Xinbei Ma, Yiting Wang, Yao Yao, Tongxin Yuan, Aston Zhang, Zhuosheng Zhang, Hai Zhao. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

The development and evaluation of Large Language Models (LLMs) have largely focused on individual capabilities. However, this overlooks the intersection of multiple abilities across different types of expertise that are often required for real-world tasks, which we term **cross capabilities**. To systematically explore this concept, we first define seven core individual capabilities and then pair them to form seven common cross capabilities, each supported by a manually constructed taxonomy. Building on these definitions, we introduce *CrossEval*, a benchmark comprising 1,400 human-annotated prompts, with 100 prompts for each individual and cross capability. To ensure reliable evaluation, we involve expert annotators to assess 4,200 model responses, gathering 8,400 human ratings with detailed explanations to serve as reference examples. Our findings reveal that current LLMs consistently exhibit the ``Law of the Weakest Link,'' where cross-capability performance is significantly constrained by the weakest component. Across 58 cross-capability scores from 17 models, 38 scores are lower than all individual capabilities, while 20 fall between strong and weak, but closer to the weaker ability. These results highlight LLMs' underperformance in cross-capability tasks, emphasizing the need to identify and improve their weakest capabilities as a key research priority. The code, benchmarks, and evaluations are available on our [project website](https://www.llm-cross-capabilities.org).

Joongwon Kim, Anirudh Goyal, Aston Zhang, Bo Xiong, Rui Hou, Melanie Kambadur, Dhruv Mahajan, Hannaneh Hajishirzi, Liang Tan. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Yue Yu, Zhengxing Chen, Aston Zhang, Liang Tan, Chenguang Zhu, Richard Yuanzhe Pang, Yundi Qian, Xuewei Wang, Suchin Gururangan, Chao Zhang, Melanie Kambadur, Dhruv Mahajan, Rui Hou. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Hypercomplex neural networks have proven to reduce the overall number of parameters while ensuring valuable performance by leveraging the properties of Clifford algebras. Recently, hypercomplex linear layers have been further improved by involving efficient parameterized Kronecker products. In this article, we define the parameterization of hypercomplex convolutional layers and introduce the family of parameterized hypercomplex neural networks (PHNNs) that are lightweight and efficient large-scale models. Our method grasps the convolution rules and the filter organization directly from data without requiring a rigidly predefined domain structure to follow. PHNNs are flexible to operate in any user-defined or tuned domain, from 1-D to [Formula: see text] regardless of whether the algebra rules are preset. Such a malleability allows processing multidimensional inputs in their natural domain without annexing further dimensions, as done, instead, in quaternion neural networks (QNNs) for 3-D inputs like color images. As a result, the proposed family of PHNNs operates with 1/n free parameters as regards its analog in the real domain. We demonstrate the versatility of this approach to multiple domains of application by performing experiments on various image datasets and audio datasets in which our method outperforms real and quaternion-valued counterparts. Full code is available at: https://github.com/eleGAN23/HyperNets.

Diffusion models that are based on iterative denoising have been recently proposed and leveraged in various generation tasks like image generation.Whereas, as a way inherently built for continuous data, existing diffusion models still have some limitations in modeling discrete data, e.g., languages.For example, the generally used Gaussian noise can not handle the discrete corruption well, and the objectives in continuous spaces fail to be stable for textual data in the diffusion process especially when the dimension is high.To alleviate these issues, we introduce a novel diffusion model for language modeling, Masked-Diffusion LM, with lower training cost and better performances, inspired by linguistic features in languages.Specifically, we design a linguistic-informed forward process which adds corruptions to the text through strategically soft-masking to better noise the textual data.Also, we directly predict the categorical distribution with cross-entropy loss function in every diffusion step to connect the continuous space and discrete space in a more efficient and straightforward way.Through experiments on 5 controlled generation tasks, we demonstrate that our Masked-Diffusion LM can achieve better generation quality than the state-of-the-art diffusion models with better efficiency.Code is available at https://github. com/SALT-NLP/Masked_Diffusioin_LM.

