Constructing drivable and photorealistic 3D head avatars has become a central task in$A R / X R$, enabling immersive and expressive user experiences. With the emergence of high-fidelity, efficient representations such as$3 D$Gaussians, recent works have pushed toward ultra-detailed head avatars. Existing approaches typically fall into two categories: rule-based analytic rigging or neural networkbased deformation fields. While effective in constrained settings, both approaches often fail to generalize to unseen expressions and poses-particularly in extreme reenactment scenarios. Other methods constrain Gaussians to the global texel space of 3DMMs to reduce rendering complexity. However, these texel-based avatars tend to underutilize the underlying mesh structure. They apply minimal analytic deformation and rely heavily on neural regressors and heuristic regularization in UV space, which weakens geometric consistency and limits extrapolation to complex, out-of-distribution deformations. To address these limitations, we introduce TexAvatars, a hybrid avatar representation that combines the explicit geometric grounding of analytic rigging with the spatial continuity of texel space. Our approach predicts local geometric attributes in UV space via CNNs, but drives 3D deformation through mesh-aware Jacobians, enabling smooth and semantically meaningful transitions across triangle boundaries. This hybrid design separates semantic modeling from geometric control, resulting in improved generalization, interpretability, and stability. Furthermore, TexAvatars captures fine-grained expression effects-including muscle-induced wrinkles, glabellar lines, and realistic mouth cavity geom-etry-with high fidelity. Our method achieves state-of-theart performance under extreme pose and expression vari-ations, demonstrating strong generalization in challenging head reenactment settings.

The increasing demand for AR/VR applications has highlighted the need for high-quality content, such as 360° live wallpapers. However, generating high-quality 360° panoramic contents remains a challenging task due to the severe distortions introduced by equirectangular projection (ERP). Existing approaches either fine-tune pretrained diffusion models on limited ERP datasets or adopt tuning-free methods that still rely on ERP latent representations, often resulting in distracting distortions near the poles. In this paper, we introduce SphereDiff, a novel approach for synthesizing 360° static and live wallpaper with state-of-the-art diffusion models without additional tuning. We define a spherical latent representation that ensures consistent quality across all perspectives, including near the poles. Then, we extend MultiDiffusion to spherical latent representation and propose a dynamic spherical latent sampling method to enable direct use of pretrained diffusion models. Moreover, we introduce distortion-aware weighted averaging to further improve the generation quality. Our method outperforms existing approaches in generating 360° static and live wallpaper, making it a robust solution for immersive AR/VR applications.

Large Language Models (LLMs) have recently emerged as powerful reasoning engines in recommender systems, generating natural-language explanations that foster user engagement. However, their recommendation performance remains limited, as they lack exposure to collaborative user-item interaction patterns. In contrast, collaborative filtering (CF) models achieve strong performance by learning from these behavioral patterns at scale. To unify the strengths of both paradigms, we propose TWiCE-Rec (Think Wise, Collaborate Effectively), a rationale-aware LLM-based recommender that incorporates collaborative user-item interactions. In the first stage, we construct a rationale dataset by applying in-context learning with self-annotated curation. A state-of-the-art LLM is guided to generate persuasive rationales that explain the causal relationship between the user’s interaction sequence and the ground-truth next item, resulting in a curated post-hoc training dataset. In the second stage, we perform multi-task instruction-tuned adaptation—based on the rationale-augmented training dataset—comprising item description generation and both non-reasoning and reasoning-based sequential recommendation, to equip the LLM with the ability to generate rationales that reflect how user preferences align with item characteristics. Finally, we aim to enhance the LLM’s recommendation performance by incorporating user-item interaction patterns derived from the CF-Rec model. To achieve this, we propose a confidence-weighted reinforcement learning strategy that adjusts rewards in proportion to both the LLM’s prediction alignment with the ground-truth and the confidence from the pretrained CF-Rec model. Our method outperforms both CF- and LLM-Rec models on Amazon datasets in terms of recommendation performance and rationale quality. In an online A/B test, it achieved about 8% higher click-through rate than existing models, demonstrating practical value.

Hairstyle transfer is a challenging task in the image editing field that modifies the hairstyle of a given face image while preserving its other appearance and background features. The existing hairstyle transfer approaches heavily rely on StyleGAN, which is pre-trained on cropped and aligned face images. Hence, they struggle to generalize under challenging conditions such as extreme variations of head poses or focal lengths. To address this issue, we propose a one-stage hairstyle transfer diffusion model, HairFusion, that applies to real-world scenarios. Specifically, we carefully design a hair-agnostic representation as the input of the model, where the original hair information is thoroughly eliminated. Next, we introduce a hair align cross-attention (Align-CA) to accurately align the reference hairstyle with the face image while considering the difference in their head poses. To enhance the preservation of the face image’s original features, we leverage adaptive hair blending during the inference, where the output’s hair regions are estimated by the cross-attention map in Align-CA and blended with non-hair areas of the face image. Our experimental results show that our method achieves state-of-the-art performance compared to the existing methods in preserving the integrity of both the transferred hairstyle and the surrounding features.

Point-based interactive colorization techniques allow users to effortlessly colorize grayscale images using user-provided color hints. However, point-based methods often face challenges when different colors are given to semantically similar areas, leading to color intermingling and unsatisfactory results—an issue we refer to as color collapse. The fundamental cause of color collapse is the inadequacy of points for defining the boundaries for each color. To mitigate color collapse, we introduce a lasso tool that can control the scope of each color hint. Additionally, we design a framework that leverages the user-provided lassos to localize the attention masks. The experimental results show that using a single lasso is as effective as applying 4.18 individual color hints and can achieve the desired outcomes in 30% less time than using points alone.

Instruction-following large language models (LLMs), such as ChatGPT, have become widely popular among everyday users.However, these models inadvertently disclose private, sensitive information to their users, underscoring the need for machine unlearning techniques to remove selective information from the models.While prior work has focused on forgetting small, random subsets of training data at the instance-level, we argue that real-world scenarios often require the removal of an entire user data, which may require a more careful maneuver.In this study, we explore entitylevel unlearning, which aims to erase all knowledge related to a target entity while preserving the remaining model capabilities.To address this, we introduce OPT-OUT, an optimal transport-based unlearning method that utilizes the Wasserstein distance from the model's initial parameters to achieve more effective and fine-grained unlearning.We also present the first Entity-Level Unlearning Dataset (ELUDe) designed to evaluate entity-level unlearning.Our empirical results demonstrate that OPT-OUT surpasses existing methods, establishing a new standard for secure and adaptable LLMs that can accommodate user data removal requests without the need for full retraining. 1

To create culturally inclusive vision-language models (VLMs), developing a benchmark that tests their ability to address culturally relevant questions is essential. Existing approaches typically rely on human annotators, making the process labor-intensive and creating a cognitive burden in generating diverse questions. To address this, we propose a semi-automated framework for constructing cultural VLM benchmarks, specifically targeting multiple-choice QA. This framework combines human-VLM collaboration, where VLMs generate questions based on guidelines, a small set of annotated examples, and relevant knowledge, followed by a verification process by native speakers. We demonstrate the effectiveness of this framework through the creation of K-Viscuit, a dataset focused on Korean culture. Our experiments on this dataset reveal that open-source models lag behind proprietary ones in understanding Korean culture, highlighting key areas for improvement. We also present a series of further analyses, including human evaluation, augmenting VLMs with external knowledge, and the evaluation beyond multiple-choice QA. Our dataset is available at https://huggingface.co/datasets/ddehun/k-viscuit.

Diffusion models have emerged as a powerful tool for generating high-quality images, videos, and 3D content. While sampling guidance techniques like CFG improve quality, they reduce diversity and motion. Autoguidance mitigates these issues but demands extra weak model training, limiting its practicality for large-scale models. In this work, we introduce Spatiotemporal Skip Guidance (STG), a simple training-free sampling guidance method for enhancing transformer-based video diffusion models. STG employs an implicit weak model via self-perturbation, avoiding the need for external models or additional training. By selectively skipping spatiotemporal layers, STG produces an aligned, degraded version of the original model to boost sample quality without compromising diversity or dynamic degree. Our contributions include: (1) introducing STG as an efficient, high-performing guidance technique for video diffusion models, (2) eliminating the need for auxiliary models by simulating a weak model through layer skipping, and (3) ensuring quality-enhanced guidance without compromising sample diversity or dynamics unlike CFG. For additional results, visit https://junhahyung.github.io/STGuidance/.

While large-scale text-to-image diffusion models enable the generation of high-quality, diverse images from text prompts, these prompts struggle to capture intricate details, such as textures, preventing the user intent from being reflected. This limitation has led to efforts to generate images conditioned on user-provided images, referred to as image prompts. Recent work modifies the self-attention mechanism to impose image conditions in generated images by replacing or concatenating the keys and values from the image prompt. This enables the self-attention layer to work like a cross-attention layer, generally used to incorporate text prompts. In this paper, we identify two common issues in existing methods of modifying self-attention that hinder diffusion models from reflecting the image prompt. By addressing these issues, we propose a novel method that generates images that properly reflect the details of image prompts. First, existing approaches often neglect the importance of image prompts in classifier-free guidance, which directs the model towards the intended conditions and away from those undesirable. Specifically, current methods use image prompts as both desired and undesired conditions, causing conflicting signals. To resolve this, we propose conflict-free guidance by using image prompts only as desired conditions, ensuring that the generated image faithfully reflects the image prompt. In addition, we observe that the two most common self-attention modifications involve a trade-off between the realism of the generated image and alignment with the image prompt, achieved by selectively using keys and values from both images. Specifically, selecting more keys and values from the image prompt improves alignment, while selecting more from the generated image enhances realism. To balance both, we propose an alternative self-attention modification method, Stratified Attention, which jointly uses keys and values from both images rather than selecting between them. Through extensive experiments across three distinct image generation tasks, we demonstrate that the proposed method outperforms existing image-prompting models in faithfully reflecting the image prompt.

With growing demand in media and social networks for personalized images, the need for advanced head-swapping techniques—integrating an entire head from the head image with the body from the body image—has increased. However, traditional head-swapping methods heavily rely on face-centered cropped data with primarily frontal-facing views, which limits their effectiveness in real-world applications. Additionally, their masking methods, designed to indicate regions requiring editing, are optimized for these types of dataset but struggle to achieve seamless blending in complex situations, such as when the original data includes features like long hair extending beyond the masked area. To overcome these limitations and enhance adaptability in diverse and complex scenarios, we propose a novel head swapping method, HID, that is robust to images including the full head and the upper body, and handles from frontal to side views, while automatically generating context-aware masks. For automatic mask generation, we introduce the IOMask, which enables seamless blending of the head and body, effectively addressing integration challenges. We further introduce the hair injection module to capture hair details with greater precision. Our experiments demonstrate that the proposed approach achieves state-of-the-art performance in head swapping, providing visually consistent and realistic results across a wide range of challenging conditions.

Masked diffusion models (MDMs) offer a promising non-autoregressive alternative for large language modeling.Standard decoding methods for MDMs, such as confidence-based sampling, select tokens independently based on individual token confidences at each diffusion step.However, we observe that this independent token selection often results in generation orders resembling sequential autoregressive processes, limiting the advantages of non-autoregressive modeling.To mitigate this pheonomenon, we propose Reward-Weighted Sampling (RWS), a novel decoding strategy that leverages an external reward model to provide a principled global signal during the iterative diffusion process.Specifically, at each diffusion step, RWS evaluates the quality of the entire intermediate sequence and scales token logits accordingly, guiding token selection by integrating global sequence-level coherence.This method selectively increases the confidence of tokens that initially have lower scores, thereby promoting a more non-autoregressive generation order.Furthermore, we provide theoretical justification showing that rewardweighted logit scaling induces beneficial rank reversals in token selection and consistently improves expected reward.Experiments demonstrate that RWS significantly promotes nonautoregressive generation orders, leading to improvements across multiple evaluation metrics.These results highlight the effectiveness of integrating global signals in enhancing both the non-autoregressive properties and overall performance of MDMs.

Recent virtual try-on approaches have advanced by finetuning pre-trained text-to-image diffusion models to leverage their powerful generative ability. However, the use of text prompts in virtual try-on remains underexplored. This paper tackles a text-editable virtual try-on task that modifies the clothing based on the provided clothing image while editing the wearing style (e.g., tucking style, fit) according to the text descriptions. In the text-editable virtual try-on, three key aspects exist: (i) designing rich text descriptions for paired person-clothing data to train the model, (ii) addressing the conflicts where textual information of the existing person's clothing interferes the generation of the new clothing, and (iii) adaptively adjust the inpainting mask aligned with the text descriptions, ensuring proper editing areas while preserving the original person's appearance irrelevant to the new clothing. To address these aspects, we propose PromptDresser, a text-editable virtual try-on model that leverages large multimodal model (LMM) assistance to enable high-quality and versatile manipulation based on generative text prompts. Our approach utilizes LMMs via in-context learning to generate detailed text descriptions for person and clothing images independently, including pose details and editing attributes using minimal human cost. Moreover, to ensure the editing areas, we adjust the inpainting mask depending on the text prompts adaptively. Our approach enhances text editability while effectively conveying clothing details that are difficult to capture through images alone, leading to improved image quality. Experiments show that PromptDresser significantly outperforms baselines, demonstrating superior text-driven control and versatile clothing manipulation. Our code is available at https://github.com/rlawjdghek/PromptDresser.

Recent advancements in diffusion-based text-to-image (T2I) models have enabled the generation of high-quality and photorealistic images from text. However, they often exhibit societal biases related to gender, race, and socioeconomic status, thereby potentially reinforcing harmful stereotypes and shaping public perception in unintended ways. While existing bias mitigation methods demonstrate effectiveness, they often encounter attribute entanglement, where adjustments to attributes relevant to the bias (i.e., target attributes) unintentionally alter attributes unassociated with the bias (i.e., non-target attributes), causing undesirable distribution shifts. To address this challenge, we introduce Entanglement-Free Attention (EFA), a method that accurately incorporates target attributes (e.g., White, Black, and Asian) while preserving non-target attributes (e.g., background) during bias mitigation. At inference time, EFA randomly samples a target attribute with equal probability and adjusts the cross-attention in selected layers to incorporate the sampled attribute, achieving a fair distribution of target attributes. Extensive experiments demonstrate that EFA outperforms existing methods in mitigating bias while preserving non-target attributes, thereby maintaining the original model's output distribution and generative capacity.

Recent advances in CV and NLP have been largely driven by scaling up the number of network parameters, despite traditional theories suggesting that larger networks are prone to overfitting. These large networks avoid overfitting by integrating components that induce a simplicity bias, guiding models toward simple and generalizable solutions. However, in deep RL, designing and scaling up networks have been less explored. Motivated by this opportunity, we present SimBa, an architecture designed to scale up parameters in deep RL by injecting a simplicity bias. SimBa consists of three components: (i) an observation normalization layer that standardizes inputs with running statistics, (ii) a residual feedforward block to provide a linear pathway from the input to output, and (iii) a layer normalization to control feature magnitudes. By scaling up parameters with SimBa, the sample efficiency of various deep RL algorithms—including off-policy, on-policy, and unsupervised methods—is consistently improved. Moreover, solely by integrating SimBa architecture into SAC, it matches or surpasses state-of-the-art deep RL methods with high computational efficiency across DMC, MyoSuite, and HumanoidBench. These results demonstrate SimBa's broad applicability and effectiveness across diverse RL algorithms and environments.

