Source author record

Shuicheng Yan

Shuicheng Yan appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Catalog footprint

What is connected

91works

21topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

Anisotropic Modality Align

Training multimodal large language models has long been limited by the scarcity of high-quality paired multimodal data. Recent studies show that the shared representation space of pretrained multimodal contrastive models can serve as a bridge, enabling models to perform multimodal training with unimodal data. However, the key premise of this paradigm remains insufficiently understood: can representations from different modalities be reliably interchanged? The core obstacle lies in the persistent Modality Gap in the shared space. In this work, we revisit the geometric nature of the modality gap. We find that modality representations already share compatible dominant semantic geometry. What truly hinders modality interchangeability is not a simple global shift, but an anisotropic residual structure concentrated along a small number of dominant directions. Based on this finding, we further propose the principle of anisotropic modality gap alignment: effective modality alignment should align with the target-modality distribution while preserving the semantic structure of the source modality. Guided by this principle, we propose an anisotropic geometric correction framework, AnisoAlign, for unpaired modality alignment. This framework leverages the internal geometric prior of the target modality and performs bounded correction on source-modality representations, thereby constructing substitute representations in the target modality. Experiments confirm its benefits in both geometric diagnostics and text-only MLLM training. Overall, this work recasts the modality gap from an empirical observation into a correctable, structured geometric phenomenon and provides a new representation alignment perspective for training multimodal models with unimodal data.

preprint2026arXiv

EvoRoute: Experience-Driven Self-Routing LLM Agent Systems

Complex agentic AI systems, powered by a coordinated ensemble of Large Language Models (LLMs), tool and memory modules, have demonstrated remarkable capabilities on intricate, multi-turn tasks. However, this success is shadowed by prohibitive economic costs and severe latency, exposing a critical, yet underexplored, trade-off. We formalize this challenge as the \textbf{Agent System Trilemma}: the inherent tension among achieving state-of-the-art performance, minimizing monetary cost, and ensuring rapid task completion. To dismantle this trilemma, we introduce EvoRoute, a self-evolving model routing paradigm that transcends static, pre-defined model assignments. Leveraging an ever-expanding knowledge base of prior experience, EvoRoute dynamically selects Pareto-optimal LLM backbones at each step, balancing accuracy, efficiency, and resource use, while continually refining its own selection policy through environment feedback. Experiments on challenging agentic benchmarks such as GAIA and BrowseComp+ demonstrate that EvoRoute, when integrated into off-the-shelf agentic systems, not only sustains or enhances system performance but also reduces execution cost by up to $80\%$ and latency by over $70\%$.

preprint2026arXiv

JavisGPT: A Unified Multi-modal LLM for Sounding-Video Comprehension and Generation

This paper presents JavisGPT, the first unified multimodal large language model (MLLM) for joint audio-video (JAV) comprehension and generation. JavisGPT has a concise encoder-LLM-decoder architecture, which has a SyncFusion module for spatio-temporal audio-video fusion and synchrony-aware learnable queries to bridge a pretrained JAV-DiT generator. This design enables temporally coherent video-audio understanding and generation from multimodal instructions. We design an effective three-stage training pipeline consisting of multimodal pretraining, audio-video fine-tuning, and large-scale instruction-tuning, to progressively build multimodal comprehension and generation from existing vision-language models. For instruction tuning, we construct JavisInst-Omni, a high-quality instruction dataset with over 200K GPT-4o-curated audio-video-text dialogues that cover diverse and multi-level comprehension and generation scenarios. On JAV comprehension and generation benchmarks, our experiments show that JavisGPT outperforms existing MLLMs, particularly in complex and temporally synchronized settings.

preprint2026arXiv

Mega-ASR: Towards In-the-wild^2 Speech Recognition via Scaling up Real-world Acoustic Simulation

Despite rapid advances in automatic speech recognition (ASR) and large audio-language models, robust recognition in real-world environments remains limited by an "acoustic robustness bottleneck": models often lose acoustic grounding and produce omissions or hallucinations under severe, compositional distortions. We propose Mega-ASR, a unified ASR-in-the-wild framework that combines scalable compound-data construction with progressive acoustic-to-semantic optimization. We introduce Voices-in-the-Wild-2M, covering 7 classic acoustic phenomena and 54 physically plausible compound scenarios, and train Mega-ASR with Acoustic-to-Semantic Progressive Supervised Fine-Tuning and Dual-Granularity WER-Gated Policy Optimization. Extensive experiments demonstrate that Mega-ASR achieves significant advantages over prior state-of-the-art systems on adverse-condition ASR benchmarks (45.69% vs. 54.01% on VOiCES R4-B-F, and 21.49% vs. 29.34% on NOIZEUS Sta-0). On complex compositional acoustic scenarios, Mega-ASR further delivers over 30% relative WER reduction against strong open- and closed-source baselines, establishing a scalable paradigm for robust ASR in-the-wild.

preprint2026arXiv

Mem-W: Latent Memory-Native GUI Agents

GUI agents are beginning to operate the web, mobile, and desktop as interactive worlds, where successful control depends on carrying forward visual, procedural, and task-level evidence beyond the fleeting present screen. Yet most agents still treat memory as an external, human-readable artifact: histories are summarized, categorized, retrieved, and reinserted as text or structured records before being encoded again by the policy. This creates a mismatch between the representational form in which experience is stored and the latent embedding sequence over which modern GUI policies actually act. We introduce Mem-W, a series of latent-memory-native GUI agents that treat memory as part of the agent's continuous context rather than as an auxiliary symbolic scaffold. Mem-W weaves both historical trajectories (as experiential memory) and in-session segments (as working memory) into compact memory tokens through a shared trajectory-to-latent compressor. These tokens are woven with the current GUI observation and local context into one continuous embedding sequence, allowing the agent to read successes, failures, and unfinished progress through the same machine-native interface. Mem-W is trained with self-distillation and outcome-aware supervision to preserve decision-relevant state while filtering memory toward evidence that truly supports task success. Across four web and mobile navigation benchmarks, Mem-W consistently improves diverse backbones and memory-enhanced baselines, with gains of up to $+30.0$, suggesting that latent-context-native memory can serve as a scalable foundation for long-horizon GUI agency.