Spurred by advancements in scale, large language models (LLMs) have demonstrated the ability to perform a variety of natural language processing (NLP) tasks zero-shot-i.e., without adaptation on downstream data.Recently, the debut of ChatGPT 1 has drawn a great deal of attention from the natural language processing (NLP) community due to the fact that it can generate high-quality responses to human input and self-correct previous mistakes based on subsequent conversations.However, it is not yet known whether ChatGPT can serve as a generalist model that can perform many NLP tasks zero-shot.In this work, we empirically analyze the zero-shot learning ability of ChatGPT by evaluating it on 20 popular NLP datasets covering 7 representative task categories.With extensive empirical studies, we demonstrate both the effectiveness and limitations of the current version of ChatGPT.We find that ChatGPT performs well on many tasks favoring reasoning capabilities (e.g., arithmetic reasoning) while it still faces challenges when solving specific tasks such as sequence tagging.We additionally provide in-depth analysis through qualitative case studies.

This work proposes POMP, a prompt pre-training method for vision-language models. Being memory and computation efficient, POMP enables the learned prompt to condense semantic information for a rich set of visual concepts with over twenty-thousand classes. Once pre-trained, the prompt with a strong transferable ability can be directly plugged into a variety of visual recognition tasks including image classification, semantic segmentation, and object detection, to boost recognition performances in a zero-shot manner. Empirical evaluation shows that POMP achieves state-of-the-art performances on 21 datasets, e.g., 67.0% average accuracy on 10 classification datasets (+3.1% compared to CoOp) and 84.4 hIoU on open-vocabulary Pascal VOC segmentation (+6.9 compared to ZSSeg).

Transformers are central in modern natural language processing and computer vision applications. Despite recent works devoted to reducing the quadratic cost of such models with respect to sequence length, dealing with ultra long sequences (e.g., $>$16K tokens) remains challenging. Applications such as answering questions based on a book or summarizing a scientific article are inefficient or infeasible. Here, we propose to significantly improve the efficiency of Transformers for ultra long sequences, by compressing the sequence into a much smaller representation at each layer. Specifically, by exploiting the fact that in many tasks, only a small subset of special tokens, which we call VIP-tokens, are most relevant to the final prediction, we propose a VIP-token centric compression (VCC) scheme which selectively compresses the sequence based on their impact on approximating the representation of the VIP-tokens. Compared with competitive baselines, our algorithm is not only efficient (achieving more than $3\times$ compute efficiency gain compared to baselines on 4K and 16K lengths), but also offers competitive/better performance on a large number of tasks. Further, we show that our algorithm scales to 128K tokens (or more) while consistently offering accuracy improvement. Code is available at https://github.com/mlpen/VCC.

Adversarial training (AT) defends deep neural networks against adversarial attacks. One challenge that limits its practical application is the performance degradation on clean samples. A major bottleneck identified by previous works is the widely used batch normalization (BN), which struggles to model the different statistics of clean and adversarial training samples in AT. Although the dominant approach is to extend BN to capture this mixture of distribution, we propose to completely eliminate this bottleneck by removing all BN layers in AT. Our normalizer-free robust training (NoFrost) method extends recent advances in normalizer-free networks to AT for its unexplored advantage on handling the mixture distribution challenge. We show that NoFrost achieves adversarial robustness with only a minor sacrifice on clean sample accuracy. On ImageNet with ResNet50, NoFrost achieves $74.06%$ clean accuracy, which drops merely $2.00%$ from standard training. In contrast, BN-based AT obtains $59.28%$ clean accuracy, suffering a significant $16.78%$ drop from standard training. In addition, NoFrost achieves a $23.56%$ adversarial robustness against PGD attack, which improves the $13.57%$ robustness in BN-based AT. We observe better model smoothness and larger decision margins from NoFrost, which make the models less sensitive to input perturbations and thus more robust. Moreover, when incorporating more data augmentations into NoFrost, it achieves comprehensive robustness against multiple distribution shifts. Code and pre-trained models are public at https://github.com/amazon-research/normalizer-free-robust-training.