Recent advancements in head avatar rendering using Gaussian primitives have achieved significantly high-fidelity results. Although precise head geometry is crucial for applications like mesh reconstruction and relighting, current methods struggle to capture intricate geometric details and render unseen poses due to their reliance on similarity transformations, which cannot handle stretch and shear transforms essential for detailed deformations of geometry. To address this, we propose SurFhead, a novel method that reconstructs riggable head geometry from RGB videos using 2D Gaussian surfels, which offer well-defined geometric properties, such as precise depth from fixed ray intersections and normals derived from their surface orientation, making them advantageous over 3D counterparts. SurFhead ensures high-fidelity rendering of both normals and images, even in extreme poses, by leveraging classical mesh-based deformation transfer and affine transformation interpolation. SurFhead introduces precise geometric deformation and blends surfels through polar decomposition of transformations, including those affecting normals. Our key contribution lies in bridging classical graphics techniques, such as mesh-based deformation, with modern Gaussian primitives, achieving state-of-the-art geometry reconstruction and rendering quality. Unlike previous avatar rendering approaches, SurFhead enables efficient reconstruction driven by Gaussian primitives while preserving high-fidelity geometry.

Adapting foundation models for specific purposes has become a standard approach to build machine learning systems for downstream applications. Yet, it is an open question which mechanisms take place during adaptation. Here we develop a new Sparse Autoencoder (SAE) for the CLIP vision transformer, named PatchSAE, to extract interpretable concepts at granular levels (e.g., shape, color, or semantics of an object) and their patch-wise spatial attributions. We explore how these concepts influence the model output in downstream image classification tasks and investigate how recent state-of-the-art prompt-based adaptation techniques change the association of model inputs to these concepts. While activations of concepts slightly change between adapted and non-adapted models, we find that the majority of gains on common adaptation tasks can be explained with the existing concepts already present in the non-adapted foundation model. This work provides a concrete framework to train and use SAEs for Vision Transformers and provides insights into explaining adaptation mechanisms.

Scaling up the model size and computation has brought consistent performance improvements in supervised learning. However, this lesson often fails to apply to reinforcement learning (RL) because training the model on non-stationary data easily leads to overfitting and unstable optimization. In response, we introduce SimbaV2, a novel RL architecture designed to stabilize optimization by (i) constraining the growth of weight and feature norm by hyperspherical normalization; and (ii) using a distributional value estimation with reward scaling to maintain stable gradients under varying reward magnitudes. Using the soft actor-critic as a base algorithm, SimbaV2 scales up effectively with larger models and greater compute, achieving state-of-the-art performance on 57 continuous control tasks across 4 domains.

Soyoung Yang, Hojun Cho, Jiyoung Lee, Sohee Yoon, Edward Choi, Jaegul Choo, Won Ik Cho. 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.

Dataset bias, where data points are skewed to certain concepts, is ubiquitous in machine learning datasets. Yet, systematically identifying these biases is challenging without costly, fine-grained attribute annotations. We present ConceptScope, a scalable and automated framework for analyzing visual datasets by discovering and quantifying human-interpretable concepts using Sparse Autoencoders trained on representations from vision foundation models. ConceptScope categorizes concepts into target, context, and bias types based on their semantic relevance and statistical correlation to class labels, enabling class-level dataset characterization, bias identification, and robustness evaluation through concept-based subgrouping. We validate that ConceptScope captures a wide range of visual concepts, including objects, textures, backgrounds, facial attributes, emotions, and actions, through comparisons with annotated datasets. Furthermore, we show that concept activations produce spatial attributions that align with semantically meaningful image regions. ConceptScope reliably detects known biases (e.g., background bias in Waterbirds) and uncovers previously unannotated ones (e.g, co-occurring objects in ImageNet), offering a practical tool for dataset auditing and model diagnostics.

Recent advances in text-to-video diffusion models have enabled high-quality video synthesis, but controllable generation remains challenging—particularly under limited data and compute. Existing fine-tuning methods for conditional generation often rely on external encoders or architectural modifications, which demand large datasets and are typically restricted to spatially aligned conditioning, limiting flexibility and scalability. In this work, we introduce Temporal In-Context Fine-Tuning (TIC-FT), an efficient and versatile approach with temporal reasoning for adapting pretrained video diffusion models to diverse conditional generation tasks. Our key idea is to concatenate condition and target frames along the temporal axis and insert intermediate buffer frames with progressively increasing noise levels. These buffer frames enable smooth transitions, aligning the fine-tuning process with the pretrained model’s temporal dynamics. TIC-FT is architecture-agnostic and achieves strong performance with as few as 10–30 training samples. We validate our method across a range of tasks—including image-to-video and video-to-video generation—using large-scale base models such as CogVideoX-5B and Wan-14B. Extensive experiments show that TIC-FT outperforms existing baselines in both condition fidelity and visual quality, while remaining highly efficient in both training and inference.

Recent efforts to apply Large Language Models (LLMs) to time-series anomaly detection (TSAD) have yielded limited success, often performing worse than even simple methods. While prior work has focused solely on downstream performance evaluation, the fundamental question—why do LLMs struggle with TSAD?—has remained largely unexplored. In this paper, we present an in-depth analysis that identifies two core challenges in understanding complex temporal dynamics and accurately localizing anomalies. To address these challenges, we propose a simple yet effective method that combines statistical decomposition with index-aware prompting. Our method outperforms 21 existing prompting strategies on the AnomLLM benchmark, achieving up to a 66.6\% improvement in F1 score. We further compare LLMs with 16 non-LLM baselines on the TSB-AD benchmark, highlighting scenarios where LLMs offer unique advantages via contextual reasoning. Our findings provide empirical insights into how and when LLMs can be effective for TSAD. The code is publicly available at: https://github.com/junwoopark92/LLM-TSAD

Recent research has unveiled the development of customizing large-scale text-to-image models. These models bind a unique subject desired by a user to a specific token, using the token to generate the subject in various contexts. However, models from previous studies also bind elements unrelated to the subject's identity, such as common backgrounds or poses in the reference images. This often leads to conflicts between the token and the context of text prompts during inference, causing the model to fail to generate both the subject and the prompted context. In this work, we approach this issue from a data scarcity perspective and propose to augment the number of reference images through a novel self-distillation framework. Our framework selects high-quality samples from images generated by a teacher model and uses them in student training. Our framework can be applied to any models that suffer from the conflicts, and we demonstrate that our framework most effectively resolves the issue through comprehensive evaluations.

The emergence of diffusion models has enabled the generation of diverse high-quality images solely from text, prompting subsequent efforts to enhance the controllability of these models. Despite the improvement in controllability, pose control remains limited to specific objects (e.g., humans) or poses (e.g., frontal view) due to the fact that pose is generally controlled via camera parameters (e.g., rotation angle) or keypoints (e.g., eyes, nose). Specifically, camera parameters-conditional pose control models generate unrealistic images depending on the object, owing to the small size of 3D datasets for training. Also, keypoint-based approaches encounter challenges in acquiring reliable key-points for various objects (e.g., church) or poses (e.g., back view). To address these limitations, we propose depth-based pose control, as depth maps are easily obtainable from a single depth estimation model regardless of objects and poses, unlike camera parameters and keypoints. However, depth-based pose control confronts issues of shape dependency, as depth maps influence not only the pose but also the shape of the generated images. To tackle this issue, we propose Skip-and-Play (SnP), designed via analysis of the impact of three components of depth-conditional Control-Net on the pose and the shape of the generated images. To be specific, based on the analysis, we selectively skip parts of the components to mitigate shape dependency on the depth map while preserving the pose. Through various experiments, we demonstrate the superiority of SnP over baselines and showcase the ability of SnP to generate images of diverse objects and poses. Remarkably, SnP exhibits the ability to generate images even when the objects in the condition (e.g., a horse) and the prompt (e.g., a hedgehog) differ from each other.

Accurately annotating multiple 3D objects in LiDAR scenes is laborious and challenging. While a few previous studies have attempted to leverage semi-automatic methods for cost-effective bounding box annotation, such methods have limitations in efficiently handling numerous multi-class objects. To effectively accelerate 3D annotation pipelines, we propose iDet3D, an efficient interactive 3D object detector. Supporting a user-friendly 2D interface, which can ease the cognitive burden of exploring 3D space to provide click interactions, iDet3D enables users to annotate the entire objects in each scene with minimal interactions. Taking the sparse nature of 3D point clouds into account, we design a negative click simulation (NCS) to improve accuracy by reducing false-positive predictions. In addition, iDet3D incorporates two click propagation techniques to take full advantage of user interactions: (1) dense click guidance (DCG) for keeping user-provided information throughout the network and (2) spatial click propagation (SCP) for detecting other instances of the same class based on the user-specified objects. Through our extensive experiments, we present that our method can construct precise annotations in a few clicks, which shows the practicality as an efficient annotation tool for 3D object detection.

Large-scale text-to-image models including Stable Diffusion are capable of generating high-fidelity photorealistic portrait images. There is an active research area dedicated to personalizing these models, aiming to synthesize specific subjects or styles using provided sets of reference images. However, despite the plausible results from these personalization methods, they tend to produce images that often fall short of realism and are not yet on a commercially viable level. This is particularly noticeable in portrait image generation, where any unnatural artifact in human faces is easily discernible due to our inherent human bias. To address this, we introduce MagiCapture, a personalization method for integrating subject and style concepts to generate high-resolution portrait images using just a few subject and style references. For instance, given a handful of random selfies, our fine-tuned model can generate high-quality portrait images in specific styles, such as passport or profile photos. The main challenge with this task is the absence of ground truth for the composed concepts, leading to a reduction in the quality of the final output and an identity shift of the source subject. To address these issues, we present a novel Attention Refocusing loss coupled with auxiliary priors, both of which facilitate robust learning within this weakly supervised learning setting. Our pipeline also includes additional post-processing steps to ensure the creation of highly realistic outputs. MagiCapture outperforms other baselines in both quantitative and qualitative evaluations and can also be generalized to other non-human objects.

Time-series anomaly detection deals with the problem of detecting anomalous timesteps by learning normality from the sequence of observations. However, the concept of normality evolves over time, leading to a "new normal problem", where the distribution of normality can be changed due to the distribution shifts between training and test data. This paper highlights the prevalence of the new normal problem in unsupervised time-series anomaly detection studies. To tackle this issue, we propose a simple yet effective test-time adaptation strategy based on trend estimation and a self-supervised approach to learning new normalities during inference. Extensive experiments on real-world benchmarks demonstrate that incorporating the proposed strategy into the anomaly detector consistently improves the model's performances compared to the existing baselines, leading to robustness to the distribution shifts.

Given a clothing image and a person image, an image-based virtual try-on aims to generate a customized image that appears natural and accurately reflects the character-istics of the clothing image. In this work, we aim to expand the applicability of the pre-trained diffusion model so that it can be utilized independently for the virtual try-on task. The main challenge is to preserve the clothing details while effectively utilizing the robust generative capability of the pre-trained model. In order to tackle these issues, we propose StableVITON, learning the semantic correspon-dence between the clothing and the human body within the latent space of the pre-trained diffusion model in an end-to-end manner. Our proposed zero cross-attention blocks not only preserve the clothing details by learning the semantic correspondence but also generate high-fidelity images by utilizing the inherent knowledge of the pre-trained model in the warping process. Through our proposed novel attention total variation loss and applying augmentation, we achieve the sharp attention map, resulting in a more precise representation of clothing details. Stable VITON out-performs the baselines in qualitative and quantitative evaluation, showing promising quality in arbitrary person images. Our code is available at https://github.com/rlawjdghek/StableVITON.

In the image classification task, deep neural networks frequently rely on bias attributes that are spuriously cor-related with a target class in the presence of dataset bias, resulting in degraded performance when applied to data without bias attributes. The task of debiasing aims to compel classifiers to learn intrinsic attributes that inher-ently define a target class rather than focusing on bias at-tributes. While recent approaches mainly focus on empha-sizing the learning of data samples without bias attributes (i.e., bias-conflicting samples) compared to samples with bias attributes (i.e., bias-aligned samples), they fall short of directly guiding models where to focus for learning in-trinsic features. To address this limitation, this paper pro-poses a method that provides the model with explicit spa-tial guidance that indicates the region of intrinsic features. We first identify the intrinsic features by investigating the class-discerning common features between a bias-aligned (BA) sample and a bias-conflicting (BC) sample (i.e., bias-contrastive pair). Next, we enhance the intrinsic features in the BA sample that are relatively under-exploited for pre-diction compared to the BC sample. To construct the bias-contrastive pair without using bias information, we intro-duce a bias-negative score that distinguishes BC samples from BA samples employing a biased model. The experi-ments demonstrate that our method achieves state-of-the-art performance on synthetic and real-world datasets with various levels of bias severity.

Despite the remarkable advancements in head reenactment, the existing methods face challenges in cross-domain head reenactment, which aims to transfer human motions to domains outside the human, including cartoon characters. It is still difficult to extract motion from out-of-domain images due to the distinct appearances, such as large eyes. Recently, previous work introduced a large-scale anime dataset called AnimeCeleb and a cross-domain head reenactment model, including an optimization-based mapping function to translate the human domain’s expressions to the anime domain. However, we found that the mapping function, which relies on a subset of expressions, imposes limitations on the mapping of various expressions. To solve this challenge, we introduce a novel expression domain translation network that transforms human expressions into anime expressions. Specifically, to maintain the geometric consistency of expressions between the input and output of the expression domain translation network, we employ a 3D geometric-aware loss function that reduces the distances between the vertices in the 3D mesh of the human and anime. By doing so, it forces high-fidelity and one-to-one mapping with respect to two cross-expression domains. Our method outperforms existing methods in both qualitative and quantitative analysis, marking a significant advancement in the field of cross-domain head reenactment.

Long-term forecasting presents unique challenges due to the time and memory complexity of handling long sequences. Existing methods, which rely on sliding windows to process long sequences, struggle to effectively capture long-term variations that are partially caught within the short window (i.e., outer-window variations). In this paper, we introduce a novel approach that overcomes this limitation by employing contrastive learning and enhanced decomposition architecture, specifically designed to focus on long-term variations. To this end, our contrastive loss incorporates global autocorrelation held in the whole time series, which facilitates the construction of positive and negative pairs in a self-supervised manner. When combined with our decomposition networks, our constrative learning significantly improves long-term forecasting performance. Extensive experiments demonstrate that our approach outperforms 14 baseline models on well-established nine long-term benchmarks, especially in challenging scenarios that require a significantly long output for forecasting. This paper not only presents a novel direction for long-term forecasting but also offers a more reliable method for effectively integrating long-term variations into time-series representation learning.

Vision Transformers (ViT), when paired with large-scale pretraining, have shown remarkable performance across various computer vision tasks, primarily due to their weak inductive bias. However, while such weak inductive bias aids in pretraining scalability, this may hinder the effective adaptation of ViTs for visuo-motor control tasks as a result of the absence of control-centric inductive biases. Such absent inductive biases include spatial locality and translation equivariance bias which convolutions naturally offer. To this end, we introduce Convolution Injector (CoIn), an add-on module that injects convolutions which are rich in locality and equivariance biases into a pretrained ViT for effective adaptation in visuo-motor control. We evaluate CoIn with three distinct types of pretrained ViTs (CLIP, MVP, VC-1) across 12 varied control tasks within three separate domains (Adroit, MetaWorld, DMC), and demonstrate that CoIn consistently enhances control task performance across all experimented environments and models, validating the effectiveness of providing pretrained ViTs with control-centric biases.

Recently, various pre-training methods have been introduced in vision-based Reinforcement Learning (RL). However, their generalization ability remains unclear due to evaluations being limited to in-distribution environments and non-unified experimental setups. To address this, we introduce the Atari Pre-training Benchmark (Atari-PB), which pre-trains a ResNet-50 model on 10 million transitions from 50 Atari games and evaluates it across diverse environment distributions. Our experiments show that pre-training objectives focused on learning task-agnostic features (e.g., identifying objects and understanding temporal dynamics) enhance generalization across different environments. In contrast, objectives focused on learning task-specific knowledge (e.g., identifying agents and fitting reward functions) improve performance in environments similar to the pre-training dataset but not in varied ones. We publicize our codes, datasets, and model checkpoints at https://github.com/dojeon-ai/Atari-PB.

