text]. We study the quality of the resulting Bethe approximation using the recently developed “loop series” framework of Chertkov and Chernyak [J. Stat. Mech. Theory Exp., 6 (2006), P06009]. As this characterization is applicable only for exact stationary points of the Bethe free energy, we provide a slightly modified characterization that holds for [Formula: see text]-approximate stationary points. We establish that for any graph on [Formula: see text] nodes with max-degree [Formula: see text] and girth larger than [Formula: see text], the multiplicative error between the number of independent sets and the Bethe approximation decays as [Formula: see text] for some [Formula: see text]. This provides a deterministic counting algorithm that leads to strictly different results compared to a recent result of Weitz [in Proceedings of the Thirty-Eighth Annual ACM Symposium on Theory of Computing, ACM Press, New York, 2006, pp. 140–149]. Finally, as a consequence of our analysis we prove that the Bethe approximation is exceedingly good for a random 3-regular graph conditioned on the shortest cycle cover conjecture of Alon and Tarsi [SIAM J. Algebr. Discrete Methods, 6 (1985), pp. 345–350] being true.

Game theoretic modeling and equilibrium analysis of congestion games have provided insights in the performance of Internet congestion control, road transportation networks, etc. Despite the long history, very little is known about their transient (non equilibrium) performance. In this paper, we are motivated to seek answers to questions such as how long does it take to reach equilibrium, when the system does operate near equilibrium in the presence of dynamics, e.g. nodes join or leave. , or the tradeoff between performance and the rate of dynamics. In this pursuit, we provide three contributions in this paper. First, a novel probabilistic model to capture realistic behaviors of agents allowing for the possibility of arbitrariness in conjunction with rationality. Second, evaluation of (a) time to converge to equilibrium under this behavior model and (b) distance to Nash equilibrium. Finally, determination of tradeoff between the rate of dynamics and quality of performance (distance to equilibrium) which leads to an interesting uncertainty principle. The novel technical ingredients involve analysis of logarithmic Sobolov constant of Markov process with time varying state space and methodically this should be of broader interest in the context of dynamical systems.

In the past year or so, an exciting progress has led to throughput optimal design of CSMA-based algorithms for wireless networks. However, such an algorithm suffers from very poor delay performance. A recent work suggests that it is impossible to design a CSMA-like simple algorithm that is throughput optimal and induces low delay for any wireless network. However, wireless networks arising in practice are formed by nodes placed, possibly arbitrarily, in some geographic area.

This paper provides proofs of the rate stability, Harris recurrence, and ε-optimality of carrier sense multiple access (CSMA) algorithms where the random access (or backoff) parameter of each node is adjusted dynamically. These algorithms require only local information and they are easy to implement. The setup is a network of wireless nodes with a fixed conflict graph that identifies pairs of nodes whose simultaneous transmissions conflict. The paper studies two algorithms. The first algorithm schedules transmissions to keep up with given arrival rates of packets. The second algorithm controls the arrivals in addition to the scheduling and attempts to maximize the sum of the utilities, in terms of the rates, of the packet flows at different nodes. For the first algorithm, the paper proves rate stability for strictly feasible arrival rates and also Harris recurrence of the queues. For the second algorithm, the paper proves the ε-optimality in terms of the utilities of the allocated rates. Both algorithms are iterative and we study two versions of each of them. In the first version, both operate with strictly local information but have relatively weaker performance guarantees; under the second version, both provide stronger performance guarantees by utilizing the additional information of the number of nodes in the network.

Motivated by applications of distributed linear estimation, distributed control, and distributed optimization, we consider the question of designing linear iterative algorithms for computing the average of numbers in a network. Specifically, our interest is in designing such an algorithm with the fastest rate of convergence given the topological constraints of the network. As the main result of this paper, we design an algorithm with the fastest possible rate of convergence using a nonreversible Markov chain on the given network graph. We construct such a Markov chain by transforming the standard Markov chain, which is obtained using the Metropolis-Hastings method. We call this novel transformationpseudo-lifting. We apply our method to graphs with geometry, or graphs with doubling dimension. Specifically, the convergence time of our algorithm (equivalently, the mixing time of our Markov chain) is proportional to the diameter of the network graph and hence optimal. As a byproduct, our result provides the fastest mixing Markov chain given the network topological constraints, and should naturally find their applications in the context of distributed optimization, estimation and control.