Existing out-of-distribution (OOD) detection methods are typically benchmarked on training sets with balanced class distributions. However, in real-world applications, it is common for the training sets to have long-tailed distributions. In this work, we first demonstrate that existing OOD detection methods commonly suffer from significant performance degradation when the training set is long-tail distributed. Through analysis, we posit that this is because the models struggle to distinguish the minority tail-class in-distribution samples, from the true OOD samples, making the tail classes more prone to be falsely detected as OOD. To solve this problem, we propose Partial and Asymmetric Supervised Contrastive Learning (PASCL), which explicitly encourages the model to distinguish between tail-class in-distribution samples and OOD samples. To further boost in-distribution classification accuracy, we propose Auxiliary Branch Finetuning, which uses two separate branches of BN and classification layers for anomaly detection and in-distribution classification, respectively. The intuition is that in-distribution and OOD anomaly data have different underlying distributions. Our method outperforms previous state-of-the-art method by $1.29%$, $1.45%$, $0.69%$ anomaly detection false positive rate (FPR) and $3.24%$, $4.06%$, $7.89%$ in-distribution classification accuracy on CIFAR10-LT, CIFAR100-LT, and ImageNet-LT, respectively. Code and pre-trained models are available at https://github.com/amazon-research/long-tailed-ood-detection.

Deep neural networks (DNNs) are vulnerable to backdoor attacks. Previous works have shown it extremely challenging to unlearn the undesired backdoor behavior from the network, since the entire network can be affected by the backdoor samples. In this paper, we propose a brand-new backdoor defense strategy, which makes it much easier to remove the harmful influence of backdoor samples from the model. Our defense strategy, \emph{Trap and Replace}, consists of two stages. In the first stage, we bait and trap the backdoors in a small and easy-to-replace subnetwork. Specifically, we add an auxiliary image reconstruction head on top of the stem network shared with a light-weighted classification head. The intuition is that the auxiliary image reconstruction task encourages the stem network to keep sufficient low-level visual features that are hard to learn but semantically correct, instead of overfitting to the easy-to-learn but semantically incorrect backdoor correlations. As a result, when trained on backdoored datasets, the backdoors are easily baited towards the unprotected classification head, since it is much more vulnerable than the shared stem, leaving the stem network hardly poisoned. In the second stage, we replace the poisoned light-weighted classification head with an untainted one, by re-training it from scratch only on a small holdout dataset with clean samples, while fixing the stem network. As a result, both the stem and the classification head in the final network are hardly affected by backdoor training samples. We evaluate our method against ten different backdoor attacks. Our method outperforms previous state-of-the-art methods by up to $20.57\%$, $9.80\%$, and $13.72\%$ attack success rate and on-average $3.14\%$, $1.80\%$, and $1.21\%$ clean classification accuracy on CIFAR10, GTSRB, and ImageNet-12, respectively. Code is available at https://github.com/VITA-Group/Trap-and-Replace-Backdoor-Defense.

This paper reviews the novel concept of a controllable variational autoencoder (ControlVAE), discusses its parameter tuning to meet application needs, derives its key analytic properties, and offers useful extensions and applications. ControlVAE is a new variational autoencoder (VAE) framework that combines automatic control theory with the basic VAE to stabilize the KL-divergence of VAE models to a specified value. It leverages a non-linear PI controller, a variant of the proportional-integral-derivative (PID) controller, to dynamically tune the weight of the KL-divergence term in the evidence lower bound (ELBO) using the output KL-divergence as feedback. This allows us to precisely control the KL-divergence to a desired value (set point) that is effective in avoiding posterior collapse and learning disentangled representations. While prior work developed alternative techniques for controlling the KL divergence, we show that our PI controller has better stability properties and thus better convergence, thereby producing better disentangled representations from finite training data. In order to improve the ELBO of ControlVAE over that of the regular VAE, we provide a simplified theoretical analysis to inform the choice of set point for the KL-divergence of ControlVAE. We evaluate the proposed method on three tasks: image generation, language modeling, and disentangled representation learning. The results show that ControlVAE can achieve much better reconstruction quality than the other methods for comparable disentanglement. On the language modeling task, our method can avoid posterior collapse (KL vanishing) and improve the diversity of generated text. Moreover, it can change the optimization trajectory, improving the ELBO and the reconstruction quality for image generation.