preprint2026arXiv

Memory in the Age of AI Agents

Memory has emerged, and will continue to remain, a core capability of foundation model-based agents. As research on agent memory rapidly expands and attracts unprecedented attention, the field has also become increasingly fragmented. Existing works that fall under the umbrella of agent memory often differ substantially in their motivations, implementations, and evaluation protocols, while the proliferation of loosely defined memory terminologies has further obscured conceptual clarity. Traditional taxonomies such as long/short-term memory have proven insufficient to capture the diversity of contemporary agent memory systems. This work aims to provide an up-to-date landscape of current agent memory research. We begin by clearly delineating the scope of agent memory and distinguishing it from related concepts such as LLM memory, retrieval augmented generation (RAG), and context engineering. We then examine agent memory through the unified lenses of forms, functions, and dynamics. From the perspective of forms, we identify three dominant realizations of agent memory, namely token-level, parametric, and latent memory. From the perspective of functions, we propose a finer-grained taxonomy that distinguishes factual, experiential, and working memory. From the perspective of dynamics, we analyze how memory is formed, evolved, and retrieved over time. To support practical development, we compile a comprehensive summary of memory benchmarks and open-source frameworks. Beyond consolidation, we articulate a forward-looking perspective on emerging research frontiers, including memory automation, reinforcement learning integration, multimodal memory, multi-agent memory, and trustworthiness issues. We hope this survey serves not only as a reference for existing work, but also as a conceptual foundation for rethinking memory as a first-class primitive in the design of future agentic intelligence.

preprint2026arXiv

Tracking the Truth: Object-Centric Spatio-Temporal Monitoring for Video Large Language Models

While multimodal large language models (MLLMs) have advanced video understanding, they remain highly prone to hallucinations in dynamic scenes. We argue this stems from a failure in spatio-temporal monitoring, the ability to persistently track object identities, states, and relations over time. Existing benchmarks obscure this deficit by relying on single final-answer evaluations for queries that can often be resolved via local visual cues or statistical priors. To rigorously diagnose this, we introduce STEMO-Bench (Spatio-TEmporal MOnitoring), a benchmark of human-verified object-centric facts that evaluates intermediate reasoning by decomposing queries into sub-questions, distinguishing genuine temporal understanding from coincidental correctness. To address failure modes exposed by STEMO, we propose STEMO-Track, a novel object-centric framework that explicitly constructs and reasons over structured object trajectories via chunk-wise state extraction and temporal aggregation. Extensive experiments demonstrate that our object-centric framework significantly reduces hallucinated answers and improves spatio-temporal reasoning consistency over state-of-the-art MLLMs.

preprint2026arXiv

UniF$^2$ace: A Unified Fine-grained Face Understanding and Generation Model

Unified multimodal models (UMMs) have emerged as a powerful paradigm in fundamental cross-modality research, demonstrating significant potential in both image understanding and generation. However, existing research in the face domain primarily faces two challenges: $\textbf{(1)}$ $\textbf{fragmentation development}$, with existing methods failing to unify understanding and generation into a single one, hindering the way to artificial general intelligence. $\textbf{(2) lack of fine-grained facial attributes}$, which are crucial for high-fidelity applications. To handle those issues, we propose $\textbf{UniF$^2$ace}$, $\textit{the first UMM specifically tailored for fine-grained face understanding and generation}$. $\textbf{First}$, we introduce a novel theoretical framework with a Dual Discrete Diffusion (D3Diff) loss, unifying masked generative models with discrete score matching diffusion and leading to a more precise approximation of the negative log-likelihood. Moreover, this D3Diff significantly enhances the model's ability to synthesize high-fidelity facial details aligned with text input. $\textbf{Second}$, we propose a multi-level grouped Mixture-of-Experts architecture, adaptively incorporating the semantic and identity facial embeddings to complement the attribute forgotten phenomenon in representation evolvement. $\textbf{Finally}$, to this end, we construct UniF$^2$aceD-1M, a large-scale dataset comprising 130K fine-grained image-caption pairs and 1M visual question-answering pairs, spanning a much wider range of facial attributes than existing datasets. Extensive experiments demonstrate that UniF$^2$ace outperforms existing models with a similar scale in both understanding and generation tasks, with 7.1\% higher Desc-GPT and 6.6\% higher VQA-score, respectively.

preprint2026arXiv

WARD: Adversarially Robust Defense of Web Agents Against Prompt Injections

Web agents can autonomously complete online tasks by interacting with websites, but their exposure to open web environments makes them vulnerable to prompt injection attacks embedded in HTML content or visual interfaces. Existing guard models still suffer from limited generalization to unseen domains and attack patterns, high false positive rates on benign content, reduced deployment efficiency due to added latency at each step, and vulnerability to adversarial attacks that evolve over time or directly target the guard itself. To address these limitations, we propose WARD (Web Agent Robust Defense against Prompt Injection), a practical guard model for secure and efficient web agents. WARD is built on WARD-Base, a large-scale dataset with around 177K samples collected from 719 high-traffic URLs and platforms, and WARD-PIG, a dedicated dataset designed for prompt injection attacks targeting the guard model. We further introduce A3T, an adaptive adversarial attack training framework that iteratively strengthens WARD through a memory-based attacker and guard co-evolution process. Extensive experiments show that WARD achieves nearly perfect recall on out-of-distribution benchmarks, maintains low false positive rates to preserve agent utility, remains robust against guard-targeted and adaptive attacks under substantial distribution shifts, and runs efficiently in parallel with the agent without introducing additional latency.