This study investigates the loss of generalization ability in neural networks, revisiting warm-starting experiments from Ash & Adams. Our empirical analysis reveals that common methods designed to enhance plasticity by maintaining trainability provide limited benefits to generalization. While reinitializing the network can be effective, it also risks losing valuable prior knowledge. To this end, we introduce the Hare & Tortoise, inspired by the brain's complementary learning system. Hare & Tortoise consists of two components: the Hare network, which rapidly adapts to new information analogously to the hippocampus, and the Tortoise network, which gradually integrates knowledge akin to the neocortex. By periodically reinitializing the Hare network to the Tortoise's weights, our method preserves plasticity while retaining general knowledge. Hare & Tortoise can effectively maintain the network's ability to generalize, which improves advanced reinforcement learning algorithms on the Atari-100k benchmark. The code is available at https://github.com/dojeon-ai/hare-tortoise.

Despite the growing prevalence of black-box pre-trained models (PTMs) such as prediction API services, there remains a significant challenge in directly applying general models to real-world scenarios due to the data distribution gap. Considering a data deficiency and constrained computational resource scenario, this paper proposes a novel parameter-efficient transfer learning framework for vision recognition models in the black-box setting. Our framework incorporates two novel training techniques. First, we align the input space (i.e., image) of PTMs to the target data distribution by generating visual prompts of spatial and frequency domain. Along with the novel spatial-frequency hybrid visual prompter, we design a novel training technique based on probabilistic clusters, which can enhance class separation in the output space (i.e., prediction probabilities). In experiments, our model demonstrates superior performance in a few-shot transfer learning setting across extensive visual recognition datasets, surpassing state-of-the-art baselines. Additionally, we show that the proposed method efficiently reduces computational costs for training and inference phases.

3D reconstruction from multi-view images is one of the fundamental challenges in computer vision and graphics. Recently, 3D Gaussian Splatting (3DGS) has emerged as a promising technique capable of real-time rendering with high-quality 3D reconstruction. This method utilizes 3D Gaussian representation and tile-based splatting techniques, bypassing the expensive neural field querying. Despite its potential, 3DGS encounters challenges such as needle-like artifacts, suboptimal geometries, and inaccurate normals caused by the Gaussians converging into anisotropic shapes with one dominant variance. We propose using the effective rank analysis to examine the shape statistics of 3D Gaussian primitives, and identify the Gaussians indeed converge into needle-like shapes with the effective rank 1. To address this, we introduce the effective rank as a regularization, which constrains the structure of the Gaussians. Our new regularization method enhances normal and geometry reconstruction while reducing needle-like artifacts. The approach can be integrated as an add-on module to other 3DGS variants, improving their quality without compromising visual fidelity. The project page is available at https://junhahyung.github.io/erankgs.github.io/.

Large language models (LLMs) have demonstrated remarkable in-context learning capabilities across diverse applications. In this work, we explore the effectiveness of LLMs for generating realistic synthetic tabular data, identifying key prompt design elements to optimize performance. We introduce EPIC, a novel approach that leverages balanced, grouped data samples and consistent formatting with unique variable mapping to guide LLMs in generating accurate synthetic data across all classes, even for imbalanced datasets. Evaluations on real-world datasets show that EPIC achieves state-of-the-art machine learning classification performance, significantly improving generation efficiency. These findings highlight the effectiveness of EPIC for synthetic tabular data generation, particularly in addressing class imbalance.

Unsupervised skill discovery is a learning paradigm that aims to acquire diverse behaviors without explicit rewards. However, it faces challenges in learning complex behaviors and often leads to learning unsafe or undesirable behaviors. For instance, in various continuous control tasks, current unsupervised skill discovery methods succeed in learning basic locomotions like standing but struggle with learning more complex movements such as walking and running. Moreover, they may acquire unsafe behaviors like tripping and rolling or navigate to undesirable locations such as pitfalls or hazardous areas. In response, we present **DoDont** (Do’s and Dont’s), an instruction-based skill discovery algorithm composed of two stages. First, in instruction learning stage, DoDont leverages action-free instruction videos to train an instruction network to distinguish desirable transitions from undesirable ones. Then, in the skill learning stage, the instruction network adjusts the reward function of the skill discovery algorithm to weight the desired behaviors. Specifically, we integrate the instruction network into a distance-maximizing skill discovery algorithm, where the instruction network serves as the distance function. Empirically, with less than 8 instruction videos, DoDont effectively learns desirable behaviors and avoids undesirable ones across complex continuous control tasks. Code and videos are available at https://mynsng.github.io/dodont/

Improving out-of-distribution (OOD) generalization during in-distribution (ID) adaptation is a primary goal of robust fine-tuning of zero-shot models beyond naive fine-tuning. However, despite decent OOD generalization performance from recent robust fine-tuning methods, confidence calibration for reliable model output has not been fully addressed. This work proposes a robust fine-tuning method that improves both OOD accuracy and confidence calibration simultaneously in vision language models. Firstly, we show that both OOD classification and OOD calibration errors have a shared upper bound consisting of two terms of ID data: 1) ID calibration error and 2) the smallest singular value of the ID input covariance matrix. Based on this insight, we design a novel framework that conducts fine-tuning with a constrained multimodal contrastive loss enforcing a larger smallest singular value, which is further guided by the self-distillation of a moving-averaged model to achieve calibrated prediction as well. Starting from empirical evidence supporting our theoretical statements, we provide extensive experimental results on ImageNet distribution shift benchmarks that demonstrate the effectiveness of our theorem and its practical implementation.

Recent advances in Neural Radiance Field (NeRF) show promising results in rendering realistic novel view images. However, NeRF and its variants assume that input images are captured using a pinhole camera and that subjects in images are always all-in-focus by tacit agreement. In this paper, we propose aperture-aware NeRF optimization and rendering methods using a thin-lens model (dubbed LensNeRF), which allows defocus images of any aperture size as input and output. To generalize a pinhole camera model to a thin-lens camera model in NeRF framework, we define multiple rays originating from the aperture area, solving world-to-pixel scale ambiguity. Also, we propose in-focus loss that assigns the given pixel color to points on the focus plane to alleviate the color ambiguity caused by the use of multiple rays. For the rigorous evaluation of the proposed method, we collect a real forward-facing dataset with different F-numbers for each viewpoint. Experimental results demonstrate that our method successfully fuses an aperture-size adjustable thin-lens camera model into the NeRF architecture, showing favorable qualitative and quantitative results compared to baseline models. The dataset will be made available.

Interactive segmentation methods have been investigated to address the potential need for additional refinement in automatic segmentation via human-in-the-loop techniques. For accurate segmentation of 3D images, we propose Slice-and-Conquer, a novel planar-to-3D pipeline formulating volumetric mask construction into two stages: 1) 2D interactive segmentation and 2) guided 3D segmentation. Specifically, the first stage enables users to focus on a single 2D slice and provides the corresponding 2D prediction results as strong shape priors. Taking the planar guidance, an accurate 3D mask can be constructed with minimal interactions. To support a flexible iterative refinement, our system recommends a next slice to annotate at the end of the second stage. Since volumetric segmentation can be completed by consecutively annotating a few recommended 2D slices, our method significantly reduces the cognitive burden of exploring volumetric space for users. Through extensive experiments on various datasets of 3D biomedical images, we demonstrate the effectiveness of the proposed pipeline.

In image classification, debiasing aims to train a classifier to be less susceptible to dataset bias, the strong correlation between peripheral attributes of data samples and a target class. For example, even if the frog class in the dataset mainly consists of frog images with a swamp background (i.e., bias aligned samples), a debiased classifier should be able to correctly classify a frog at a beach (i.e., bias conflicting samples). Recent debiasing approaches commonly use two components for debiasing, a biased model fB and a debiased model fD. fB is trained to focus on bias aligned samples (i.e., overfitted to the bias) while fD is mainly trained with bias conflicting samples by concentrating on samples which fB fails to learn, leading fD to be less susceptible to the dataset bias. While the state of the art debiasing techniques have aimed to better train fD, we focus on training fB, an overlooked component until now. Our empirical analysis reveals that removing the bias conflicting samples from the training set for fB is important for improving the debiasing performance of fD. This is due to the fact that the bias conflicting samples work as noisy samples for amplifying the bias for fB since those samples do not include the bias attribute. To this end, we propose a simple yet effective data sample selection method which removes the bias conflicting samples to construct a bias amplified dataset for training fB. Our data sample selection method can be directly applied to existing reweighting based debiasing approaches, obtaining consistent performance boost and achieving the state of the art performance on both synthetic and real-world datasets.

Despite the extensive applications of relation extraction (RE) tasks in various domains, little has been explored in the historical context, which contains promising data across hundreds and thousands of years. To promote the historical RE research, we present HistRED constructed from Yeonhaengnok. Yeonhaengnok is a collection of records originally written in Hanja, the classical Chinese writing, which has later been translated into Korean. HistRED provides bilingual annotations such that RE can be performed on Korean and Hanja texts. In addition, HistRED supports various self-contained subtexts with different lengths, from a sentence level to a document level, supporting diverse context settings for researchers to evaluate the robustness of their RE models. To demonstrate the usefulness of our dataset, we propose a bilingual RE model that leverages both Korean and Hanja contexts to predict relations between entities. Our model outperforms monolingual baselines on HistRED, showing that employing multiple language contexts supplements the RE predictions.

In this paper, we introduce ST-RAP, a novel Spatio-Temporal framework for Real estate APpraisal. ST-RAP employs a hierarchical architecture with a heterogeneous graph neural network to encapsulate temporal dynamics and spatial relationships simultaneously. Through comprehensive experiments on a large-scale real estate dataset, ST-RAP outperforms previous methods, demonstrating the significant benefits of integrating spatial and temporal aspects in real estate appraisal. Our code and dataset are available at https://github.com/dojeon-ai/STRAP.

This paper investigates Cross-Domain Sequential Recommendation (CDSR), a promising method that uses information from multiple domains (more than three) to generate accurate and diverse recommendations, and takes into account the sequential nature of user interactions. The effectiveness of these systems often depends on the complex interplay among the multiple domains. In this dynamic landscape, the problem of negative transfer arises, where heterogeneous knowledge between dissimilar domains leads to performance degradation due to differences in user preferences across these domains. As a remedy, we propose a new CDSR framework that addresses the problem of negative transfer by assessing the extent of negative transfer from one domain to another and adaptively assigning low weight values to the corresponding prediction losses. To this end, the amount of negative transfer is estimated by measuring the marginal contribution of each domain to model performance based on a cooperative game theory. In addition, a hierarchical contrastive learning approach that incorporates information from the sequence of coarse-level categories into that of fine-level categories (e.g., item level) when implementing contrastive learning was developed to mitigate negative transfer. Despite the potentially low relevance between domains at the fine-level, there may be higher relevance at the category level due to its generalised and broader preferences. We show that our model is superior to prior works in terms of model performance on two real-world datasets across ten different domains.

3D content manipulation is an important computer vision task with many real-world applications (e.g., product design, cartoon generation, and 3D Avatar editing). Recently proposed 3D CANs can generate diverse photorealistic 3D-aware contents using Neural Radiance fields (NeRF). However, manipulation of NeRF still remains a challenging problem since the visual quality tends to degrade after manipulation and suboptimal control handles such as 2D semantic maps are used for manipulations. While text-guided manipulations have shown potential in 3D editing, such approaches often lack locality. To overcome these problems, we propose Local Editing NeRF (LENeRF), which only requires text inputs for fine-grained and localized manipulation. Specifically, we present three add-on modules of LENeRF, the Latent Residual Mapper, the Attention Field Network, and the Deformation Network, which are jointly usedfor local manipulations of 3D features by estimating a 3D attention field. The 3D attention field is learned in an unsupervised way, by distilling the zero-shot mask generation capability of CLIP to the 3D space with multi-view guidance. We conduct diverse experiments and thorough evaluations both quantitatively and qualitatively.11We will make our code publicly available.

A User Next Location Prediction (UNLP) task, which predicts the next location that a user will move to given his/her trajectory, is an indispensable task for a wide range of applications. Previous studies using large-scale trajectory datasets in a single server have achieved remarkable performance in UNLP task. However, in real-world applications, legal and ethical issues have been raised regarding privacy concerns leading to restrictions against sharing human trajectory datasets to any other server. In response, Federated Learning (FL) has emerged to address the personal privacy issue by collaboratively training multiple clients (i.e., users) and then aggregating them. While previous studies employed FL for UNLP, they are still unable to achieve reliable performance because of the heterogeneity of clients' mobility. To tackle this problem, we propose the Federated Learning for Geographic Information (FedGeo), a FL framework specialized for UNLP, which alleviates the heterogeneity of clients' mobility and guarantees personal privacy protection. Firstly, we incorporate prior global geographic adjacency information to the local client model, since the spatial correlation between locations is trained partially in each client who has only a heterogeneous subset of the overall trajectories in FL. We also introduce a novel aggregation method that minimizes the gap between client models to solve the problem of client drift caused by differences between client models when learning with their heterogeneous data. Lastly, we probabilistically exclude clients with extremely heterogeneous data from the FL process by focusing on clients who visit relatively diverse locations. We show that FedGeo is superior to other FL methods for model performance in UNLP task. We also validated our model in a real-world application using our own customers' mobile phones and the FL agent system.

X-ray computed tomography (CT) is one of the most common imaging techniques used to diagnose various diseases in the medical field. Its high contrast sensitivity and spatial resolution allow the physician to observe details of body parts such as bones, soft tissue, blood vessels, etc. As it involves potentially harmful radiation exposure to patients and surgeons, however, reconstructing 3D CT volume from perpendicular 2D X-ray images is considered a promising alternative, thanks to its lower radiation risk and better accessibility. This is highly challenging though, since it requires reconstruction of 3D anatomical information from 2D images with limited views, where all the information is overlapped. In this paper, we propose PerX2CT, a novel CT reconstruction framework from X-ray that reflects the perspective projection scheme. Our proposed method provides a different combination of features for each coordinate which implicitly allows the model to obtain information about the 3D location. We reveal the potential to reconstruct the selected part of CT with high resolution by properly using the coordinate-wise local and global features. Our approach shows potential for use in clinical applications with low computational complexity and fast inference time, demonstrating superior performance than baselines in multiple evaluation metrics.

Video-to-video translation aims to generate video frames of a target domain from an input video. Despite its usefulness, the existing networks require enormous computations, necessitating their model compression for wide use. While there exist compression methods that improve computational efficiency in various image/video tasks, a generally-applicable compression method for video-to-video translation has not been studied much. In response, we present Shortcut-V2V, a general-purpose compression framework for video-to-video translation. Shortcut-V2V avoids full inference for every neighboring video frame by approximating the intermediate features of a current frame from those of the previous frame. Moreover, in our framework, a newly-proposed block called AdaBD adaptively blends and deforms features of neighboring frames, which makes more accurate predictions of the intermediate features possible. We conduct quantitative and qualitative evaluations using well-known video-to-video translation models on various tasks to demonstrate the general applicability of our framework. The results show that Shortcut-V2V achieves comparable performance compared to the original video-to-video translation model while saving 3.2-5.7× computational cost and 7.8-44× memory at test time. Our code and videos are available at https://shortcut-v2v.github.io/.

As recent advances in Neural Radiance Fields (NeRF) have enabled high-fidelity 3D face reconstruction and novel view synthesis, its manipulation also became an essential task in 3D vision. However, existing manipulation methods require extensive human labor, such as a user-provided semantic mask and manual attribute search unsuitable for non-expert users. Instead, our approach is designed to require a single text to manipulate a face reconstructed with NeRF. To do so, we first train a scene manipulator, a latent code-conditional deformable NeRF, over a dynamic scene to control a face deformation using the latent code. However, representing a scene deformation with a single latent code is unfavorable for compositing local deformations observed in different instances. As so, our proposed Position-conditional Anchor Compositor (PAC) learns to represent a manipulated scene with spatially varying latent codes. Their renderings with the scene manipulator are then optimized to yield high cosine similarity to a target text in CLIP embedding space for text-driven manipulation. To the best of our knowledge, our approach is the first to address the text-driven manipulation of a face reconstructed with NeRF. Extensive results, comparisons, and ablation studies demonstrate the effectiveness of our approach.