With graphs rapidly growing in size and deeper graph neural networks (GNNs) emerging, the training and inference of GNNs become increasingly expensive. Existing network weight pruning algorithms cannot address the main space and computational bottleneck in GNNs, caused by the size and connectivity of the graph. To this end, this paper first presents a unified GNN sparsification (UGS) framework that simultaneously prunes the graph adjacency matrix and the model weights, for effectively accelerating GNN inference on large-scale graphs. Leveraging this new tool, we further generalize the recently popular lottery ticket hypothesis to GNNs for the first time, by defining a graph lottery ticket (GLT) as a pair of core sub-dataset and sparse sub-network, which can be jointly identified from the original GNN and the full dense graph by iteratively applying UGS. Like its counterpart in convolutional neural networks, GLT can be trained in isolation to match the performance of training with the full model and graph, and can be drawn from both randomly initialized and self-supervised pre-trained GNNs. Our proposal has been experimentally verified across various GNN architectures and diverse tasks, on both small-scale graph datasets (Cora, Citeseer and PubMed), and large-scale datasets from the challenging Open Graph Benchmark (OGB). Specifically, for node classification, our found GLTs achieve the same accuracies with 20% 98% MACs saving on small graphs and 25% 85% MACs saving on large ones. For link prediction, GLTs lead to 48% 97% and 70% MACs saving on small and large graph datasets, respectively, without compromising predictive performance. Codes are at https://github.com/VITA-Group/Unified-LTH-GNN.

Recent advances in machine learning have largely benefited from the massive accessible training data. However, large-scale data sharing has raised great privacy concerns. In this work, we propose a novel privacy-preserving data Generative model based on the PATE framework (G-PATE), aiming to train a scalable differentially private data generator that preserves high generated data utility. Our approach leverages generative adversarial nets to generate data, combined with private aggregation among different discriminators to ensure strong privacy guarantees. Compared to existing approaches, G-PATE significantly improves the use of privacy budgets. In particular, we train a student data generator with an ensemble of teacher discriminators and propose a novel private gradient aggregation mechanism to ensure differential privacy on all information that flows from teacher discriminators to the student generator. In addition, with random projection and gradient discretization, the proposed gradient aggregation mechanism is able to effectively deal with high-dimensional gradient vectors. Theoretically, we prove that G-PATE ensures differential privacy for the data generator. Empirically, we demonstrate the superiority of G-PATE over prior work through extensive experiments. We show that G-PATE is the first work being able to generate high-dimensional image data with high data utility under limited privacy budgets ($\varepsilon \le 1$). Our code is available at https://github.com/AI-secure/G-PATE.

While standard recurrent neural networks explicitly impose a chain structure on different forms of data, they do not have an explicit bias towards recursive self-instantiation where the extent of recursion is dynamic. Given diverse and even growing data modalities (e.g., logic, algorithmic input and output, music, code, images, and language) that can be expressed in sequences and may benefit from more architectural flexibility, we propose the self-instantiated recurrent unit (Self-IRU) with a novel inductive bias towards dynamic soft recursion. On one hand, theSelf-IRU is characterized by recursive self-instantiation via its gating functions, i.e., gating mechanisms of the Self-IRU are controlled by instances of the Self-IRU itself, which are repeatedly invoked in a recursive fashion. On the other hand, the extent of the Self-IRU recursion is controlled by gates whose values are between 0 and 1 and may vary across the temporal dimension of sequences, enabling dynamic soft recursion depth at each time step. The architectural flexibility and effectiveness of our proposed approach are demonstrated across multiple data modalities. For example, the Self-IRU achieves state-of-the-art performance on the logical inference dataset [Bowman et al., 2014] even when comparing with competitive models that have access to ground-truth syntactic information.

Modeling user interests is crucial in real-world recommender systems. In this paper, we present a new user interest representation model for personalized recommendation. Specifically, the key novelty behind our model is that it explicitly models user interests as a hypercuboid instead of a point in the space. In our approach, the recommendation score is learned by calculating a compositional distance between the user hypercuboid and the item. This helps to alleviate the potential geometric inflexibility of existing collaborative filtering approaches, enabling a greater extent of modeling capability. Furthermore, we present two variants of hypercuboids to enhance the capability in capturing the diversities of user interests. A neural architecture is also proposed to facilitate user hypercuboid learning by capturing the activity sequences (e.g., buy and rate) of users. We demonstrate the effectiveness of our proposed model via extensive experiments on both public and commercial datasets. Empirical results show that our approach achieves very promising results, outperforming existing state-of-the-art.