preprint2024arXiv

Spikformer V2: Join the High Accuracy Club on ImageNet with an SNN Ticket

Spiking Neural Networks (SNNs), known for their biologically plausible architecture, face the challenge of limited performance. The self-attention mechanism, which is the cornerstone of the high-performance Transformer and also a biologically inspired structure, is absent in existing SNNs. To this end, we explore the potential of leveraging both self-attention capability and biological properties of SNNs, and propose a novel Spiking Self-Attention (SSA) and Spiking Transformer (Spikformer). The SSA mechanism eliminates the need for softmax and captures the sparse visual feature employing spike-based Query, Key, and Value. This sparse computation without multiplication makes SSA efficient and energy-saving. Further, we develop a Spiking Convolutional Stem (SCS) with supplementary convolutional layers to enhance the architecture of Spikformer. The Spikformer enhanced with the SCS is referred to as Spikformer V2. To train larger and deeper Spikformer V2, we introduce a pioneering exploration of Self-Supervised Learning (SSL) within the SNN. Specifically, we pre-train Spikformer V2 with masking and reconstruction style inspired by the mainstream self-supervised Transformer, and then finetune the Spikformer V2 on the image classification on ImageNet. Extensive experiments show that Spikformer V2 outperforms other previous surrogate training and ANN2SNN methods. An 8-layer Spikformer V2 achieves an accuracy of 80.38% using 4 time steps, and after SSL, a 172M 16-layer Spikformer V2 reaches an accuracy of 81.10% with just 1 time step. To the best of our knowledge, this is the first time that the SNN achieves 80+% accuracy on ImageNet. The code will be available at Spikformer V2.

preprint2022arXiv

$O(N^2)$ Universal Antisymmetry in Fermionic Neural Networks

Fermionic neural network (FermiNet) is a recently proposed wavefunction Ansatz, which is used in variational Monte Carlo (VMC) methods to solve the many-electron Schrödinger equation. FermiNet proposes permutation-equivariant architectures, on which a Slater determinant is applied to induce antisymmetry. FermiNet is proved to have universal approximation capability with a single determinant, namely, it suffices to represent any antisymmetric function given sufficient parameters. However, the asymptotic computational bottleneck comes from the Slater determinant, which scales with $O(N^3)$ for $N$ electrons. In this paper, we substitute the Slater determinant with a pairwise antisymmetry construction, which is easy to implement and can reduce the computational cost to $O(N^2)$. We formally prove that the pairwise construction built upon permutation-equivariant architectures can universally represent any antisymmetric function. Besides, this universality can be achieved via continuous approximators when we aim to represent ground-state wavefunctions.

preprint2022arXiv

Geometry-Guided Progressive NeRF for Generalizable and Efficient Neural Human Rendering

In this work we develop a generalizable and efficient Neural Radiance Field (NeRF) pipeline for high-fidelity free-viewpoint human body synthesis under settings with sparse camera views. Though existing NeRF-based methods can synthesize rather realistic details for human body, they tend to produce poor results when the input has self-occlusion, especially for unseen humans under sparse views. Moreover, these methods often require a large number of sampling points for rendering, which leads to low efficiency and limits their real-world applicability. To address these challenges, we propose a Geometry-guided Progressive NeRF (GP-NeRF). In particular, to better tackle self-occlusion, we devise a geometry-guided multi-view feature integration approach that utilizes the estimated geometry prior to integrate the incomplete information from input views and construct a complete geometry volume for the target human body. Meanwhile, for achieving higher rendering efficiency, we introduce a progressive rendering pipeline through geometry guidance, which leverages the geometric feature volume and the predicted density values to progressively reduce the number of sampling points and speed up the rendering process. Experiments on the ZJU-MoCap and THUman datasets show that our method outperforms the state-of-the-arts significantly across multiple generalization settings, while the time cost is reduced > 70% via applying our efficient progressive rendering pipeline.

preprint2022arXiv

Improving Vision Transformers by Revisiting High-frequency Components

The transformer models have shown promising effectiveness in dealing with various vision tasks. However, compared with training Convolutional Neural Network (CNN) models, training Vision Transformer (ViT) models is more difficult and relies on the large-scale training set. To explain this observation we make a hypothesis that \textit{ViT models are less effective in capturing the high-frequency components of images than CNN models}, and verify it by a frequency analysis. Inspired by this finding, we first investigate the effects of existing techniques for improving ViT models from a new frequency perspective, and find that the success of some techniques (e.g., RandAugment) can be attributed to the better usage of the high-frequency components. Then, to compensate for this insufficient ability of ViT models, we propose HAT, which directly augments high-frequency components of images via adversarial training. We show that HAT can consistently boost the performance of various ViT models (e.g., +1.2% for ViT-B, +0.5% for Swin-B), and especially enhance the advanced model VOLO-D5 to 87.3% that only uses ImageNet-1K data, and the superiority can also be maintained on out-of-distribution data and transferred to downstream tasks. The code is available at: https://github.com/jiawangbai/HAT.

preprint2022arXiv

Inception Transformer

Recent studies show that Transformer has strong capability of building long-range dependencies, yet is incompetent in capturing high frequencies that predominantly convey local information. To tackle this issue, we present a novel and general-purpose Inception Transformer, or iFormer for short, that effectively learns comprehensive features with both high- and low-frequency information in visual data. Specifically, we design an Inception mixer to explicitly graft the advantages of convolution and max-pooling for capturing the high-frequency information to Transformers. Different from recent hybrid frameworks, the Inception mixer brings greater efficiency through a channel splitting mechanism to adopt parallel convolution/max-pooling path and self-attention path as high- and low-frequency mixers, while having the flexibility to model discriminative information scattered within a wide frequency range. Considering that bottom layers play more roles in capturing high-frequency details while top layers more in modeling low-frequency global information, we further introduce a frequency ramp structure, i.e. gradually decreasing the dimensions fed to the high-frequency mixer and increasing those to the low-frequency mixer, which can effectively trade-off high- and low-frequency components across different layers. We benchmark the iFormer on a series of vision tasks, and showcase that it achieves impressive performance on image classification, COCO detection and ADE20K segmentation. For example, our iFormer-S hits the top-1 accuracy of 83.4% on ImageNet-1K, much higher than DeiT-S by 3.6%, and even slightly better than much bigger model Swin-B (83.3%) with only 1/4 parameters and 1/3 FLOPs. Code and models will be released at https://github.com/sail-sg/iFormer.