While recent 3D-aware generative models have shown photo-realistic image synthesis with multi-view consistency, the synthesized image quality degrades depending on the camera pose (e.g., a face with a blurry and noisy boundary at a side viewpoint). Such degradation is mainly caused by the difficulty of learning both pose consistency and photo-realism simultaneously from a dataset with heavily imbalanced poses. In this paper, we propose SideGAN, a novel 3D GAN training method to generate photo-realistic images irrespective of the camera pose, especially for faces of side-view angles. To ease the challenging problem of learning photo-realistic and pose-consistent image synthesis, we split the problem into two subproblems, each of which can be solved more easily. Specifically, we formulate the problem as a combination of two simple discrimination problems, one of which learns to discriminate whether a synthesized image looks real or not, and the other learns to discriminate whether a synthesized image agrees with the camera pose. Based on this, we propose a dual-branched discriminator with two discrimination branches. We also propose a pose-matching loss to learn the pose consistency of 3D GANs. In addition, we present a pose sampling strategy to increase learning opportunities for steep angles in a pose-imbalanced dataset. With extensive validation, we demonstrate that our approach enables 3D GANs to generate high-quality geometries and photo-realistic images irrespective of the camera pose.

Despite recent advancements in deep learning, deep neural networks continue to suffer from performance degradation when applied to new data that differs from training data. Test-time adaptation (TTA) aims to address this challenge by adapting a model to unlabeled data at test time. TTA can be applied to pretrained networks without modifying their training procedures, enabling them to utilize a well-formed source distribution for adaptation. One possible approach is to align the representation space of test samples to the source distribution (i.e., feature alignment). However, performing feature alignment in TTA is especially challenging in that access to labeled source data is restricted during adaptation. That is, a model does not have a chance to learn test data in a class-discriminative manner, which was feasible in other adaptation tasks (e.g., unsupervised domain adaptation) via supervised losses on the source data. Based on this observation, we propose a simple yet effective feature alignment loss, termed as Class-Aware Feature Alignment (CAFA), which simultaneously 1) encourages a model to learn target representations in a class-discriminative manner and 2) effectively mitigates the distribution shifts at test time. Our method does not require any hyper-parameters or additional losses, which are required in previous approaches. We conduct extensive experiments on 6 different datasets and show our proposed method consistently outperforms existing baselines.

Test-time adaptation (TTA) methods, which generally rely on the model’s predictions (e.g., entropy minimization) to adapt the source pretrained model to the unlabeled target domain, suffer from noisy signals originating from 1) incorrect or 2) open-set predictions. Long-term stable adaptation is hampered by such noisy signals, so training models without such error accumulation is crucial for practical TTA. To address these issues, including open-set TTA, we propose a simple yet effective sample selection method inspired by the following crucial empirical finding. While entropy minimization compels the model to increase the probability of its predicted label (i.e., confidence values), we found that noisy samples rather show decreased confidence values. To be more specific, entropy minimization attempts to raise the confidence values of an individual sample’s prediction, but individual confidence values may rise or fall due to the influence of signals from numerous other predictions (i.e., wisdom of crowds). Due to this fact, noisy signals misaligned with such ‘wisdom of crowds’, generally found in the correct signals, fail to raise the individual confidence values of wrong samples, despite attempts to increase them. Based on such findings, we filter out the samples whose confidence values are lower in the adapted model than in the original model, as they are likely to be noisy. Our method is widely applicable to existing TTA methods and improves their long-term adaptation performance in both image classification (e.g., 49.4% reduced error rates with TENT) and semantic segmentation (e.g., 11.7% gain in mIoU with TENT).

Existing text-to-image generation approaches have set high standards for photorealism and text-image correspondence, largely benefiting from web-scale text-image datasets, which can include up to 5 billion pairs. However, text-to-image generation models trained on domain-specific datasets, such as urban scenes, medical images, and faces, still suffer from low text-image correspondence due to the lack of text-image pairs. Additionally, collecting billions of text-image pairs for a specific domain can be time-consuming and costly. Thus, ensuring high text-image correspondence without relying on web-scale text-image datasets remains a challenging task. In this paper, we present a novel approach for enhancing text-image correspondence by leveraging available semantic layouts. Specifically, we propose a Gaussian-categorical diffusion process that simultaneously generates both images and corresponding layout pairs. Our experiments reveal that we can guide text-to-image generation models to be aware of the semantics of different image regions, by training the model to generate semantic labels for each pixel. We demonstrate that our approach achieves higher text-image correspondence compared to existing text-to-image generation approaches in the Multi-Modal CelebA-HQ and the Cityscapes dataset, where text-image pairs are scarce. Codes are available at https://pmh9960.github.io/research/GCDP.

Test-time adaptation (TTA) aims to adapt a pre-trained model to the target domain in a batch-by-batch manner during inference. While label distributions often exhibit imbalances in real-world scenarios, most previous TTA approaches typically assume that both source and target domain datasets have balanced label distribution. Due to the fact that certain classes appear more frequently in certain domains (e.g., buildings in cities, trees in forests), it is natural that the label distribution shifts as the domain changes. However, we discover that the majority of existing TTA methods fail to address the coexistence of covariate and label shifts. To tackle this challenge, we propose a novel label shift adapter that can be incorporated into existing TTA approaches to deal with label shifts during the TTA process effectively. Specifically, we estimate the label distribution of the target domain to feed it into the label shift adapter. Subsequently, the label shift adapter produces optimal parameters for target label distribution. By predicting only the parameters for a part of the pre-trained source model, our approach is computationally efficient and can be easily applied, regardless of the model architectures. Through extensive experiments, we demonstrate that integrating our strategy with TTA approaches leads to substantial performance improvements under the joint presence of label and covariate shifts.

Recently, unsupervised representation learning (URL) has improved the sample efficiency of Reinforcement Learning (RL) by pretraining a model from a large unlabeled dataset. The underlying principle of these methods is to learn temporally predictive representations by predicting future states in the latent space. However, an important challenge of this approach is the representational collapse, where the subspace of the latent representations collapses into a low-dimensional manifold. To address this issue, we propose a novel URL framework that causally predicts future states while increasing the dimension of the latent manifold by decorrelating the features in the latent space. Through extensive empirical studies, we demonstrate that our framework effectively learns predictive representations without collapse, which significantly improves the sample efficiency of state-of-the-art URL methods on the Atari 100k benchmark. The code is available at https://github.com/dojeon-ai/SimTPR.

In the field of unsupervised skill discovery (USD), a major challenge is limited exploration, primarily due to substantial penalties when skills deviate from their initial trajectories. To enhance exploration, recent methodologies employ auxiliary rewards to maximize the epistemic uncertainty or entropy of states. However, we have identified that the effectiveness of these rewards declines as the environmental complexity rises. Therefore, we present a novel USD algorithm, skill **disco**very with gui**dance** (**DISCO-DANCE**), which (1) selects the guide skill that possesses the highest potential to reach unexplored states, (2) guides other skills to follow guide skill, then (3) the guided skills are dispersed to maximize their discriminability in unexplored states. Empirical evaluation demonstrates that DISCO-DANCE outperforms other USD baselines in challenging environments, including two navigation benchmarks and a continuous control benchmark. Qualitative visualizations and code of DISCO-DANCE are available at https://mynsng.github.io/discodance/.

In Reinforcement Learning (RL), enhancing sample efficiency is crucial, particularly in scenarios when data acquisition is costly and risky. In principle, off-policy RL algorithms can improve sample efficiency by allowing multiple updates per environment interaction. However, these multiple updates often lead the model to overfit to earlier interactions, which is referred to as the loss of plasticity. Our study investigates the underlying causes of this phenomenon by dividing plasticity into two aspects. Input plasticity, which denotes the model's adaptability to changing input data, and label plasticity, which denotes the model's adaptability to evolving input-output relationships. Synthetic experiments on the CIFAR-10 dataset reveal that finding smoother minima of loss landscape enhances input plasticity, whereas refined gradient propagation improves label plasticity. Leveraging these findings, we introduce the **PLASTIC** algorithm, which harmoniously combines techniques to address both concerns. With minimal architectural modifications, PLASTIC achieves competitive performance on benchmarks including Atari-100k and Deepmind Control Suite. This result emphasizes the importance of preserving the model's plasticity to elevate the sample efficiency in RL. The code is available at https://github.com/dojeon-ai/plastic.

Estimating the energetic properties of molecular systems is a critical task in material design. Machine learning has shown remarkable promise on this task over classical force fields, but a fully data-driven approach suffers from limited labeled data; not just the amount of available data lacks, but the distribution of labeled examples is highly skewed to stable states. In this work, we propose a molecular representation learning method that extrapolates well beyond the training distribution, powered by physics-driven parameter estimation from classical energy equations and self-supervised learning inspired from masked language modeling. To ensure reliability of the proposed model, we introduce a series of novel evaluation schemes in multifaceted ways, beyond the energy or force accuracy that has been dominantly used. From extensive experiments, we demonstrate that the proposed method is effective in discovering molecular structures, outperforming other baselines. Furthermore, we extrapolate it to the chemical reaction pathways beyond stable states, taking a step towards physically reliable molecular representation learning.

Existing deep interactive colorization models have focused on ways to utilize various types of interactions, such as point-wise color hints, scribbles, or natural-language texts, as methods to reflect a user’s intent at runtime. However, another approach, which actively informs the user of the most effective regions to give hints for sketch image colorization, has been under-explored. This paper proposes a novel model-guided deep interactive colorization framework that reduces the required amount of user interactions, by prioritizing the regions in a colorization model. Our method, called GuidingPainter, prioritizes these regions where the model most needs a color hint, rather than just relying on the user’s manual decision on where to give a color hint. In our extensive experiments, we show that our approach outperforms existing interactive colorization methods in terms of the conventional metrics, such as PSNR and FID, and reduces required amount of interactions.

Machine learning-based algorithms using fully convolutional networks (FCNs) have been a promising option for medical image segmentation. However, such deep networks silently fail if input samples are drawn far from the training data distribution, thus causing critical problems in automatic data processing pipelines. To overcome such out-of-distribution (OoD) problems, we propose a novel OoD score formulation and its regularization strategy by applying an auxiliary add-on classifier to an intermediate layer of an FCN, where the auxiliary module is helfpul for analyzing the encoder output features by taking their class information into account. Our regularization strategy train the module along with the FCN via the principle of outlier exposure so that our model can be trained to distinguish OoD samples from normal ones without modifying the original network architecture. Our extensive experiment results demonstrate that the proposed approach can successfully conduct effective OoD detection without loss of segmentation performance. In addition, our module can provide reasonable explanation maps along with OoD scores, which can enable users to analyze the reliability of predictions.

Recent generative models based on neural radiance fields (NeRF) achieve the generation of diverse 3D-aware images. Despite the success, their applicability can be further expanded by incorporating with various types of user-specified conditions such as text and images. In this paper, we propose a novel approach called the conditional generative neural radiance fields (CG-NeRF), which generates multi-view images that reflect multimodal input conditions such as images or text. However, generating 3D-aware images from multimodal conditions bears several challenges. First, each condition type has different amount of information - e.g., the amount of information in text and color images are significantly different. Furthermore, the pose-consistency is often violated when diversifying the generated images from input conditions. Addressing such challenges, we propose 1) a unified architecture that effectively handles multiple types of conditions, and 2) the pose-consistent diversity loss for generating various images while maintaining the view consistency. Experimental results show that the proposed method maintains consistent image quality on various multimodal condition types and achieves superior fidelity and diversity compared to the existing NeRF-based generative models.

Point-interactive image colorization aims to colorize grayscale images when a user provides the colors for specific locations. It is essential for point-interactive colorization methods to appropriately propagate user-provided colors (i.e., user hints) in the entire image to obtain a reasonably colorized image with minimal user effort. However, existing approaches often produce partially colorized results due to the inefficient design of stacking convolutional layers to propagate hints to distant relevant regions. To address this problem, we present iColoriT, a novel point-interactive colorization Vision Transformer capable of propagating user hints to relevant regions, leveraging the global receptive field of Transformers. The self-attention mechanism of Transformers enables iColoriT to selectively colorize relevant regions with only a few local hints. Our approach colorizes images in real-time by utilizing pixel shuffling, an efficient upsampling technique that replaces the decoder architecture. Also, in order to mitigate the artifacts caused by pixel shuffling with large upsampling ratios, we present the local stabilizing layer. Extensive quantitative and qualitative results demonstrate that our approach highly out-performs existing methods for point-interactive colorization, producing accurately colorized images with a user's minimal effort. Official codes are available at https://pmh9960.github.io/research/iColoriT/.

This special section of theIEEE Transactions on Visualization and Computer Graphics (IEEE TVCG)presents the five most highly rated papers from the 2023 IEEE Pacific Visualization Symposium (IEEE PacificVis), hosted in Seoul, Korea from April 18 to Apr 21, 2023. IEEE PacificVis, sponsored by the IEEE Visualization and Graphics Technical Committee (VGTC), aims to foster greater exchange between visualization researchers and practitioners, especially in the Asia-Pacific region. This forum has grown to be a truly international event, attracting submissions and attendees from many countries, not only in the Asia-Pacific but also in Europe, America, and beyond. Thus, IEEE PacificVis is serving the additional purpose of sharing the latest advances in the field of visualization with researchers and practitioners in the region and, also, introducing research developments from the region to the broader international visualization research community.

With deep learning (DL) outperforming conventional methods for different tasks, much effort has been devoted to utilizing DL in various domains. Researchers and developers in the traffic domain have also designed and improved DL models for forecasting tasks such as estimation of traffic speed and time of arrival. However, there exist many challenges in analyzing DL models due to the black-box property of DL models and complexity of traffic data (i.e., spatio-temporal dependencies). Collaborating with domain experts, we design a visual analytics system, AttnAnalyzer, that enables users to explore how DL models make predictions by allowing effective spatio-temporal dependency analysis. The system incorporates dynamic time warping (DTW) and Granger causality tests for computational spatio-temporal dependency analysis while providing map, table, line chart, and pixel views to assist user to perform dependency and model behavior analysis. For the evaluation, we present three case studies showing how AttnAnalyzer can effectively explore model behaviors and improve model performance in two different road networks. We also provide domain expert feedback.

Recently, deep learning-based methods have drawn huge attention due to their simple yet high performance without domain knowledge in sound classification and localization tasks. However, a lack of gun sounds in existing datasets has been a major obstacle to implementing a support system to spot criminals from their gunshots by leveraging deep learning models. Since the occurrence of gunshot is rare and unpredictable, it is impractical to collect gun sounds in the real world. As an alternative, gun sounds can be obtained from an FPS game that is designed to mimic real-world warfare. The recent FPS game offers a realistic environment where we can safely collect gunshot data while simulating even dangerous situations. By exploiting the advantage of the game environment, we construct a gunshot dataset, namely BGG, for the firearm classification and gunshot localization tasks. The BGG dataset consists of 37 different types of firearms, distances, and directions between the sound source and a receiver. We carefully verify that the in-game gunshot data has sufficient information to identify the location and type of gunshots by training several sound classification and localization baselines on the BGG dataset. Afterward, we demonstrate that the accuracy of real-world firearm classification and localization tasks can be enhanced by utilizing the BGG dataset.

Predicting traffic conditions is tremendously challenging since every road is highly dependent on each other, both spatially and temporally. Recently, to capture this spatial and temporal dependency, specially designed architectures such as graph convolutional networks and temporal convolutional networks have been introduced. While there has been remarkable progress in traffic forecasting, we found that deep-learning-based traffic forecasting models still fail in certain patterns, mainly in event situations (e.g., rapid speed drops). Although it is commonly accepted that these failures are due to unpredictable noise, we found that these failures can be corrected by considering previous failures. Specifically, we observe autocorrelated errors in these failures, which indicates that some predictable information remains. In this study, to capture the correlation of errors, we introduce ResCAL, a residual estimation module for traffic forecasting, as a widely applicable add-on module to existing traffic forecasting models. Our ResCAL calibrates the prediction of the existing models in real time by estimating future errors using previous errors and graph signals. Extensive experiments on METR-LA and PEMS-BAY demonstrate that our ResCAL can correctly capture the correlation of errors and correct the failures of various traffic forecasting models in event situations.