In E-commerce, a key challenge in text generation is to find a good trade-off between word diversity and accuracy (relevance) in order to make generated text appear more natural and human-like. In order to improve the relevance of generated results, conditional text generators were developed that use input keywords or attributes to produce the corresponding text. Prior work, however, do not finely control the diversity of automatically generated sentences. For example, it does not control the order of keywords to put more relevant ones first. Moreover, it does not explicitly control the balance between diversity and accuracy. To remedy these problems, we propose a fine-grained controllable generative model, called Apex, that uses an algorithm borrowed from automatic control (namely, a variant of the proportional, integral, and derivative (PID) controller) to precisely manipulate the diversity/accuracy trade-off of generated text. The algorithm is injected into a Conditional Variational Autoencoder (CVAE), allowing Apex to control both (i) the order of keywords in the generated sentences (conditioned on the input keywords and their order), and (ii) the trade-off between diversity and accuracy. Evaluation results on real world datasets 1 show that the proposed method outperforms existing generative models in terms of diversity and relevance. Moreover, it achieves about 97% accuracy in the control of the order of keywords.

Variational Autoencoders (VAE) and their variants have been widely used in a variety of applications, such as dialog generation, image generation and disentangled representation learning. However, the existing VAE models may suffer from KL vanishing in language modeling and low reconstruction quality for disentangling. To address these issues, we propose a novel controllable variational autoencoder framework, ControlVAE, that combines a controller, inspired by automatic control theory, with the basic VAE to improve the performance of resulting generative models. Specifically, we design a new non-linear PI controller, a variant of the proportional-integral-derivative (PID) control, to automatically tune the hyperparameter (weight) added in the VAE objective using the output KL-divergence as feedback during model training. The framework is evaluated using three applications; namely, language modeling, disentangled representation learning, and image generation. The results show that ControlVAE can achieve much better reconstruction quality than the competitive methods for the comparable disentanglement performance. For language modeling, it not only averts the KL-vanishing, but also improves the diversity of generated text. Finally, we also demonstrate that ControlVAE improves the reconstruction quality for image generation compared to the original VAE.

GitHub has become a popular social application platform, where a large number of users post their open source projects. In particular, an increasing number of researchers release repositories of source code related to their research papers in order to attract more people to follow their work. Motivated by this trend, we describe a novel item-item cross-platform recommender system, paper2repo, that recommends relevant repositories on GitHub that match a given paper in an academic search system such as Microsoft Academic. The key challenge is to identify the similarity between an input paper and its related repositories across the two platforms, without the benefit of human labeling. Towards that end, paper2repo integrates text encoding and constrained graph convolutional networks (GCN) to automatically learn and map the embeddings of papers and repositories into the same space, where proximity offers the basis for recommendation. To make our method more practical in real life systems, labels used for model training are computed automatically from features of user actions on GitHub. In machine learning, such automatic labeling is often called distant supervision. To the authors’ knowledge, this is the first distant-supervised cross-platform (paper to repository) matching system. We evaluate the performance of paper2repo on real-world data sets collected from GitHub and Microsoft Academic. Results demonstrate that it outperforms other state of the art recommendation methods.

We present GluonCV and GluonNLP, the deep learning toolkits for computer vision and natural language processing based on Apache MXNet (incubating). These toolkits provide state-of-the-art pre-trained models, training scripts, and training logs, to facilitate rapid prototyping and promote reproducible research. We also provide modular APIs with flexible building blocks to enable efficient customization. Leveraging the MXNet ecosystem, the deep learning models in GluonCV and GluonNLP can be deployed onto a variety of platforms with different programming languages. The Apache 2.0 license has been adopted by GluonCV and GluonNLP to allow for software distribution, modification, and usage.

Yi Tay, Shuohang Wang, Anh Tuan Luu, Jie Fu, Minh C. Phan, Xingdi Yuan, Jinfeng Rao, Siu Cheung Hui, Aston Zhang. Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. 2019.