preprint2022arXiv

MetaFormer Is Actually What You Need for Vision

Transformers have shown great potential in computer vision tasks. A common belief is their attention-based token mixer module contributes most to their competence. However, recent works show the attention-based module in Transformers can be replaced by spatial MLPs and the resulted models still perform quite well. Based on this observation, we hypothesize that the general architecture of the Transformers, instead of the specific token mixer module, is more essential to the model's performance. To verify this, we deliberately replace the attention module in Transformers with an embarrassingly simple spatial pooling operator to conduct only basic token mixing. Surprisingly, we observe that the derived model, termed as PoolFormer, achieves competitive performance on multiple computer vision tasks. For example, on ImageNet-1K, PoolFormer achieves 82.1% top-1 accuracy, surpassing well-tuned Vision Transformer/MLP-like baselines DeiT-B/ResMLP-B24 by 0.3%/1.1% accuracy with 35%/52% fewer parameters and 50%/62% fewer MACs. The effectiveness of PoolFormer verifies our hypothesis and urges us to initiate the concept of "MetaFormer", a general architecture abstracted from Transformers without specifying the token mixer. Based on the extensive experiments, we argue that MetaFormer is the key player in achieving superior results for recent Transformer and MLP-like models on vision tasks. This work calls for more future research dedicated to improving MetaFormer instead of focusing on the token mixer modules. Additionally, our proposed PoolFormer could serve as a starting baseline for future MetaFormer architecture design. Code is available at https://github.com/sail-sg/poolformer.

preprint2022arXiv

Modern Augmented Reality: Applications, Trends, and Future Directions

Augmented reality (AR) is one of the relatively old, yet trending areas in the intersection of computer vision and computer graphics with numerous applications in several areas, from gaming and entertainment, to education and healthcare. Although it has been around for nearly fifty years, it has seen a lot of interest by the research community in the recent years, mainly because of the huge success of deep learning models for various computer vision and AR applications, which made creating new generations of AR technologies possible. This work tries to provide an overview of modern augmented reality, from both application-level and technical perspective. We first give an overview of main AR applications, grouped into more than ten categories. We then give an overview of around 100 recent promising machine learning based works developed for AR systems, such as deep learning works for AR shopping (clothing, makeup), AR based image filters (such as Snapchat's lenses), AR animations, and more. In the end we discuss about some of the current challenges in AR domain, and the future directions in this area.

preprint2022arXiv

Mugs: A Multi-Granular Self-Supervised Learning Framework

In self-supervised learning, multi-granular features are heavily desired though rarely investigated, as different downstream tasks (e.g., general and fine-grained classification) often require different or multi-granular features, e.g.~fine- or coarse-grained one or their mixture. In this work, for the first time, we propose an effective MUlti-Granular Self-supervised learning (Mugs) framework to explicitly learn multi-granular visual features. Mugs has three complementary granular supervisions: 1) an instance discrimination supervision (IDS), 2) a novel local-group discrimination supervision (LGDS), and 3) a group discrimination supervision (GDS). IDS distinguishes different instances to learn instance-level fine-grained features. LGDS aggregates features of an image and its neighbors into a local-group feature, and pulls local-group features from different crops of the same image together and push them away for others. It provides complementary instance supervision to IDS via an extra alignment on local neighbors, and scatters different local-groups separately to increase discriminability. Accordingly, it helps learn high-level fine-grained features at a local-group level. Finally, to prevent similar local-groups from being scattered randomly or far away, GDS brings similar samples close and thus pulls similar local-groups together, capturing coarse-grained features at a (semantic) group level. Consequently, Mugs can capture three granular features that often enjoy higher generality on diverse downstream tasks over single-granular features, e.g.~instance-level fine-grained features in contrastive learning. By only pretraining on ImageNet-1K, Mugs sets new SoTA linear probing accuracy 82.1$\%$ on ImageNet-1K and improves previous SoTA by $1.1\%$. It also surpasses SoTAs on other tasks, e.g. transfer learning, detection and segmentation.

preprint2022arXiv

PnP-DETR: Towards Efficient Visual Analysis with Transformers

Recently, DETR pioneered the solution of vision tasks with transformers, it directly translates the image feature map into the object detection result. Though effective, translating the full feature map can be costly due to redundant computation on some area like the background. In this work, we encapsulate the idea of reducing spatial redundancy into a novel poll and pool (PnP) sampling module, with which we build an end-to-end PnP-DETR architecture that adaptively allocates its computation spatially to be more efficient. Concretely, the PnP module abstracts the image feature map into fine foreground object feature vectors and a small number of coarse background contextual feature vectors. The transformer models information interaction within the fine-coarse feature space and translates the features into the detection result. Moreover, the PnP-augmented model can instantly achieve various desired trade-offs between performance and computation with a single model by varying the sampled feature length, without requiring to train multiple models as existing methods. Thus it offers greater flexibility for deployment in diverse scenarios with varying computation constraint. We further validate the generalizability of the PnP module on panoptic segmentation and the recent transformer-based image recognition model ViT and show consistent efficiency gain. We believe our method makes a step for efficient visual analysis with transformers, wherein spatial redundancy is commonly observed. Code will be available at \url{https://github.com/twangnh/pnp-detr}.