Semantically meaningful sentence embeddings are important for numerous tasks in natural language processing. To obtain such embeddings, recent studies explored the idea of utilizing synthetically generated data from pretrained language models(PLMs) as a training corpus. However, PLMs often generate sentences different from the ones written by human. We hypothesize that treating all these synthetic examples equally for training can have an adverse effect on learning semantically meaningful embeddings. To analyze this, we first train a classifier that identifies machine-written sentences and observe that the linguistic features of the sentences identified as written by a machine are significantly different from those of human-written sentences. Based on this, we propose a novel approach that first trains the classifier to measure the importance of each sentence. The distilled information from the classifier is then used to train a reliable sentence embedding model. Through extensive evaluation on four real-world datasets, we demonstrate that our model trained on synthetic data generalizes well and outperforms the baselines.

Hojun Cho, Dohee Kim, Seungwoo Ryu, ChaeHun Park, Hyungjong Noh, Jeong-in Hwang, Minseok Choi, Edward Choi, Jaegul Choo. Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing. 2022.

In retrieval-based dialogue systems, a response selection model acts as a ranker to select the most appropriate response among several candidates.However, such selection models tend to rely on context-response content similarity, which makes models vulnerable to adversarial responses that are semantically similar but not relevant to the dialogue context.Recent studies have shown that leveraging these adversarial responses as negative training samples is useful for improving the discriminating power of the selection model.Nevertheless, collecting human-written adversarial responses is expensive, and existing synthesizing methods often have limited scalability.To overcome these limitations, this paper proposes a simple but efficient method for generating adversarial negative responses leveraging a large-scale language model.Experimental results on dialogue selection tasks show that our method outperforms other methods of synthesizing adversarial negative responses.These results suggest that our method can be an effective alternative to human annotators in generating adversarial responses.Our dataset and generation code is available at https://github.com/leenw23/ generating-negatives-by-gpt3.Create five irrelevant responses containing keywords of the given dialogue context: ### Dialogue context: """ A: Yes, it's the same every day … B: I'll look forward to that, I can't stand… """ Create five irrelevant responses containing keywords of the given dialogue context: 1. { other examples } ... ...

Single-image 3D human reconstruction aims to reconstruct the 3D textured surface of the human body given a single image. While implicit function-based methods recently achieved reasonable reconstruction performance, they still bear limitations showing degraded quality in both surface geometry and texture from an unobserved view. In response, to generate a realistic textured surface, we propose ReFu, a coarse-to-fine approach that refines the projected backside view image and fuses the refined image to predict the final human body. To suppress the diffused occupancy that causes noise in projection images and reconstructed meshes, we propose to train occupancy probability by simultaneously utilizing 2D and 3D supervisions with occupancy-based volume rendering. We also introduce a refinement architecture that generates detail-preserving backside-view images with front-to-back warping. Extensive experiments demonstrate that our method achieves state-of-the-art performance in 3D human reconstruction from a single image, showing enhanced geometry and texture quality from an unobserved view.

Conditional generative adversarial networks (cGANs) have shown superior results in class-conditional generation tasks. To simultaneously control multiple conditions, cGANs require multi-label training datasets, where multiple labels can be assigned to each data instance. Nevertheless, the tremendous annotation cost limits the accessibility of multi-label datasets in real-world scenarios. Therefore, in this study we explore the practical setting called the single positive setting, where each data instance is annotated by only one positive label with no explicit negative labels. To generate multi-label data in the single positive setting, we propose a novel sampling approach called single-to-multi-label (S2M) sampling, based on the Markov chain Monte Carlo method. As a widely applicable “add-on” method, our proposed S2M sampling method enables existing unconditional and conditional GANs to draw high-quality multi-label data with a minimal annotation cost. Extensive experiments on real image datasets verify the effectiveness and correctness of our method, even when compared to a model trained with fully annotated datasets.

Time series forecasting has become a critical task due to its high practicality in real-world applications such as traffic, energy consumption, economics and finance, and disease analysis. Recent deep-learning-based approaches have shown remarkable success in time series forecasting. Nonetheless, due to the dynamics of time series data, deep networks still suffer from unstable training and overfitting. Inconsistent patterns appearing in real-world data lead the model to be biased to a particular pattern, thus limiting the generalization. In this work, we introduce the dynamic error bounds on training loss to address the overfitting issue in time series forecasting. Consequently, we propose a regularization method called WaveBound which estimates the adequate error bounds of training loss for each time step and feature at each iteration. By allowing the model to focus less on unpredictable data, WaveBound stabilizes the training process, thus significantly improving generalization. With the extensive experiments, we show that WaveBound consistently improves upon the existing models in large margins, including the state-of-the-art model.

Optical Chemical Structure Recognition (OCSR) deals with the translation from chemical images to molecular structures, this being the main way chemical compounds are depicted in scientific documents. Traditionally, rule-based methods have followed a framework based on the detection of chemical entities, such as atoms and bonds, followed by a compound structure reconstruction step. Recently, neural architectures analog to image captioning have been explored to solve this task, yet they still show to be data inefficient, using millions of examples just to show performances comparable with traditional methods. Looking to motivate and benchmark new approaches based on atomic-level entities detection and graph reconstruction, we present CEDe, a unique collection of chemical entity bounding boxes manually curated by experts for scientific literature datasets. These annotations combine to more than 700,000 chemical entity bounding boxes with the necessary information for structure reconstruction. Also, a large synthetic dataset containing one million molecular images and annotations is released in order to explore transfer-learning techniques that could help these architectures perform better under low-data regimes. Benchmarks show that detection-reconstruction based models can achieve performances on par with or better than image captioning-like models, even with 100x fewer training examples.

Interactive Recommender Systems (IRSs) have attracted a lot of attention, due to their ability to model interactive processes between users and recommender systems. Numerous approaches have adopted Reinforcement Learning (RL) algorithms, as these can directly maximize users' cumulative rewards. In IRS, researchers commonly utilize publicly available review datasets to compare and evaluate algorithms. However, user feedback provided in public datasets merely includes instant responses (e.g., a rating), with no inclusion of delayed responses (e.g., the dwell time and the lifetime value). Thus, the question remains whether these review datasets are an appropriate choice to evaluate the long-term effects in IRS. In this work, we revisited experiments on IRS with review datasets and compared RL-based models with a simple reward model that greedily recommends the item with the highest one-step reward. Following extensive analysis, we can reveal three main findings: First, a simple greedy reward model consistently outperforms RL-based models in maximizing cumulative rewards. Second, applying higher weighting to long-term rewards leads to degradation of recommendation performance. Third, user feedbacks have mere long-term effects in the benchmark datasets. Based on our findings, we conclude that a dataset has to be carefully verified and that a simple greedy baseline should be included for a proper evaluation of RL-based IRS approaches. Our code and dataset are available at https://github.com/dojeon-ai/irs_validation.

This paper presents a personalized character recommendation system for Multiplayer Online Battle Arena (MOBA) games which are considered as one of the most popular online video game genres around the world. When playing MOBA games, players go through a draft stage, where they alternately select a virtual character to play. When drafting, players select characters by not only considering their character preferences, but also the synergy and competence of their team’s character combination. However, the complexity of drafting induces difficulties for beginners to choose the appropriate characters based on the characters of their team while considering their own champion preferences. To alleviate this problem, we propose DraftRec, a novel hierarchical model which recommends characters by considering each player’s champion preferences and the interaction between the players. DraftRec consists of two networks: the player network and the match network. The player network captures the individual player’s champion preference, and the match network integrates the complex relationship between the players and their respective champions. We train and evaluate our model from a manually collected 280,000 matches of League of Legends and a publicly available 50,000 matches of Dota2. Empirically, our method achieved state-of-the-art performance in character recommendation and match outcome prediction task. Furthermore, a comprehensive user survey confirms that DraftRec provides convincing and satisfying recommendations to the real-world players. Our code and dataset are available at https://github.com/dojeon-ai/DraftRec .

Video generation models often operate under the assumption of fixed frame rates, which leads to suboptimal performance when it comes to handling flexible frame rates (e.g., increasing the frame rate of the more dynamic portion of the video as well as handling missing video frames). To resolve the restricted nature of existing video generation models' ability to handle arbitrary timesteps, we propose continuous-time video generation by combining neural ODE (Vid-ODE) with pixel-level video processing techniques. Using ODE-ConvGRU as an encoder, a convolutional version of the recently proposed neural ODE, which enables us to learn continuous-time dynamics, Vid-ODE can learn the spatio-temporal dynamics of input videos of flexible frame rates. The decoder integrates the learned dynamics function to synthesize video frames at any given timesteps, where the pixel-level composition technique is used to maintain the sharpness of individual frames. With extensive experiments on four real-world video datasets, we verify that the proposed Vid-ODE outperforms state-of-the-art approaches under various video generation settings, both within the trained time range (interpolation) and beyond the range (extrapolation). To the best of our knowledge, Vid-ODE is the first work successfully performing continuous-time video generation using real-world videos.

Cheonbok Park, Yunwon Tae, TaeHee Kim, Soyoung Yang, Mohammad Azam Khan, Lucy Park, Jaegul Choo. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Success of training deep learning models largely depends on the amount and quality of training data. Although numerous data augmentation techniques have already been pro- posed for certain domains such as computer vision where simple schemes such as rotation and flipping have been shown to be effective, other domains such as time-series data have a relatively smaller set of augmentation techniques readily available. Data imbalance is a phenomenon often observed in real-world data. However, a simple oversampling technique may make a model vulnerable to overfitting, so a proper data augmentation is desired. To tackle these problems, we propose a novel data augmentation method that utilizes the latent vectors of an autoencoder in a novel way. When input data are perturbed in its latent space, their reconstructed data retains properties similar to the original one. In con- trast, adversarial augmentation is a technique to train robust deep neural networks against unforeseen data shifts or corruptions by providing a downstream model with samples that are difficult to predict. Our method adversarially perturbs input data in its latent space so that the augmented data is diverse and conducive to reducing test error of a downstream model. The experimental results demonstrated that our method achieves the right balance, significantly modifying the input data to help generalization while retaining its realism.

Enhancing the generalization capability of deep neural networks to unseen domains is crucial for safety-critical applications in the real world such as autonomous driving. To address this issue, this paper proposes a novel instance selective whitening loss to improve the robustness of the segmentation networks for unseen domains. Our approach disentangles the domain-specific style and domain-invariant content encoded in higher-order statistics (i.e., feature covariance) of the feature representations and selectively removes only the style information causing domain shift. As shown in Fig. 1, our method provides reasonable predictions for (a) low-illuminated, (b) rainy, and (c) unseen structures. These types of images are not included in the training dataset, where the baseline shows a significant performance drop, contrary to ours. Being simple yet effective, our approach improves the robustness of various backbone networks without additional computational cost. We conduct extensive experiments in urban-scene segmentation and show the superiority of our approach to existing work. Our code is available at this link1.

The task of image-based virtual try-on aims to transfer a target clothing item onto the corresponding region of a person, which is commonly tackled by fitting the item to the desired body part and fusing the warped item with the person. While an increasing number of studies have been conducted, the resolution of synthesized images is still limited to low (e.g., 256×192), which acts as the critical limitation against satisfying online consumers. We argue that the limitation stems from several challenges: as the resolution increases, the artifacts in the misaligned areas between the warped clothes and the desired clothing regions become noticeable in the final results; the architectures used in existing methods have low performance in generating high-quality body parts and maintaining the texture sharpness of the clothes. To address the challenges, we propose a novel virtual try-on method called VITON-HD that successfully synthesizes 1024×768 virtual try-on images. Specifically, we first prepare the segmentation map to guide our virtual try-on synthesis, and then roughly fit the target clothing item to a given person’s body. Next, we propose ALIgnment-Aware Segment (ALIAS) normalization and ALIAS generator to handle the misaligned areas and preserve the details of 1024×768 inputs. Through rigorous comparison with existing methods, we demonstrate that VITON-HD highly surpasses the baselines in terms of synthesized image quality both qualitatively and quantitatively.

Facial attribute editing aims to manipulate the image with the desired attribute while preserving the other details. Recently, generative adversarial networks along with the encoder-decoder architecture have been utilized for this task owing to their ability to create realistic images. However, the existing methods for the unpaired dataset cannot still preserve the attribute-irrelevant regions properly due to the absence of the ground truth image. This work proposes a novel, intuitive loss function called the CAM-consistency loss, which improves the consistency of an input image in image translation. While the existing cycle-consistency loss ensures that the image can be translated back, our approach makes the model further preserve the attribute-irrelevant regions even in a single translation to another domain by using the Grad-CAM output computed from the discriminator. Our CAM-consistency loss directly optimizes such a Grad-CAM output from the discriminator during training, in order to properly capture which local regions the generator should change while keeping the other regions unchanged. In this manner, our approach allows the generator and the discriminator to collaborate with each other to improve the image translation quality. In our experiments, we validate the effectiveness and versatility of our proposed CAM-consistency loss by applying it to several representative models for facial image editing, such as StarGAN, AttGAN, and STGAN.

During the fine-tuning phase of transfer learning, the pretrained vocabulary remains unchanged, while model parameters are updated.The vocabulary generated based on the pretrained data is suboptimal for downstream data when domain discrepancy exists.We propose to consider the vocabulary as an optimizable parameter, allowing us to update the vocabulary by expanding it with domain-specific vocabulary based on a tokenization statistic.Furthermore, we preserve the embeddings of the added words from overfitting to downstream data by utilizing knowledge learned from a pretrained language model with a regularization term.Our method achieved consistent performance improvements on diverse domains (i.e.,

This paper proposes a novel generative model called PUGAN, which progressively synthesizes high-quality audio in a raw waveform. Progressive upsampling GAN (PUGAN) leverages the progressive generation of higher-resolution output by stacking multiple encoder-decoder architectures. Compared to an existing state-of-the-art model called WaveGAN, which uses a single decoder architecture, our model generates audio signals and converts them to a higher resolution in a progressive manner, while using a significantly smaller number of parameters, e.g., 3.17x smaller for 16 kHz output, than WaveGAN. Our experiments show that the audio signals can be generated in real time with a comparable quality to that of WaveGAN in terms of the inception scores and human perception.

Identifying unexpected objects on roads in semantic segmentation (e.g., identifying dogs on roads) is crucial in safetycritical applications. Existing approaches use images of unexpected objects from external datasets or require additional training (e.g., retraining segmentation networks or training an extra network), which necessitate a non-trivial amount of labor intensity or lengthy inference time. One possible alternative is to use prediction scores of a pretrained network such as the max logits (i.e., maximum values among classes before the final softmax layer) for detecting such objects. However, the distribution of max logits of each predicted class is significantly different from each other, which degrades the performance of identifying unexpected objects in urban-scene segmentation. To address this issue, we propose a simple yet effective approach that standardizes the max logits in order to align the different distributions and reflect the relative meanings of max logits within each predicted class. Moreover, we consider the local regions from two different perspectives based on the intuition that neighboring pixels share similar semantic information. In contrast to previous approaches, our method does not utilize any external datasets or require additional training, which makes our method widely applicable to existing pretrained segmentation models. Such a straightforward approach achieves a new state-of-the-art performance on the publicly available Fishyscapes Lost & Found leader-board with a large margin. Our code is publicly available at this link1.