Many state-of-the-art neural models for NLP are heavily parameterized and thus memory inefficient.This paper proposes a series of lightweight and memory efficient neural architectures for a potpourri of natural language processing (NLP) tasks.To this end, our models exploit computation using Quaternion algebra and hypercomplex spaces, enabling not only expressive inter-component interactions but also significantly (75%) reduced parameter size due to lesser degrees of freedom in the Hamilton product.We propose Quaternion variants of models, giving rise to new architectures such as the Quaternion attention Model and Quaternion Transformer.Extensive experiments on a battery of NLP tasks demonstrates the utility of proposed Quaternion-inspired models, enabling up to 75% reduction in parameter size without significant loss in performance.

Traditionally, many text-mining tasks treat individual word-tokens as the finest meaningful semantic granularity. However, in many languages and specialized corpora, words are composed by concatenating semantically meaningful subword structures. Word-level analysis cannot leverage the semantic information present in such subword structures. With regard to word embedding techniques, this leads to not only poor embeddings for infrequent words in long-tailed text corpora but also weak capabilities for handling out-of-vocabulary words. In this paper we propose MorphMine for unsupervised morpheme segmentation. MorphMine applies a parsimony criterion to hierarchically segment words into the fewest number of morphemes at each level of the hierarchy. This leads to longer shared morphemes at each level of segmentation. Experiments show that MorphMine segments words in a variety of languages into human-verified morphemes. Additionally, we experimentally demonstrate that utilizing MorphMine morphemes to enrich word embeddings consistently improves embedding quality on a variety of of embedding evaluations and a downstream language modeling task.

This paper proposes Quaternion Collaborative Filtering (QCF), a novel representation learning method for recommendation. Our proposed QCF relies on and exploits computation with Quaternion algebra, benefiting from the expressiveness and rich representation learning capability of Hamilton products. Quaternion representations, based on hypercomplex numbers, enable rich inter-latent dependencies between imaginary components. This encourages intricate relations to be captured when learning user-item interactions, serving as a strong inductive bias as compared with the real-space inner product. All in all, we conduct extensive experiments on six real-world datasets, demonstrating the effectiveness of Quaternion algebra in recommender systems. The results exhibit that QCF outperforms a wide spectrum of strong neural baselines on all datasets. Ablative experiments confirm the effectiveness of Hamilton-based composition over multi-embedding composition in real space.

Attentional models are distinctly characterized by their ability to learn relative importance, i.e., assigning a different weight to input values. This paper proposes a new quasi-attention that is compositional in nature, i.e., learning whether to \textit{add}, \textit{subtract} or \textit{nullify} a certain vector when learning representations. This is strongly contrasted with vanilla attention, which simply re-weights input tokens. Our proposed \textit{Compositional De-Attention} (CoDA) is fundamentally built upon the intuition of both similarity and dissimilarity (negative affinity) when computing affinity scores, benefiting from a greater extent of expressiveness. We evaluate CoDA on six NLP tasks, i.e. open domain question answering, retrieval/ranking, natural language inference, machine translation, sentiment analysis and text2code generation. We obtain promising experimental results, achieving state-of-the-art performance on several tasks/datasets.

Recent advances in deep-learning-based applications have attracted a growing attention from the IoT community. These highly capable learning models have shown significant improvements in expected accuracy of various sensory inference tasks. One important and yet overlooked direction remains to provide uncertainty estimates in deep learning outputs. Since robustness and reliability of sensory inference results are critical to IoT systems, uncertainty estimates are indispensable for IoT applications. To address this challenge, we develop ApDeepSense, an effective and efficient deep learning uncertainty estimation method for resource-constrained IoT devices. ApDeepSense leverages an implicit Bayesian approximation that links neural networks to deep Gaussian processes, allowing output uncertainty to be quantified. Our approach is shown to significantly reduce the execution time and energy consumption of uncertainty estimation thanks to a novel layer-wise approximation that replaces the traditional computationally intensive sampling-based uncertainty estimation methods. ApDeepSense is designed for neural net-works trained using dropout; one of the most widely used regularization methods in deep learning. No additional training is needed for uncertainty estimation purposes. We evaluate ApDeepSense using four IoT applications on Intel Edison devices. Results show that ApDeepSense can reduce around 88.9% of the execution time and 90.0% of the energy consumption, while producing more accurate uncertainty estimates compared with state-of-the-art methods.