preprint2022arXiv

Robustness and Accuracy Could Be Reconcilable by (Proper) Definition

The trade-off between robustness and accuracy has been widely studied in the adversarial literature. Although still controversial, the prevailing view is that this trade-off is inherent, either empirically or theoretically. Thus, we dig for the origin of this trade-off in adversarial training and find that it may stem from the improperly defined robust error, which imposes an inductive bias of local invariance -- an overcorrection towards smoothness. Given this, we advocate employing local equivariance to describe the ideal behavior of a robust model, leading to a self-consistent robust error named SCORE. By definition, SCORE facilitates the reconciliation between robustness and accuracy, while still handling the worst-case uncertainty via robust optimization. By simply substituting KL divergence with variants of distance metrics, SCORE can be efficiently minimized. Empirically, our models achieve top-rank performance on RobustBench under AutoAttack. Besides, SCORE provides instructive insights for explaining the overfitting phenomenon and semantic input gradients observed on robust models. Code is available at https://github.com/P2333/SCORE.

preprint2022arXiv

Self-Promoted Supervision for Few-Shot Transformer

The few-shot learning ability of vision transformers (ViTs) is rarely investigated though heavily desired. In this work, we empirically find that with the same few-shot learning frameworks, \eg~Meta-Baseline, replacing the widely used CNN feature extractor with a ViT model often severely impairs few-shot classification performance. Moreover, our empirical study shows that in the absence of inductive bias, ViTs often learn the low-qualified token dependencies under few-shot learning regime where only a few labeled training data are available, which largely contributes to the above performance degradation. To alleviate this issue, for the first time, we propose a simple yet effective few-shot training framework for ViTs, namely Self-promoted sUpervisioN (SUN). Specifically, besides the conventional global supervision for global semantic learning SUN further pretrains the ViT on the few-shot learning dataset and then uses it to generate individual location-specific supervision for guiding each patch token. This location-specific supervision tells the ViT which patch tokens are similar or dissimilar and thus accelerates token dependency learning. Moreover, it models the local semantics in each patch token to improve the object grounding and recognition capability which helps learn generalizable patterns. To improve the quality of location-specific supervision, we further propose two techniques:~1) background patch filtration to filtrate background patches out and assign them into an extra background class; and 2) spatial-consistent augmentation to introduce sufficient diversity for data augmentation while keeping the accuracy of the generated local supervisions. Experimental results show that SUN using ViTs significantly surpasses other few-shot learning frameworks with ViTs and is the first one that achieves higher performance than those CNN state-of-the-arts.

preprint2022arXiv

Value-Consistent Representation Learning for Data-Efficient Reinforcement Learning

Deep reinforcement learning (RL) algorithms suffer severe performance degradation when the interaction data is scarce, which limits their real-world application. Recently, visual representation learning has been shown to be effective and promising for boosting sample efficiency in RL. These methods usually rely on contrastive learning and data augmentation to train a transition model for state prediction, which is different from how the model is used in RL--performing value-based planning. Accordingly, the learned representation by these visual methods may be good for recognition but not optimal for estimating state value and solving the decision problem. To address this issue, we propose a novel method, called value-consistent representation learning (VCR), to learn representations that are directly related to decision-making. More specifically, VCR trains a model to predict the future state (also referred to as the ''imagined state'') based on the current one and a sequence of actions. Instead of aligning this imagined state with a real state returned by the environment, VCR applies a $Q$-value head on both states and obtains two distributions of action values. Then a distance is computed and minimized to force the imagined state to produce a similar action value prediction as that by the real state. We develop two implementations of the above idea for the discrete and continuous action spaces respectively. We conduct experiments on Atari 100K and DeepMind Control Suite benchmarks to validate their effectiveness for improving sample efficiency. It has been demonstrated that our methods achieve new state-of-the-art performance for search-free RL algorithms.

preprint2022arXiv

Video Graph Transformer for Video Question Answering

This paper proposes a Video Graph Transformer (VGT) model for Video Quetion Answering (VideoQA). VGT's uniqueness are two-fold: 1) it designs a dynamic graph transformer module which encodes video by explicitly capturing the visual objects, their relations, and dynamics for complex spatio-temporal reasoning; and 2) it exploits disentangled video and text Transformers for relevance comparison between the video and text to perform QA, instead of entangled cross-modal Transformer for answer classification. Vision-text communication is done by additional cross-modal interaction modules. With more reasonable video encoding and QA solution, we show that VGT can achieve much better performances on VideoQA tasks that challenge dynamic relation reasoning than prior arts in the pretraining-free scenario. Its performances even surpass those models that are pretrained with millions of external data. We further show that VGT can also benefit a lot from self-supervised cross-modal pretraining, yet with orders of magnitude smaller data. These results clearly demonstrate the effectiveness and superiority of VGT, and reveal its potential for more data-efficient pretraining. With comprehensive analyses and some heuristic observations, we hope that VGT can promote VQA research beyond coarse recognition/description towards fine-grained relation reasoning in realistic videos. Our code is available at https://github.com/sail-sg/VGT.

preprint2021arXiv

A Survey on Concept Factorization: From Shallow to Deep Representation Learning

The quality of learned features by representation learning determines the performance of learning algorithms and the related application tasks (such as high-dimensional data clustering). As a relatively new paradigm for representation learning, Concept Factorization (CF) has attracted a great deal of interests in the areas of machine learning and data mining for over a decade. Lots of effective CF based methods have been proposed based on different perspectives and properties, but note that it still remains not easy to grasp the essential connections and figure out the underlying explanatory factors from exiting studies. In this paper, we therefore survey the recent advances on CF methodologies and the potential benchmarks by categorizing and summarizing the current methods. Specifically, we first re-view the root CF method, and then explore the advancement of CF-based representation learning ranging from shallow to deep/multilayer cases. We also introduce the potential application areas of CF-based methods. Finally, we point out some future directions for studying the CF-based representation learning. Overall, this survey provides an insightful overview of both theoretical basis and current developments in the field of CF, which can also help the interested researchers to understand the current trends of CF and find the most appropriate CF techniques to deal with particular applications.