Deep neural networks often make decisions based on the spurious correlations inherent in the dataset, failing to generalize in an unbiased data distribution. Although previous approaches pre-define the type of dataset bias to prevent the network from learning it, recognizing the bias type in the real dataset is often prohibitive. This paper proposes a novel bias-tailored augmentation-based approach, BiaSwap, for learning debiased representation without requiring supervision on the bias type. Assuming that the bias corresponds to the easy-to-learn attributes, we sort the training images based on how much a biased classifier can exploits them as shortcut and divide them into bias-guiding and bias-contrary samples in an unsupervised manner. Afterwards, we integrate the style-transferring module of the image translation model with the class activation maps of such biased classifier, which enables to primarily transfer the bias attributes learned by the classifier. Therefore, given the pair of bias-guiding and bias-contrary, BiaSwap generates the bias-swapped image which contains the bias attributes from the bias-contrary images, while preserving bias-irrelevant ones in the bias-guiding images. Given such augmented images, BiaSwap demonstrates the superiority in debiasing against the existing baselines over both synthetic and real-world datasets. Even without careful supervision on the bias, BiaSwap achieves a remarkable performance on both unbiased and bias-guiding samples, implying the improved generalization capability of the model.

Deep neural networks for automatic image colorization often suffer from the color-bleeding artifact, a problematic color spreading near the boundaries between adjacent objects. Such color-bleeding artifacts debase the reality of generated outputs, limiting the applicability of colorization models in practice. Although previous approaches have attempted to address this problem in an automatic manner, they tend to work only in limited cases where a high contrast of gray-scale values are given in an input image. Alternatively, leveraging user interactions would be a promising approach for solving this color-breeding artifacts. In this paper, we propose a novel edge-enhancing network for the regions of interest via simple user scribbles indicating where to enhance. In addition, our method requires a minimal amount of effort from users for their satisfactory enhancement. Experimental results demonstrate that our interactive edge-enhancing approach effectively improves the color-bleeding artifacts compared to the existing baselines across various datasets.

To tackle ever-increasing city traffic congestion problems, researchers have proposed deep learning models to aid decision-makers in the traffic control domain. Although the proposed models have been remarkably improved in recent years, there are still questions that need to be answered before deploying models. For example, it is difficult to figure out which models provide state-of-the-art performance, as recently proposed models have often been evaluated with different datasets and experiment environments. It is also difficult to determine which models would work when traffic conditions change abruptly (e.g., rush hour). In this work, we conduct two experiments to answer the two questions. In the first experiment, we conduct an experiment with the state-of-the-art models and the identical public datasets to compare model performance under a consistent experiment environment. We then extract a set of temporal regions in the datasets, whose speeds change abruptly and use these regions to explore model performance with difficult intervals. The experiment results indicate that Graph-WaveNet and GMAN show better performance in general. We also find that prediction models tend to have varying performances with data and intervals, which calls for in-depth analysis of models on difficult intervals for real-world deployment.

The hair and beauty industry is a fast-growing industry. This led to the development of various applications, such as virtual hair dyeing or hairstyle translations, to satisfy the customer needs. Although several hairstyle datasets are available for these applications, they often consist of a relatively small number of images with low resolution, thus limiting their performance on high-quality hair editing. In response, we introduce a novel large-scale Korean hairstyle dataset, K-hairstyle, containing 500,000 high-resolution images. In addition, K-hairstyle includes various hair attributes annotated by Korean expert hairstylists as well as hair segmentation masks. We validate the effectiveness of our dataset via several applications, such as hair dyeing, hairstyle translation, and hairstyle classification. K-hairstyle is publicly available at https://psh01087.github.io/K-Hairstyle/.

Kyeongpil Kang, Kyohoon Jin, Soyoung Yang, Soojin Jang, Jaegul Choo, Youngbin Kim. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2021.

Image classification models tend to make decisions based on peripheral attributes of data items that have strong correlation with a target variable (i.e., dataset bias). These biased models suffer from the poor generalization capability when evaluated on unbiased datasets. Existing approaches for debiasing often identify and emphasize those samples with no such correlation (i.e., bias-conflicting) without defining the bias type in advance. However, such bias-conflicting samples are significantly scarce in biased datasets, limiting the debiasing capability of these approaches. This paper first presents an empirical analysis revealing that training with "diverse" bias-conflicting samples beyond a given training set is crucial for debiasing as well as the generalization capability. Based on this observation, we propose a novel feature-level data augmentation technique in order to synthesize diverse bias-conflicting samples. To this end, our method learns the disentangled representation of (1) the intrinsic attributes (i.e., those inherently defining a certain class) and (2) bias attributes (i.e., peripheral attributes causing the bias), from a large number of bias-aligned samples, the bias attributes of which have strong correlation with the target variable. Using the disentangled representation, we synthesize bias-conflicting samples that contain the diverse intrinsic attributes of bias-aligned samples by swapping their latent features. By utilizing these diversified bias-conflicting features during the training, our approach achieves superior classification accuracy and debiasing results against the existing baselines on both synthetic and real-world datasets.

Capturing human annotators' subjective responses in image annotation has become crucial as vision-based classifiers expand the range of application areas. While there has been significant progress in image annotation interface design in general, relatively little research has been conducted to understand how to elicit reliable and cost-efficient human annotation when the nature of the task includes a certain level of subjectivity. To bridge this gap, we aim to understand how different sampling methods in image batch labeling, a design that allows human annotators to label a batch of images simultaneously, can impact human annotation performances. In particular, we developed three different strategies in forming image batches: (1) uncertainty-based labeling (UL) that prioritizes images that a classifier predicts with the highest uncertainty, (2) certainty-based labeling (CL), a reverse strategy of UL, and (3) random, a baseline approach that randomly selects images. Although UL and CL solely select images to be labeled from a classifier's point of view, we hypothesized that human-side perception and labeling performance may also vary depending on the different sampling strategies. In our study, we observed that participants were able to recognize a different level of perceived cognitive load across three conditions (CL the easiest while UL the most difficult). We also observed a trade-off between annotation task effectiveness (CL and UL more reliable than random) and task efficiency (UL the most efficient while CL the least efficient). Based on the results, we discuss the implications of design and possible future research directions of image batch labeling.

To mitigate the pain of manually tuning hyperparameters of deep neural networks, automated machine learning (AutoML) methods have been developed to search for an optimal set of hyperparameters in large combinatorial search spaces. However, the search results of AutoML methods significantly depend on initial configurations, making it a non-trivial task to find a proper configuration. Therefore, human intervention via a visual analytic approach bears huge potential in this task. In response, we propose HyperTendril, a web-based visual analytics system that supports user-driven hyperparameter tuning processes in a model-agnostic environment. HyperTendril takes a novel approach to effectively steering hyperparameter optimization through an iterative, interactive tuning procedure that allows users to refine the search spaces and the configuration of the AutoML method based on their own insights from given results. Using HyperTendril, users can obtain insights into the complex behaviors of various hyperparameter search algorithms and diagnose their configurations. In addition, HyperTendril supports variable importance analysis to help the users refine their search spaces based on the analysis of relative importance of different hyperparameters and their interaction effects. We present the evaluation demonstrating how HyperTendril helps users steer their tuning processes via a longitudinal user study based on the analysis of interaction logs and in-depth interviews while we deploy our system in a professional industrial environment.

Predicting road traffic speed is a challenging task due to different types of roads, abrupt speed change and spatial dependencies between roads; it requires the modeling of dynamically changing spatial dependencies among roads and temporal patterns over long input sequences. This paper proposes a novel spatio-temporal graph attention (ST-GRAT) that effectively captures the spatio-temporal dynamics in road networks. The novel aspects of our approach mainly include spatial attention, temporal attention, and spatial sentinel vectors. The spatial attention takes the graph structure information (e.g., distance between roads) and dynamically adjusts spatial correlation based on road states. The temporal attention is responsible for capturing traffic speed changes, and the sentinel vectors allow the model to retrieve new features from spatially correlated nodes or preserve existing features. The experimental results show that ST-GRAT outperforms existing models, especially in difficult conditions where traffic speeds rapidly change (e.g., rush hours). We additionally provide a qualitative study to analyze when and where ST-GRAT tended to make accurate predictions during rush-hour times.

Despite remarkable success in unpaired image-to-image translation, existing systems still require a large amount of labeled images. This is a bottleneck for their real-world applications; in practice, a model trained on labeled CelebA dataset does not work well for test images from a different distribution -- greatly limiting their application to unlabeled images of a much larger quantity. In this paper, we attempt to alleviate this necessity for labeled data in the facial image translation domain. We aim to explore the degree to which you can discover novel attributes from unlabeled faces and perform high-quality translation. To this end, we use prior knowledge about the visual world as guidance to discover novel attributes and transfer them via a novel normalization method. Experiments show that our method trained on unlabeled data produces high-quality translations, preserves identity, and be perceptually realistic, as good as, or better than, state-of-the-art methods trained on labeled data.

This paper exploits the intrinsic features of urban-scene images and proposes a general add-on module, called height-driven attention networks (HANet), for improving semantic segmentation for urban-scene images. It emphasizes informative features or classes selectively according to the vertical position of a pixel. The pixel-wise class distributions are significantly different from each other among horizontally segmented sections in the urban-scene images. Likewise, urban-scene images have their own distinct characteristics, but most semantic segmentation networks do not reflect such unique attributes in the architecture. The proposed network architecture incorporates the capability exploiting the attributes to handle the urban scene dataset effectively. We validate the consistent performance (mIoU) increase of various semantic segmentation models on two datasets when HANet is adopted. This extensive quantitative analysis demonstrates that adding our module to existing models is easy and cost-effective. Our method achieves a new state-of-the-art performance on the Cityscapes benchmark with a large margin among ResNet101 based segmentation models. Also, we show that the proposed model is coherent with the facts observed in the urban scene by visualizing and interpreting the attention map. Our code and trained models are publicly available.

This paper tackles the automatic colorization task of a sketch image given an already-colored reference image. Colorizing a sketch image is in high demand in comics, animation, and other content creation applications, but it suffers from information scarcity of a sketch image. To address this, a reference image can render the colorization process in a reliable and user-driven manner. However, it is difficult to prepare for a training data set that has a sufficient amount of semantically meaningful pairs of images as well as the ground truth for a colored image reflecting a given reference (e.g., coloring a sketch of an originally blue car given a reference green car). To tackle this challenge, we propose to utilize the identical image with geometric distortion as a virtual reference, which makes it possible to secure the ground truth for a colored output image. Furthermore, it naturally provides the ground truth for dense semantic correspondence, which we utilize in our internal attention mechanism for color transfer from reference to sketch input. We demonstrate the effectiveness of our approach in various types of sketch image colorization via quantitative as well as qualitative evaluation against existing methods.

Many machine learning algorithms are trained and evaluated by splitting data from a single source into training and test sets. While such focus on in-distribution learning scenarios has led to interesting advancement, it has not been able to tell if models are relying on dataset biases as shortcuts for successful prediction (e.g., using snow cues for recognising snowmobiles), resulting in biased models that fail to generalise when the bias shifts to a different class. The cross-bias generalisation problem has been addressed by de-biasing training data through augmentation or re-sampling, which are often prohibitive due to the data collection cost (e.g., collecting images of a snowmobile on a desert) and the difficulty of quantifying or expressing biases in the first place. In this work, we propose a novel framework to train a de-biased representation by encouraging it to be different from a set of representations that are biased by design. This tactic is feasible in many scenarios where it is much easier to define a set of biased representations than to define and quantify bias. We demonstrate the efficacy of our method across a variety of synthetic and real-world biases; our experiments show that the method discourages models from taking bias shortcuts, resulting in improved generalisation. Source code is available at https://github.com/clovaai/rebias.

Multivariate time-series prediction is a common task, but it often becomes challenging due to missing data caused by unreliable sensors and other issues. In fact, inaccurate imputation of missing values can degrade the downstream prediction performance, so it may be better not to rely on the estimated values of missing data. Furthermore, observed data may contain noise, so denoising them can be helpful for the main task at hand. In response, we propose a novel approach that can automatically utilize the optimal combination of the observed and the estimated values to generate not only complete, but also noise-reduced data by our own gating mechanism. We evaluate our model on incomplete real-world time-series datasets and achieved state-of-the-art performance. Moreover, we present in-depth studies using a carefully designed, synthetic multivariate time-series dataset to verify the effectiveness of the proposed model. The ablation studies and the experimental analysis of the proposed gating mechanism show that it works as an effective denoising and imputation method for time-series classification tasks.

Probabilistic topic models, which can discover hidden patterns in documents, have been extensively studied. However, rather than learning from a single document collection, numerous real-world applications demand a comprehensive understanding of the relationships among various document sets. To address such needs, this article proposes a new model that can identify the common and discriminative aspects of multiple datasets. Specifically, our proposed method is a Bayesian approach that represents each document as a combination of common topics (shared across all document sets) and distinctive topics (distributions over words that are exclusive to a particular dataset). Through extensive experiments, we demonstrate the effectiveness of our method compared with state-of-the-art models. The proposed model can be useful for “comparative thinking” analysis in real-world document collections.

Generalized nucleus segmentation techniques can contribute greatly to reducing the time to develop and validate visual biomarkers for new digital pathology datasets. We summarize the results of MoNuSeg 2018 Challenge whose objective was to develop generalizable nuclei segmentation techniques in digital pathology. The challenge was an official satellite event of the MICCAI 2018 conference in which 32 teams with more than 80 participants from geographically diverse institutes participated. Contestants were given a training set with 30 images from seven organs with annotations of 21,623 individual nuclei. A test dataset with 14 images taken from seven organs, including two organs that did not appear in the training set was released without annotations. Entries were evaluated based on average aggregated Jaccard index (AJI) on the test set to prioritize accurate instance segmentation as opposed to mere semantic segmentation. More than half the teams that completed the challenge outperformed a previous baseline. Among the trends observed that contributed to increased accuracy were the use of color normalization as well as heavy data augmentation. Additionally, fully convolutional networks inspired by variants of U-Net, FCN, and Mask-RCNN were popularly used, typically based on ResNet or VGG base architectures. Watershed segmentation on predicted semantic segmentation maps was a popular post-processing strategy. Several of the top techniques compared favorably to an individual human annotator and can be used with confidence for nuclear morphometrics.

The problem of generating a set of diverse paraphrase sentences while (1) not compromising the original meaning of the original sentence, and (2) imposing diversity in various semantic aspects, such as a lexical or syntactic structure, is examined. Existing work on paraphrase generation has focused more on the former, and the latter was trained as a fixed style transfer, such as transferring from positive to negative sentiments, even at the cost of losing semantics. In this work, we consider style transfer as a means of imposing diversity, with a paraphrasing correctness constraint that the target sentence must remain a paraphrase of the original sentence. However, our goal is to maximize the diversity for a set of k generated paraphrases, denoted as the diversified paraphrase (DP) problem. Our key contribution is deciding the style guidance at generation towards the direction of increasing the diversity of output with respect to those generated previously. As pre-materializing training data for all style decisions is impractical, we train with biased data, but with debiasing guidance. Compared to state-of-the-art methods, our proposed model can generate more diverse and yet semantically consistent paraphrase sentences. That is, our model, trained with the MSCOCO dataset, achieves the highest embedding scores, .94/.95/.86, similar to state-of-the-art results, but with a lower mBLEU score (more diverse) by 8.73%.

Document labeling is a critical step in building various machine learning applications. However, the step can be time-consuming and arduous, requiring a significant amount of human efforts. To support an efficient document labeling environment, we present a system called Attentive Interactive Labeling Assistant (AILA). In its core, AILA uses Interactive Attention Module (IAM), a novel module that visually highlights words in a document that labelers may pay attention to when labeling a document. IAM utilizes attention-based Deep Neural Networks which not only support a prediction of which words to highlight but also enable labelers to indicate words that should be assigned a high attention weight while labeling to improve the future quality of word prediction.We evaluated the labeling efficiency and the accuracy by comparing the conditions with and without IAM in our study. The results showed that participants' labeling efficiency increased significantly under the condition with IAM than the condition without IAM, while the two conditions maintained roughly the same labeling accuracy.