Recent advances in deep learning motivate the use of deep neutral networks in sensing applications, but their excessive resource needs on constrained embedded devices remain an important impediment. A recently explored solution space lies in compressing (approximating or simplifying) deep neural networks in some manner before use on the device. We propose a new compression solution, called DeepIoT, that makes two key contributions in that space. First, unlike current solutions geared for compressing specific types of neural networks, DeepIoT presents a unified approach that compresses all commonly used deep learning structures for sensing applications, including fully-connected, convolutional, and recurrent neural networks, as well as their combinations. Second, unlike solutions that either sparsify weight matrices or assume linear structure within weight matrices, DeepIoT compresses neural network structures into smaller dense matrices by finding the minimum number of non-redundant hidden elements, such as filters and dimensions required by each layer, while keeping the performance of sensing applications the same. Importantly, it does so using an approach that obtains a global view of parameter redundancies, which is shown to produce superior compression. The compressed model generated by DeepIoT can directly use existing deep learning libraries that run on embedded and mobile systems without further modifications. We conduct experiments with five different sensing-related tasks on Intel Edison devices. DeepIoT outperforms all compared baseline algorithms with respect to execution time and energy consumption by a significant margin. It reduces the size of deep neural networks by 90% to 98.9%. It is thus able to shorten execution time by 71.4% to 94.5%, and decrease energy consumption by 72.2% to 95.7%. These improvements are achieved without loss of accuracy. The results underscore the potential of DeepIoT for advancing the exploitation of deep neural networks on resource-constrained embedded devices.

Mobile sensing and computing applications usually require time-series inputs from sensors, such as accelerometers, gyroscopes, and magnetometers. Some applications, such as tracking, can use sensed acceleration and rate of rotation to calculate displacement based on physical system models. Other applications, such as activity recognition, extract manually designed features from sensor inputs for classification. Such applications face two challenges. On one hand, on-device sensor measurements are noisy. For many mobile applications, it is hard to find a distribution that exactly describes the noise in practice. Unfortunately, calculating target quantities based on physical system and noise models is only as accurate as the noise assumptions. Similarly, in classification applications, although manually designed features have proven to be effective, it is not always straightforward to find the most robust features to accommodate diverse sensor noise patterns and heterogeneous user behaviors. To this end, we propose DeepSense, a deep learning framework that directly addresses the aforementioned noise and feature customization challenges in a unified manner. DeepSense integrates convolutional and recurrent neural networks to exploit local interactions among similar mobile sensors, merge local interactions of different sensory modalities into global interactions, and extract temporal relationships to model signal dynamics. DeepSense thus provides a general signal estimation and classification framework that accommodates a wide range of applications. We demonstrate the effectiveness of DeepSense using three representative and challenging tasks: car tracking with motion sensors, heterogeneous human activity recognition, and user identification with biometric motion analysis. DeepSense significantly outperforms the state-of-the-art methods for all three tasks. In addition, we show that DeepSense is feasible to implement on smartphones and embedded devices thanks to its moderate energy consumption and low latency.

In a collection of documents, such as news articles or tweets, various events take place over time. The event detection problem aims at discovering significant events that have not been mentioned before the detection time. When these events occur, we observe that topic distributions of documents will diverge notably. However, event detection from such divergence may be hampered by noises. In this paper, we propose TopicDiver, a novel method to address the event detection problem. TopicDiver models topic distributions of documents over time and filters out noises while capturing the useful divergence between such distributions. The direct exploitation of topic distribution over time sets our work apart from existing studies on event detection. We conduct comprehensive experiments under different settings on news and Twitter data. The experimental results demonstrate that TopicDiver outperforms the baseline models in the measures for accuracy across various settings.