preprint2021arXiv

ConvBERT: Improving BERT with Span-based Dynamic Convolution

Pre-trained language models like BERT and its variants have recently achieved impressive performance in various natural language understanding tasks. However, BERT heavily relies on the global self-attention block and thus suffers large memory footprint and computation cost. Although all its attention heads query on the whole input sequence for generating the attention map from a global perspective, we observe some heads only need to learn local dependencies, which means the existence of computation redundancy. We therefore propose a novel span-based dynamic convolution to replace these self-attention heads to directly model local dependencies. The novel convolution heads, together with the rest self-attention heads, form a new mixed attention block that is more efficient at both global and local context learning. We equip BERT with this mixed attention design and build a ConvBERT model. Experiments have shown that ConvBERT significantly outperforms BERT and its variants in various downstream tasks, with lower training cost and fewer model parameters. Remarkably, ConvBERTbase model achieves 86.4 GLUE score, 0.7 higher than ELECTRAbase, while using less than 1/4 training cost. Code and pre-trained models will be released.

preprint2021arXiv

ORDNet: Capturing Omni-Range Dependencies for Scene Parsing

Learning to capture dependencies between spatial positions is essential to many visual tasks, especially the dense labeling problems like scene parsing. Existing methods can effectively capture long-range dependencies with self-attention mechanism while short ones by local convolution. However, there is still much gap between long-range and short-range dependencies, which largely reduces the models' flexibility in application to diverse spatial scales and relationships in complicated natural scene images. To fill such a gap, we develop a Middle-Range (MR) branch to capture middle-range dependencies by restricting self-attention into local patches. Also, we observe that the spatial regions which have large correlations with others can be emphasized to exploit long-range dependencies more accurately, and thus propose a Reweighed Long-Range (RLR) branch. Based on the proposed MR and RLR branches, we build an Omni-Range Dependencies Network (ORDNet) which can effectively capture short-, middle- and long-range dependencies. Our ORDNet is able to extract more comprehensive context information and well adapt to complex spatial variance in scene images. Extensive experiments show that our proposed ORDNet outperforms previous state-of-the-art methods on three scene parsing benchmarks including PASCAL Context, COCO Stuff and ADE20K, demonstrating the superiority of capturing omni-range dependencies in deep models for scene parsing task.

preprint2020arXiv

AdversarialNAS: Adversarial Neural Architecture Search for GANs

Neural Architecture Search (NAS) that aims to automate the procedure of architecture design has achieved promising results in many computer vision fields. In this paper, we propose an AdversarialNAS method specially tailored for Generative Adversarial Networks (GANs) to search for a superior generative model on the task of unconditional image generation. The AdversarialNAS is the first method that can search the architectures of generator and discriminator simultaneously in a differentiable manner. During searching, the designed adversarial search algorithm does not need to comput any extra metric to evaluate the performance of the searched architecture, and the search paradigm considers the relevance between the two network architectures and improves their mutual balance. Therefore, AdversarialNAS is very efficient and only takes 1 GPU day to search for a superior generative model in the proposed large search space ($10^{38}$). Experiments demonstrate the effectiveness and superiority of our method. The discovered generative model sets a new state-of-the-art FID score of $10.87$ and highly competitive Inception Score of $8.74$ on CIFAR-10. Its transferability is also proven by setting new state-of-the-art FID score of $26.98$ and Inception score of $9.63$ on STL-10. Code is at: \url{https://github.com/chengaopro/AdversarialNAS}.

preprint2020arXiv

Compressed DenseNet for Lightweight Character Recognition

Convolutional Recurrent Neural Network (CRNN) is a popular network for recognizing texts in images. Advances like the variant of CRNN, such as Dense Convolutional Network with Connectionist Temporal Classification, has reduced the running time of the network, but exposing the inner computation cost and weight size of the convolutional networks as a bottleneck. Specifically, the DenseNet based models utilize the dense blocks as the core module, but the inner features are combined in the form of concatenation in dense blocks. As such, the number of channels of combined features delivered as the input of the layers close to the output and the relevant computational cost grows rapidly with the dense blocks getting deeper. This will severely bring heavy computational cost and big weight size, which restrict the depth of dense blocks. In this paper, we propose a compressed convolution block called Lightweight Dense Block (LDB). To reduce the computing cost and weight size, we re-define and re-design the way of combining internal features of the dense blocks. LDB is a convolutional block similarly as dense block, but it can reduce the computation cost and weight size to (1/L, 2/L), compared with original ones, where L is the number of layers in blocks. Moreover, LDB can be used to replace the original dense block in any DenseNet based models. Based on the LDBs, we propose a Compressed DenseNet (CDenseNet) for the lightweight character recognition. Extensive experiments demonstrate that CDenseNet can effectively reduce the weight size while delivering the promising recognition results.

preprint2020arXiv

Convolutional Dictionary Pair Learning Network for Image Representation Learning

Both the Dictionary Learning (DL) and Convolutional Neural Networks (CNN) are powerful image representation learning systems based on different mechanisms and principles, however whether we can seamlessly integrate them to improve the per-formance is noteworthy exploring. To address this issue, we propose a novel generalized end-to-end representation learning architecture, dubbed Convolutional Dictionary Pair Learning Network (CDPL-Net) in this paper, which integrates the learning schemes of the CNN and dictionary pair learning into a unified framework. Generally, the architecture of CDPL-Net includes two convolutional/pooling layers and two dictionary pair learn-ing (DPL) layers in the representation learning module. Besides, it uses two fully-connected layers as the multi-layer perception layer in the nonlinear classification module. In particular, the DPL layer can jointly formulate the discriminative synthesis and analysis representations driven by minimizing the batch based reconstruction error over the flatted feature maps from the convolution/pooling layer. Moreover, DPL layer uses l1-norm on the analysis dictionary so that sparse representation can be delivered, and the embedding process will also be robust to noise. To speed up the training process of DPL layer, the efficient stochastic gradient descent is used. Extensive simulations on real databases show that our CDPL-Net can deliver enhanced performance over other state-of-the-art methods.