Deep neural networks improved the accuracy of sequential recommendation approach which takes into account the sequential patterns of user logs, e.g., a purchase history of a user. However, incorporating only the individual's recent logs may not be sufficient in properly reflecting global preferences and trends across all users and items. In response, we propose a self-attentive sequential recommender system with topic modeling-based category embedding as a novel approach to exploit global information in the process of sequential recommendation. Our self-attention module effectively leverages the sequential patterns from the user's recent history. In addition, our novel category embedding approach, which utilizes the information computed by topic modeling, efficiently captures global information that the user generally prefers. Furthermore, to provide diverse recommendations as well as to prevent overfitting, our model also incorporates a vector obtained by random sampling. Experimental studies show that our model outperforms state-of-the-art sequential recommendation models, and that category embedding effectively provides global preference information.

Recently, unsupervised exemplar-based image-to-image translation, conditioned on a given exemplar without the paired data, has accomplished substantial advancements. In order to transfer the information from an exemplar to an input image, existing methods often use a normalization technique, e.g., adaptive instance normalization, that controls the channel-wise statistics of an input activation map at a particular layer, such as the mean and the variance. Meanwhile, style transfer approaches similar task to image translation by nature, demonstrated superior performance by using the higher-order statistics such as covariance among channels in representing a style. In detail, it works via whitening (given a zero-mean input feature, transforming its covariance matrix into the identity). followed by coloring (changing the covariance matrix of the whitened feature to those of the style feature). However, applying this approach in image translation is computationally intensive and error-prone due to the expensive time complexity and its non-trivial backpropagation. In response, this paper proposes an end-to-end approach tailored for image translation that efficiently approximates this transformation with our novel regularization methods. We further extend our approach to a group-wise form for memory and time efficiency as well as image quality. Extensive qualitative and quantitative experiments demonstrate that our proposed method is fast, both in training and inference, and highly effective in reflecting the style of an exemplar.

Despite recent advancements in deep learning-based automatic colorization, they are still limited when it comes to few-shot learning. Existing models require a significant amount of training data. To tackle this issue, we present a novel memory-augmented colorization model MemoPainter that can produce high-quality colorization with limited data. In particular, our model is able to capture rare instances and successfully colorize them. Also, we propose a novel threshold triplet loss that enables unsupervised training of memory networks without the need for class labels. Experiments show that our model has superior quality in both few-shot and one-shot colorization tasks.

Fuxiang Chen, Seung-won Hwang, Jaegul Choo, Jung-Woo Ha, Sunghun Kim. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.

Abstract The majority of visualizations on the web are still stored as raster images, making them inaccessible to visually impaired users. We propose a deep‐neural‐network‐based approach that automatically recognizes key elements in a visualization, including a visualization type, graphical elements, labels, legends, and most importantly, the original data conveyed in the visualization. We leverage such extracted information to provide visually impaired people with the reading of the extracted information. Based on interviews with visually impaired users, we built a Google Chrome extension designed to work with screen reader software to automatically decode charts on a webpage using our pipeline. We compared the performance of the back‐end algorithm with existing methods and evaluated the utility using qualitative feedback from visually impaired users.

We have recently seen many successful applications of recurrent neural networks (RNNs) on electronic medical records (EMRs), which contain histories of patients' diagnoses, medications, and other various events, in order to predict the current and future states of patients. Despite the strong performance of RNNs, it is often challenging for users to understand why the model makes a particular prediction. Such black-box nature of RNNs can impede its wide adoption in clinical practice. Furthermore, we have no established methods to interactively leverage users' domain expertise and prior knowledge as inputs for steering the model. Therefore, our design study aims to provide a visual analytics solution to increase interpretability and interactivity of RNNs via a joint effort of medical experts, artificial intelligence scientists, and visual analytics researchers. Following the iterative design process between the experts, we design, implement, and evaluate a visual analytics tool called RetainVis, which couples a newly improved, interpretable, and interactive RNN-based model called RetainEX and visualizations for users' exploration of EMR data in the context of prediction tasks. Our study shows the effective use of RetainVis for gaining insights into how individual medical codes contribute to making risk predictions, using EMRs of patients with heart failure and cataract symptoms. Our study also demonstrates how we made substantial changes to the state-of-the-art RNN model called RETAIN in order to make use of temporal information and increase interactivity. This study will provide a useful guideline for researchers that aim to design an interpretable and interactive visual analytics tool for RNNs.

Detecting anomalous events of a particular area in a timely manner is an important task. Geo-tagged social media data are useful resource for this task; however, the abundance of everyday language in them makes this task still challenging. To address such challenges, we present TopicOnTiles, a visual analytics system that can reveal information relevant to anomalous events in a multi-level tile-based map interface by using social media data. To this end, we adopt and improve a recently proposed topic modeling method that can extract spatio-temporally exclusive topics corresponding to a particular region and a time point. Furthermore, we utilize a tile-based map interface to efficiently handle large-scale data in parallel. Our user interface effectively highlights anomalous tiles using our novel glyph visualization that encodes the degree of anomaly computed by our exclusive topic modeling processes. To show the effectiveness of our system, we present several usage scenarios using real-world datasets as well as comprehensive user study results.

Recent studies have shown remarkable success in image-to-image translation for two domains. However, existing approaches have limited scalability and robustness in handling more than two domains, since different models should be built independently for every pair of image domains. To address this limitation, we propose StarGAN, a novel and scalable approach that can perform image-to-image translations for multiple domains using only a single model. Such a unified model architecture of StarGAN allows simultaneous training of multiple datasets with different domains within a single network. This leads to StarGAN's superior quality of translated images compared to existing models as well as the novel capability of flexibly translating an input image to any desired target domain. We empirically demonstrate the effectiveness of our approach on a facial attribute transfer and a facial expression synthesis tasks.

Machine reading comprehension helps machines learn to utilize most of the human knowledge written in the form of text.Existing approaches made a significant progress comparable to human-level performance, but they are still limited in understanding, up to a few paragraphs, failing to properly comprehend lengthy document.In this paper, we propose a novel deep neural network architecture to handle a long-range dependency in RC tasks.In detail, our method has two novel aspects: (1) an advanced memory-augmented architecture and (2) an expanded gated recurrent unit with dense connections that mitigate potential information distortion occurring in the memory.Our proposed architecture is widely applicable to other models.We have performed extensive experiments with well-known benchmark datasets such as TriviaQA, QUASAR-T, and SQuAD.The experimental results demonstrate that the proposed method outperforms existing methods, especially for lengthy documents.

It is well-known that GANs are difficult to train, and several different techniques have been proposed in order to stabilize their training. In this paper, we propose a novel training method called manifold-matching, and a new GAN model called manifold-matching GAN (MMGAN). MMGAN finds two manifolds representing the vector representations of real and fake images. If these two manifolds match, it means that real and fake images are statistically identical. To assist the manifold-matching task, we also use i) kernel tricks to find better manifold structures, ii) moving-averaged manifolds across mini-batches, and iii) a regularizer based on correlation matrix to suppress mode collapse. We conduct in-depth experiments with three image datasets and compare with several state-of-the-art GAN models. 32.4% of images generated by the proposed MMGAN are recognized as fake images during our user study (16% enhancement compared to other state-of-the-art model). MMGAN achieved an unsupervised inception score of 7.8 for CIFAR-10.

Recently, generative adversarial networks (GANs) have shown promising performance in generating realistic images. However, they often struggle in learning complex underlying modalities in a given dataset, resulting in poor-quality generated images. To mitigate this problem, we present a novel approach called mixture of experts GAN (MEGAN), an ensemble approach of multiple generator networks. Each generator network in MEGAN specializes in generating images with a particular subset of modalities, e.g., an image class. Instead of incorporating a separate step of handcrafted clustering of multiple modalities, our proposed model is trained through an end-to-end learning of multiple generators via gating networks, which is responsible for choosing the appropriate generator network for a given condition. We adopt the categorical reparameterization trick for a categorical decision to be made in selecting a generator while maintaining the flow of the gradients. We demonstrate that individual generators learn different and salient subparts of the data and achieve a multiscale structural similarity (MS-SSIM) score of 0.2470 for CelebA and a competitive unsupervised inception score of 8.33 in CIFAR-10.

Being a prevalent form of social communications on the Internet, billions of short texts are generated everyday. Discovering knowledge from them has gained a lot of interest from both industry and academia. The short texts have a limited contextual information, and they are sparse, noisy and ambiguous, and hence, automatically learning topics from them remains an important challenge. To tackle this problem, in this paper, we propose a semantics-assisted non-negative matrix factorization (SeaNMF) model to discover topics for the short texts. It effectively incorporates the word-context semantic correlations into the model, where the semantic relationships between the words and their contexts are learned from the skip-gram view of the corpus. The SeaNMF model is solved using a block coordinate descent algorithm. We also develop a sparse variant of the SeaNMF model which can achieve a better model interpretability. Extensive quantitative evaluations on various real-world short text datasets demonstrate the superior performance of the proposed models over several other state-of-the-art methods in terms of topic coherence and classification accuracy. The qualitative semantic analysis demonstrates the interpretability of our models by discovering meaningful and consistent topics. With a simple formulation and the superior performance, SeaNMF can be an effective standard topic model for short texts.

Abstract Embedding and visualizing large‐scale high‐dimensional data in a two‐dimensional space is an important problem, because such visualization can reveal deep insights of complex data. However, most of the existing embedding approaches run on an excessively high precision, even when users want to obtain a brief insight from a visualization of large‐scale datasets, ignoring the fact that in the end, the outputs are embedded onto a fixed‐range pixel‐based screen space. Motivated by this observation and directly considering the properties of screen space in an embedding algorithm, we propose Pixel‐Aligned Stochastic Neighbor Embedding (PixelSNE), a highly efficient screen resolution‐driven 2D embedding method which accelerates Barnes‐Hut tree‐based t‐distributed stochastic neighbor embedding (BH‐SNE), which is known to be a state‐of‐the‐art 2D embedding method. Our experimental results show a significantly faster running time for PixelSNE compared to BH‐SNE for various datasets while maintaining comparable embedding quality.

In this article, we present an interactive visual information retrieval and recommendation system, called VisIRR, for large-scale document discovery. VisIRR effectively combines the paradigms of (1) a passive pull through query processes for retrieval and (2) an active push that recommends items of potential interest to users based on their preferences. Equipped with an efficient dynamic query interface against a large-scale corpus, VisIRR organizes the retrieved documents into high-level topics and visualizes them in a 2D space, representing the relationships among the topics along with their keyword summary. In addition, based on interactive personalized preference feedback with regard to documents, VisIRR provides document recommendations from the entire corpus, which are beyond the retrieved sets. Such recommended documents are visualized in the same space as the retrieved documents, so that users can seamlessly analyze both existing and newly recommended ones. This article presents novel computational methods, which make these integrated representations and fast interactions possible for a large-scale document corpus. We illustrate how the system works by providing detailed usage scenarios. Additionally, we present preliminary user study results for evaluating the effectiveness of the system.

Central to many text analysis methods is the notion of a concept: a set of semantically related keywords characterizing a specific object, phenomenon, or theme. Advances in word embedding allow building a concept from a small set of seed terms. However, naive application of such techniques may result in false positive errors because of the polysemy of natural language. To mitigate this problem, we present a visual analytics system called ConceptVector that guides a user in building such concepts and then using them to analyze documents. Document-analysis case studies with real-world datasets demonstrate the fine-grained analysis provided by ConceptVector. To support the elaborate modeling of concepts, we introduce a bipolar concept model and support for specifying irrelevant words. We validate the interactive lexicon building interface by a user study and expert reviews. Quantitative evaluation shows that the bipolar lexicon generated with our methods is comparable to human-generated ones.

In this paper, we list the goals for and the pros and cons of guidance, and we discuss the role that it can play not only in key low-level visualization tasks but also the more sophisticated model-generation tasks of visual analytics. Recent advances in artificial intelligence, particularly in machine learning, have led to high hopes regarding the possibilities of using automatic techniques to perform some of the tasks that are currently done manually using visualization by data analysts. However, visual analytics remains a complex activity, combining many different subtasks. Some of these tasks are relatively low-level, and it is clear how automation could play a role—for example, classification and clustering of data. Other tasks are much more abstract and require significant human creativity, for example, linking insights gleaned from a variety of disparate and heterogeneous data artifacts to build support for decision making. In this paper, we outline the potential applications of guidance, as well as the inputs to guidance. We discuss challenges in implementing guidance, including the inputs to guidance systems and how to provide guidance to users. We propose potential methods for evaluating the quality of guidance at different phases in the analytic process and introduce the potential negative effects of guidance as a source of bias in analytic decision making.

One of the key advantages of visual analytics is its capability to leverage both humans's visual perception and the power of computing. A big obstacle in integrating machine learning with visual analytics is its high computing cost. To tackle this problem, this paper presents PIVE (Per-Iteration Visualization Environment) that supports real-time interactive visualization with machine learning. By immediately visualizing the intermediate results from algorithm iterations, PIVE enables users to quickly grasp insights and interact with the intermediate output, which then affects subsequent algorithm iterations. In addition, we propose a widely-applicable interaction methodology that allows efficient incorporation of user feedback into virtually any iterative computational method without introducing additional computational cost. We demonstrate the application of PIVE for various dimension reduction algorithms such as multidimensional scaling and t-SNE and clustering and topic modeling algorithms such as k-means and latent Dirichlet allocation.

Understanding newly emerging events or topics associated with a particular region of a given day can provide deep insight on the critical events occurring in highly evolving metropolitan cities. We propose herein a novel topic modeling approach on text documents with spatio-temporal information (e.g., when and where a document was published) such as location-based social media data to discover prevalent topics or newly emerging events with respect to an area and a time point. We consider a map view composed of regular grids or tiles with each showing topic keywords from documents of the corresponding region. To this end, we present a tilebased spatio-temporally exclusive topic modeling approach called STExNMF, based on a novel nonnegative matrix factorization (NMF) technique. STExNMF mainly works based on the two following stages: (1) first running a standard NMF of each tile to obtain general topics of the tile and (2) running a spatiotemporally exclusive NMF on a weighted residual matrix. These topics likely reveal information on newly emerging events or topics of interest within a region. We demonstrate the advantages of our approach using the geo-tagged Twitter data of New York City. We also provide quantitative comparisons in terms of the topic quality, spatio-temporal exclusiveness, topic variation, and qualitative evaluations of our method using several usage scenarios. In addition, we present a fast topic modeling technique of our model by leveraging parallel computing.

Detecting buffer overruns from a source code is one of the most common and yet challenging tasks in program analysis. Current approaches based on rigid rules and handcrafted features are limited in terms of flexible applicability and robustness due to diverse bug patterns and characteristics existing in sophisticated real-world software programs. In this paper, we propose a novel, data-driven approach that is completely end-to-end without requiring any hand-crafted features, thus free from any program language-specific structural limitations. In particular, our approach leverages a recently proposed neural network model called memory networks that have shown the state-of-the-art performances mainly in question-answering tasks. Our experimental results using source code samples demonstrate that our proposed model is capable of accurately detecting different types of buffer overruns. We also present in-depth analyses on how a memory network can learn to understand the semantics in programming languages solely from raw source codes, such as tracing variables of interest, identifying numerical values, and performing their quantitative comparisons.

Nonnegative matrix factorization (NMF) has been increasingly popular for topic modeling of large-scale documents. However, the resulting topics often represent only general, thus redundant information about the data rather than minor, but potentially meaningful information to users. To tackle this problem, we propose a novel ensemble model of nonnegative matrix factorization for discovering high-quality local topics. Our method leverages the idea of an ensemble model to successively perform NMF given a residual matrix obtained from previous stages and generates a sequence of topic sets. The novelty of our method lies in the fact that it utilizes the residual matrix inspired by a state-of-the-art gradient boosting model and applies a sophisticated local weighting scheme on the given matrix to enhance the locality of topics, which in turn delivers high-quality, focused topics of interest to users.