We study the composite minimization problem where the objective function is the sum of two convex functions: one is the sum of a finite number of strongly convex and smooth functions, and the other is a general convex function that is non-differentiable. Specifically, we consider the case where the non-differentiable function is block separable and admits a simple proximal mapping for each block. This type of composite optimization is common in many data mining and machine learning problems, and can be solved by block coordinate descent algorithms. We propose an accelerated stochastic block coordinate descent (ASBCD) algorithm, which incorporates the incrementally averaged partial derivative into the stochastic partial derivative and exploits optimal sampling. We prove that ASBCD attains a linear rate of convergence. In contrast to uniform sampling, we reveal that the optimal non-uniform sampling can be employed to achieve a lower iteration complexity. Experimental results on different large-scale real data sets support our theory.

We study the new mobile query auto-completion (QAC) problem to exploit mobile devices' exclusive signals, such as those related to mobile applications (apps). We propose AppAware, a novel QAC model using installed app and recently opened app signals to suggest queries for matching input prefixes on mobile devices. To overcome the challenge of noisy and voluminous signals, AppAware optimizes composite objectives with a lighter processing cost at a linear rate of convergence. We conduct experiments on a large commercial data set of mobile queries and apps. Installed app and recently opened app signals consistently and significantly boost the accuracy of various baseline QAC models on mobile devices.

While many studies have been conducted on query understanding, there is limited understanding on why users start searches and how to predict search intent. In this paper, we propose to study this important but less explored problem. Our key intuition is that searches are triggered by different pre-search contexts, but the triggering relations are often hidden. For example, a user may search "bitcoin" because of a news article or an email the user just read, but the system does not know which of the pre-search contexts (the news article or the email) is the triggering source. Following this intuition, we conduct an in-depth analysis of pre-search context on a large-scale user log, which not only verifies the hidden triggering relations in the real world but also identifies a set of important characteristics of pre-search context and their triggered queries. Since the hidden triggering relations make it challenging to directly use pre-search context for intent prediction, we develop a mixture generative model to learn without any supervision how queries are triggered by different types of pre-search context. Further, we discuss how to apply our model to improve query prediction and query auto-completion. Our experiments on a large-scale of real-world data show that our model could accurately predict user search intent with pre-search context and improve upon the state-of-the-art methods significantly.

Query auto-completion (QAC) facilitates user query composition by suggesting queries given query prefix inputs. In 2014, global users of Yahoo! Search saved more than 50% keystrokes when submitting English queries by selecting suggestions of QAC. Users' preference of queries can be inferred during user-QAC interactions, such as dwelling on suggestion lists for a long time without selecting query suggestions ranked at the top. However, the wealth of such implicit negative feedback has not been exploited for designing QAC models. Most existing QAC models rank suggested queries for given prefixes based on certain relevance scores.

The study of online social networks has attracted increasing interest. However, concerns are raised for the privacy risks of user data since they have been frequently shared among researchers, advertisers, and application developers. To solve this problem, a number of anonymization algorithms have been recently developed for protecting the privacy of social graphs. In this article, we proposed a graph node similarity measurement in consideration with both graph structure and descriptive information, and a deanonymization algorithm based on the measurement. Using the proposed algorithm, we evaluated the privacy risks of several typical anonymization algorithms on social graphs with thousands of nodes from Microsoft Academic Search, LiveJournal, and the Enron email dataset, and a social graph with millions of nodes from Tencent Weibo. Our results showed that the proposed algorithm was efficient and effective to deanonymize social graphs without any initial seed mappings. Based on the experiments, we also pointed out suggestions on how to better maintain the data utility while preserving privacy.

Anonymized user datasets are often released for research or indus-try applications. As an example, t.qq.com released its anonymized users ’ profile, social interaction, and recommendation log data in KDD Cup 2012 to call for recommendation algorithms. Since the entities (users and so on) and edges (links among entities) are of multiple types, the released social network is a heterogeneous in-formation network. Prior work has shown how privacy can be com-promised in homogeneous information networks by the use of spe-cific types of graph patterns. We show how the extra information derived from heterogeneity can be used to relax these assumptions. To characterize and demonstrate this added threat, we formally de-fine privacy risk in an anonymized heterogeneous information net-work to identify the vulnerability in the possible way such data are released, and further present a new de-anonymization attack that exploits the vulnerability. Our attack successfully de-anonymized most individuals involved in the data—for an anonymized 1,000-user t.qq.com network of density 0.01, the attack precision is over 90 % with a 2.3-million-user auxiliary network.