preprint2020arXiv

Dual Adversarial Auto-Encoders for Clustering

As a powerful approach for exploratory data analysis, unsupervised clustering is a fundamental task in computer vision and pattern recognition. Many clustering algorithms have been developed, but most of them perform unsatisfactorily on the data with complex structures. Recently, Adversarial Auto-Encoder (AAE) shows effectiveness on tackling such data by combining Auto-Encoder (AE) and adversarial training, but it cannot effectively extract classification information from the unlabeled data. In this work, we propose Dual Adversarial Auto-encoder (Dual-AAE) which simultaneously maximizes the likelihood function and mutual information between observed examples and a subset of latent variables. By performing variational inference on the objective function of Dual-AAE, we derive a new reconstruction loss which can be optimized by training a pair of Auto-encoders. Moreover, to avoid mode collapse, we introduce the clustering regularization term for the category variable. Experiments on four benchmarks show that Dual-AAE achieves superior performance over state-of-the-art clustering methods. Besides, by adding a reject option, the clustering accuracy of Dual-AAE can reach that of supervised CNN algorithms. Dual-AAE can also be used for disentangling style and content of images without using supervised information.

preprint2020arXiv

Highly Efficient Salient Object Detection with 100K Parameters

Salient object detection models often demand a considerable amount of computation cost to make precise prediction for each pixel, making them hardly applicable on low-power devices. In this paper, we aim to relieve the contradiction between computation cost and model performance by improving the network efficiency to a higher degree. We propose a flexible convolutional module, namely generalized OctConv (gOctConv), to efficiently utilize both in-stage and cross-stages multi-scale features, while reducing the representation redundancy by a novel dynamic weight decay scheme. The effective dynamic weight decay scheme stably boosts the sparsity of parameters during training, supports learnable number of channels for each scale in gOctConv, allowing 80% of parameters reduce with negligible performance drop. Utilizing gOctConv, we build an extremely light-weighted model, namely CSNet, which achieves comparable performance with about 0.2% parameters (100k) of large models on popular salient object detection benchmarks.

preprint2020arXiv

Learning Target Domain Specific Classifier for Partial Domain Adaptation

Unsupervised domain adaptation~(UDA) aims at reducing the distribution discrepancy when transferring knowledge from a labeled source domain to an unlabeled target domain. Previous UDA methods assume that the source and target domains share an identical label space, which is unrealistic in practice since the label information of the target domain is agnostic. This paper focuses on a more realistic UDA scenario, i.e. partial domain adaptation (PDA), where the target label space is subsumed to the source label space. In the PDA scenario, the source outliers that are absent in the target domain may be wrongly matched to the target domain (technically named negative transfer), leading to performance degradation of UDA methods. This paper proposes a novel Target Domain Specific Classifier Learning-based Domain Adaptation (TSCDA) method. TSCDA presents a soft-weighed maximum mean discrepancy criterion to partially align feature distributions and alleviate negative transfer. Also, it learns a target-specific classifier for the target domain with pseudo-labels and multiple auxiliary classifiers, to further address classifier shift. A module named Peers Assisted Learning is used to minimize the prediction difference between multiple target-specific classifiers, which makes the classifiers more discriminant for the target domain. Extensive experiments conducted on three PDA benchmark datasets show that TSCDA outperforms other state-of-the-art methods with a large margin, e.g. $4\%$ and $5.6\%$ averagely on Office-31 and Office-Home, respectively.

preprint2020arXiv

Multilayer Collaborative Low-Rank Coding Network for Robust Deep Subspace Discovery

For subspace recovery, most existing low-rank representation (LRR) models performs in the original space in single-layer mode. As such, the deep hierarchical information cannot be learned, which may result in inaccurate recoveries for complex real data. In this paper, we explore the deep multi-subspace recovery problem by designing a multilayer architecture for latent LRR. Technically, we propose a new Multilayer Collabora-tive Low-Rank Representation Network model termed DeepLRR to discover deep features and deep subspaces. In each layer (>2), DeepLRR bilinearly reconstructs the data matrix by the collabo-rative representation with low-rank coefficients and projection matrices in the previous layer. The bilinear low-rank reconstruc-tion of previous layer is directly fed into the next layer as the input and low-rank dictionary for representation learning, and is further decomposed into a deep principal feature part, a deep salient feature part and a deep sparse error. As such, the coher-ence issue can be also resolved due to the low-rank dictionary, and the robustness against noise can also be enhanced in the feature subspace. To recover the sparse errors in layers accurately, a dynamic growing strategy is used, as the noise level will be-come smaller for the increase of layers. Besides, a neighborhood reconstruction error is also included to encode the locality of deep salient features by deep coefficients adaptively in each layer. Extensive results on public databases show that our DeepLRR outperforms other related models for subspace discovery and clustering.

preprint2020arXiv

Recapture as You Want

With the increasing prevalence and more powerful camera systems of mobile devices, people can conveniently take photos in their daily life, which naturally brings the demand for more intelligent photo post-processing techniques, especially on those portrait photos. In this paper, we present a portrait recapture method enabling users to easily edit their portrait to desired posture/view, body figure and clothing style, which are very challenging to achieve since it requires to simultaneously perform non-rigid deformation of human body, invisible body-parts reasoning and semantic-aware editing. We decompose the editing procedure into semantic-aware geometric and appearance transformation. In geometric transformation, a semantic layout map is generated that meets user demands to represent part-level spatial constraints and further guides the semantic-aware appearance transformation. In appearance transformation, we design two novel modules, Semantic-aware Attentive Transfer (SAT) and Layout Graph Reasoning (LGR), to conduct intra-part transfer and inter-part reasoning, respectively. SAT module produces each human part by paying attention to the semantically consistent regions in the source portrait. It effectively addresses the non-rigid deformation issue and well preserves the intrinsic structure/appearance with rich texture details. LGR module utilizes body skeleton knowledge to construct a layout graph that connects all relevant part features, where graph reasoning mechanism is used to propagate information among part nodes to mine their relations. In this way, LGR module infers invisible body parts and guarantees global coherence among all the parts. Extensive experiments on DeepFashion, Market-1501 and in-the-wild photos demonstrate the effectiveness and superiority of our approach. Video demo is at: \url{https://youtu.be/vTyq9HL6jgw}.