Topic modeling, which reveals underlying topics of a document corpus, has been actively adopted in visual analytics for large-scale document collections. However, due to its significant processing time and non-interactive nature, topic modeling has so far not been tightly integrated into a visual analytics workflow. Instead, most such systems are limited to utilizing a fixed, initial set of topics. Motivated by this gap in the literature, we propose a novel interaction technique called TopicLens that allows a user to dynamically explore data through a lens interface where topic modeling and the corresponding 2D embedding are efficiently computed on the fly. To support this interaction in real time while maintaining view consistency, we propose a novel efficient topic modeling method and a semi-supervised 2D embedding algorithm. Our work is based on improving state-of-the-art methods such as nonnegative matrix factorization and t-distributed stochastic neighbor embedding. Furthermore, we have built a web-based visual analytics system integrated with TopicLens. We use this system to measure the performance and the visualization quality of our proposed methods. We provide several scenarios showcasing the capability of TopicLens using real-world datasets.

Visual analytics techniques help users explore high-dimensional data. However, it is often challenging for users to express their domain knowledge in order to steer the underlying data model, especially when they have little attribute-level knowledge. Furthermore, users' complex, high-level domain knowledge, compared to low-level attributes, posits even greater challenges. To overcome these challenges, we introduce a technique to interpret a user's drawings with an interactive, nonlinear axis mapping approach called AxiSketcher. This technique enables users to impose their domain knowledge on a visualization by allowing interaction with data entries rather than with data attributes. The proposed interaction is performed through directly sketching lines over the visualization. Using this technique, users can draw lines over selected data points, and the system forms the axes that represent a nonlinear, weighted combination of multidimensional attributes. In this paper, we describe our techniques in three areas: 1) the design space of sketching methods for eliciting users' nonlinear domain knowledge; 2) the underlying model that translates users' input, extracts patterns behind the selected data points, and results in nonlinear axes reflecting users' complex intent; and 3) the interactive visualization for viewing, assessing, and reconstructing the newly formed, nonlinear axes.

Nonnegative matrix factorization (NMF) has been widely applied in many domains. In document analysis, it has been increasingly used in topic modeling applications, where a set of underlying topics are revealed by a low-rank factor matrix from NMF. However, it is often the case that the resulting topics give only general topic information in the data, which tends not to convey much information. To tackle this problem, we propose a novel ensemble model of nonnegative matrix factorization for discovering high-quality local topics. Our method leverages the idea of an ensemble model, which has been successful in supervised learning, into an unsupervised topic modeling context. That is, our model successively performs NMF given a residual matrix obtained from previous stages and generates a sequence of topic sets. Our algorithm for updating the input matrix has novelty in two aspects. The first lies in utilizing the residual matrix inspired by a state-of-the-art gradient boosting model, and the second stems from applying a sophisticated local weighting scheme on the given matrix to enhance the locality of topics, which in turn delivers high-quality, focused topics of interest to users. We evaluate our proposed method by comparing it against other topic modeling methods, such as a few variants of NMF and latent Dirichlet allocation, in terms of various evaluation measures representing topic coherence, diversity, coverage, computing time, and so on. We also present qualitative evaluation on the topics discovered by our method using several real-world data sets.

Scatterplots are effective visualization techniques for multidimensional data that use two (or three) axes to visualize data items as a point at its corresponding x and y Cartesian coordinates. Typically, each axis is bound to a single data attribute. Interactive exploration occurs by changing the data attributes bound to each of these axes. In the case of using scatterplots to visualize the outputs of dimension reduction techniques, the x and y axes are combinations of the true, high-dimensional data. For these spatializations, the axes present usability challenges in terms of interpretability and interactivity. That is, understanding the axes and interacting with them to make adjustments can be challenging. In this paper, we present InterAxis, a visual analytics technique to properly interpret, define, and change an axis in a user-driven manner. Users are given the ability to define and modify axes by dragging data items to either side of the x or y axes. from which the system computes a linear combination of data attributes and binds it to the axis. Further, users can directly tune the positive and negative contribution to these complex axes by using the visualization of data attributes that correspond to each axis. We describe the details of our technique and demonstrate the intended usage through two scenarios.

Through online health communities (OHCs), patients and caregivers exchange their illness experiences and strategies for overcoming the illness, and provide emotional support. To facilitate healthy and lively conversations in these communities, their members should be continuously monitored and nurtured by OHC administrators. The main challenge of OHC administrators' tasks lies in understanding the diverse dimensions of conversation threads that lead to productive discussions in their communities. In this paper, we present a design study in which three domain expert groups participated, an OHC researcher and two OHC administrators of online health communities, which was conducted to find with a visual analytic solution. Through our design study, we characterized the domain goals of OHC administrators and derived tasks to achieve these goals. As a result of this study, we propose a system called VisOHC, which visualizes individual OHC conversation threads as collapsed boxes-a visual metaphor of conversation threads. In addition, we augmented the posters' reply authorship network with marks and/or beams to show conversation dynamics within threads. We also developed unique measures tailored to the characteristics of OHCs, which can be encoded for thread visualizations at the users' requests. Our observation of the two administrators while using VisOHC showed that it supports their tasks and reveals interesting insights into online health communities. Finally, we share our methodological lessons on probing visual designs together with domain experts by allowing them to freely encode measurements into visual variables.

Analyzing text documents has been a key research topic in many areas. Countless approaches have been proposed to tackle this problem, and they are largely categorized into fully automated approaches (via statistical techniques) or human-involved exploratory ones (via interactive visualization). The primary purpose of this workshop is to bring together researchers from both sides and provide them with opportunities to discuss ways to harmonize the power of these two complementary approaches. The combination will allow us to push the boundary of text analytics. The detailed workshop schedule, proceedings, and agenda will be available at http://www.textvis.org.

Understanding large-scale document collections in an efficient manner is an important problem. Usually, document data are associated with other information (e.g., an author's gender, age, and location) and their links to other entities (e.g., co-authorship and citation networks). For the analysis of such data, we often have to reveal common as well as discriminative characteristics of documents with respect to their associated information, e.g., male- vs. female-authored documents, old vs. new documents, etc. To address such needs, this paper presents a novel topic modeling method based on joint nonnegative matrix factorization, which simultaneously discovers common as well as discriminative topics given multiple document sets. Our approach is based on a block-coordinate descent framework and is capable of utilizing only the most representative, thus meaningful, keywords in each topic through a novel pseudo-deflation approach. We perform both quantitative and qualitative evaluations using synthetic as well as real-world document data sets such as research paper collections and nonprofit micro-finance data. We show our method has a great potential for providing in-depth analyses by clearly identifying common and discriminative topics among multiple document sets.

Non-profit Micro-finance organizations provide loaning opportunities to eradicate poverty by financially equipping impoverished, yet skilled entrepreneurs who are in desperate need of an institution that lends to those who have little. Kiva.org, a widely-used crowd-funded micro-financial service, provides researchers with an extensive amount of publicly available data containing a rich set of heterogeneous information regarding micro-financial transactions. Our objective in this paper is to identify the key factors that encourage people to make micro-financing donations, and ultimately, to keep them actively involved. In our contribution to further promote a healthy micro-finance ecosystem, we detail our personalized loan recommendation system which we formulate as a supervised learning problem where we try to predict how likely a given lender will fund a new loan. We construct the features for each data item by utilizing the available connectivity relationships in order to integrate all the available Kiva data sources. For those lenders with no such relationships, e.g., first-time lenders, we propose a novel method of feature construction by computing joint nonnegative matrix factorizations. Utilizing gradient boosting tree methods, a state-of-the-art prediction model, we are able to achieve up to 0.92 AUC (area under the curve) value, which shows the potential of our methods for practical deployment. Finally, we point out several interesting phenomena on lenders' social behaviors in micro-finance activities.

Micro-finance organizations provide non-profit lending opportunities to mitigate poverty by financially supporting impoverished, yet skilled entrepreneurs who are in desperate need of an institution that lends to them. In Kiva.org, a widely-used crowd-funded micro-financial service, a vast amount of micro-financial activities are done by lending teams, and thus, understanding their diverse characteristics is crucial in maintaining a healthy micro-finance ecosystem. As the first step for this goal, we model different lending teams by using a maximum-entropy distribution approach based on a wealthy set of heterogeneous information regarding micro-financial transactions available at Kiva. Based on this approach, we achieved a competitive performance in predicting the lending activities for the top 200 teams. Furthermore, we provide deep insight about the characteristics of lending teams by analyzing the resulting team-specific lending models. We found that lending teams are generally more careful in selecting loans by a loan's geo-location, a borrower's gender, a field partner's reliability, etc., when compared to lenders without team affiliations. In addition, we identified interesting lending behaviors of different lending teams based on lenders' background and interest such as their ethnic, religious, linguistic, educational, regional, and occupational aspects. Finally, using our proposed model, we tackled a novel problem of lending team recommendation and showed its promising performance results.

Topic modeling has been widely used for analyzing text document collections. Recently, there have been significant advancements in various topic modeling techniques, particularly in the form of probabilistic graphical modeling. State-of-the-art techniques such as Latent Dirichlet Allocation (LDA) have been successfully applied in visual text analytics. However, most of the widely-used methods based on probabilistic modeling have drawbacks in terms of consistency from multiple runs and empirical convergence. Furthermore, due to the complicatedness in the formulation and the algorithm, LDA cannot easily incorporate various types of user feedback. To tackle this problem, we propose a reliable and flexible visual analytics system for topic modeling called UTOPIAN (User-driven Topic modeling based on Interactive Nonnegative Matrix Factorization). Centered around its semi-supervised formulation, UTOPIAN enables users to interact with the topic modeling method and steer the result in a user-driven manner. We demonstrate the capability of UTOPIAN via several usage scenarios with real-world document corpuses such as InfoVis/VAST paper data set and product review data sets.

Investigators across many disciplines and organizations must sift through large collections of text documents to understand and piece together information. Whether they are fighting crime, curing diseases, deciding what car to buy, or researching a new field, inevitably investigators will encounter text documents. Taking a visual analytics approach, we integrate multiple text analysis algorithms with a suite of interactive visualizations to provide a flexible and powerful environment that allows analysts to explore collections of documents while sensemaking. Our particular focus is on the process of integrating automated analyses with interactive visualizations in a smooth and fluid manner. We illustrate this integration through two example scenarios: an academic researcher examining InfoVis and VAST conference papers and a consumer exploring car reviews while pondering a purchase decision. Finally, we provide lessons learned toward the design and implementation of visual analytics systems for document exploration and understanding.

Heterogeneous data sets are typically represented in different feature spaces, making it difficult to analyze relationships spanning different data sets even when they are semantically related. Data fusion via space alignment can remedy this task by integrating multiple data sets lying in different spaces into one common space. Given a set of reference correspondence data that share the same semantic meaning across different spaces, space alignment attempts to place the corresponding reference data as close together as possible, and accordingly, the entire data are aligned in a common space. Space alignment involves optimizing two potentially conflicting criteria: minimum deformation of the original relationships and maximum alignment between the different spaces. To solve this problem, we provide a novel graph embedding framework for space alignment, which converts each data set into a graph and assigns zero distance between reference correspondence pairs resulting in a single graph. We propose a graph embedding method for fusion based on nonmetric multidimensional scaling (MDS). Its criteria using the rank order rather than the distance allows nonmetric MDS to effectively handle both deformation and alignment. Experiments using parallel data sets demonstrate that our approach works well in comparison to existing methods such as constrained Laplacian eigenmaps, Procrustes analysis, and tensor decomposition. We also present standard cross-domain information retrieval tests as well as interesting visualization examples using space alignment.

Abstract Clustering plays an important role in many large‐scale data analyses providing users with an overall understanding of their data. Nonetheless, clustering is not an easy task due to noisy features and outliers existing in the data, and thus the clustering results obtained from automatic algorithms often do not make clear sense. To remedy this problem, automatic clustering should be complemented with interactive visualization strategies. This paper proposes an interactive visual analytics system for document clustering, called iVisClustering, based on a widely‐used topic modeling method, latent Dirichlet allocation (LDA). iVisClustering provides a summary of each cluster in terms of its most representative keywords and visualizes soft clustering results in parallel coordinates. The main view of the system provides a 2D plot that visualizes cluster similarities and the relation among data items with a graph‐based representation. iVisClustering provides several other views, which contain useful interaction methods. With help of these visualization modules, we can interactively refine the clustering results in various ways. Keywords can be adjusted so that they characterize each cluster better. In addition, our system can filter out noisy data and re‐cluster the data accordingly. Cluster hierarchy can be constructed using a tree structure and for this purpose, the system supports cluster‐level interactions such as sub‐clustering, removing unimportant clusters, merging the clusters that have similar meanings, and moving certain clusters to any other node in the tree structure. Furthermore, the system provides document‐level interactions such as moving mis‐clustered documents to another cluster and removing useless documents. Finally, we present how interactive clustering is performed via iVisClustering by using real‐world document data sets.

One of the most widely-used nonlinear data embedding methods is ISOMAP. Based on a manifold learning framework, ISOMAP has a parameter k or ∈ that controls how many edges a neighborhood graph has. However, a suitable parameter value is often difficult to determine because of a time-consuming optimization process based on certain criteria, which may not be clearly justified. When ISOMAP is used to visualize data, users might want to test different parameter values in order to gain various insights about data, but such interaction between humans and such visualizations requires reasonably efficient updating, even for large-scale data. To tackle these problems, we propose an efficient updating algorithm for ISOMAP with parameter changes, called p-ISOMAP. We present not only a complexity analysis but also an empirical running time comparison, which show the advantage of p-ISOMAP. We also show interesting visualization applications of p-ISOMAP and demonstrate how to discover various characteristics of data through visualizations using different parameter values.

This paper demonstrates the applicability of the recently proposed supervised dimension reduction, hierarchical linear discriminant analysis (h-LDA) to a well-known spatial localization technique in signal processing, beamforming. The main motivation of h-LDA is to overcome the drawback of LDA that each cluster is modeled as a unimodal Gaussian distribution. For this purpose, h-LDA extends the variance decomposition in LDA to the subcluster level, and modifies the definition of the within-cluster scatter matrix. In this paper, we present an efficient h-LDA algorithm for over-sampled data, where the data dimension is larger than the dimension of the data vectors. The new algorithm utilizes the Cholesky decomposition based on a generalized singular value decomposition framework. Furthermore, we analyze the data model of h-LDA by relating it to the two-way multivariate analysis of variance (MANOVA), which fits well within the context of beamforming applications. Although beamforming has been generally dealt with as a regression problem, we propose a novel way of viewing beamforming as a classification problem, and apply a supervised dimension reduction, which allows the classifier to achieve better accuracy. Our experimental results show that h-LDA outperforms several dimension reduction methods such as LDA and kernel discriminant analysis, and regression approaches such as the regularized least squares and kernelized support vector regression.

Linear discriminant analysis (LDA) is a widely-used feature extraction method in classification. However, the original LDA has limitations due to the assumption of a unimodal structure for each cluster, which is satisfied in many applications such as facial image data when variations such as angle and illumination can significantly influence the images of the same person. In this paper, we propose a novel method, hierarchical LDA(h-LDA), which takes into account hierarchical subcluster structures in the data sets. Our experiments show that regularized h-LDA produces better accuracy than LDA, PCA, and tensorFaces.

Distance preserving dimension reduction (DPDR) using the singular value decomposition has recently been introduced. In this paper, for disease diagnosis using gene or protein expression data, we present empirical comparison results between DPDR and other various dimension reduction (DR) methods (i.e. PC A, MDS, Isomap, and LLE) when using support vector machines with radial basis function kernel. Our results show that DPDR outperforms, as a whole, other DR methods in terms of classification accuracy, but at the same time, it gives significant efficiency compared with other methods since it has no parameter to be optimized. Based on these empirical results, we reach a promising conclusion that DPDR is one of the best DR methods at hand for modeling an efficient and distortion- free classifier for gene or protein expression data.