preprint2019arXiv

RC-DARTS: Resource Constrained Differentiable Architecture Search

Recent advances show that Neural Architectural Search (NAS) method is able to find state-of-the-art image classification deep architectures. In this paper, we consider the one-shot NAS problem for resource constrained applications. This problem is of great interest because it is critical to choose different architectures according to task complexity when the resource is constrained. Previous techniques are either too slow for one-shot learning or does not take the resource constraint into consideration. In this paper, we propose the resource constrained differentiable architecture search (RC-DARTS) method to learn architectures that are significantly smaller and faster while achieving comparable accuracy. Specifically, we propose to formulate the RC-DARTS task as a constrained optimization problem by adding the resource constraint. An iterative projection method is proposed to solve the given constrained optimization problem. We also propose a multi-level search strategy to enable layers at different depths to adaptively learn different types of neural architectures. Through extensive experiments on the Cifar10 and ImageNet datasets, we show that the RC-DARTS method learns lightweight neural architectures which have smaller model size and lower computational complexity while achieving comparable or better performances than the state-of-the-art methods.

preprint2019arXiv

Very Long Natural Scenery Image Prediction by Outpainting

Comparing to image inpainting, image outpainting receives less attention due to two challenges in it. The first challenge is how to keep the spatial and content consistency between generated images and original input. The second challenge is how to maintain high quality in generated results, especially for multi-step generations in which generated regions are spatially far away from the initial input. To solve the two problems, we devise some innovative modules, named Skip Horizontal Connection and Recurrent Content Transfer, and integrate them into our designed encoder-decoder structure. By this design, our network can generate highly realistic outpainting prediction effectively and efficiently. Other than that, our method can generate new images with very long sizes while keeping the same style and semantic content as the given input. To test the effectiveness of the proposed architecture, we collect a new scenery dataset with diverse, complicated natural scenes. The experimental results on this dataset have demonstrated the efficacy of our proposed network. The code and dataset are available from https://github.com/z-x-yang/NS-Outpainting.

preprint2017arXiv

Peak-Piloted Deep Network for Facial Expression Recognition

Objective functions for training of deep networks for face-related recognition tasks, such as facial expression recognition (FER), usually consider each sample independently. In this work, we present a novel peak-piloted deep network (PPDN) that uses a sample with peak expression (easy sample) to supervise the intermediate feature responses for a sample of non-peak expression (hard sample) of the same type and from the same subject. The expression evolving process from non-peak expression to peak expression can thus be implicitly embedded in the network to achieve the invariance to expression intensities. A special purpose back-propagation procedure, peak gradient suppression (PGS), is proposed for network training. It drives the intermediate-layer feature responses of non-peak expression samples towards those of the corresponding peak expression samples, while avoiding the inverse. This avoids degrading the recognition capability for samples of peak expression due to interference from their non-peak expression counterparts. Extensive comparisons on two popular FER datasets, Oulu-CASIA and CK+, demonstrate the superiority of the PPDN over state-ofthe-art FER methods, as well as the advantages of both the network structure and the optimization strategy. Moreover, it is shown that PPDN is a general architecture, extensible to other tasks by proper definition of peak and non-peak samples. This is validated by experiments that show state-of-the-art performance on pose-invariant face recognition, using the Multi-PIE dataset.

preprint2016arXiv

A Focused Dynamic Attention Model for Visual Question Answering

Visual Question and Answering (VQA) problems are attracting increasing interest from multiple research disciplines. Solving VQA problems requires techniques from both computer vision for understanding the visual contents of a presented image or video, as well as the ones from natural language processing for understanding semantics of the question and generating the answers. Regarding visual content modeling, most of existing VQA methods adopt the strategy of extracting global features from the image or video, which inevitably fails in capturing fine-grained information such as spatial configuration of multiple objects. Extracting features from auto-generated regions -- as some region-based image recognition methods do -- cannot essentially address this problem and may introduce some overwhelming irrelevant features with the question. In this work, we propose a novel Focused Dynamic Attention (FDA) model to provide better aligned image content representation with proposed questions. Being aware of the key words in the question, FDA employs off-the-shelf object detector to identify important regions and fuse the information from the regions and global features via an LSTM unit. Such question-driven representations are then combined with question representation and fed into a reasoning unit for generating the answers. Extensive evaluation on a large-scale benchmark dataset, VQA, clearly demonstrate the superior performance of FDA over well-established baselines.

preprint2016arXiv

A Unified Alternating Direction Method of Multipliers by Majorization Minimization

Accompanied with the rising popularity of compressed sensing, the Alternating Direction Method of Multipliers (ADMM) has become the most widely used solver for linearly constrained convex problems with separable objectives. In this work, we observe that many previous variants of ADMM update the primal variable by minimizing different majorant functions with their convergence proofs given case by case. Inspired by the principle of majorization minimization, we respectively present the unified frameworks and convergence analysis for the Gauss-Seidel ADMMs and Jacobian ADMMs, which use different historical information for the current updating. Our frameworks further generalize previous ADMMs to the ones capable of solving the problems with non-separable objectives by minimizing their separable majorant surrogates. We also show that the bound which measures the convergence speed of ADMMs depends on the tightness of the used majorant function. Then several techniques are introduced to improve the efficiency of ADMMs by tightening the majorant functions. In particular, we propose the Mixed Gauss-Seidel and Jacobian ADMM (M-ADMM) which alleviates the slow convergence issue of Jacobian ADMMs by absorbing merits of the Gauss-Seidel ADMMs. M-ADMM can be further improved by using backtracking, wise variable partition and fully exploiting the structure of the constraint. Beyond the guarantee in theory, numerical experiments on both synthesized and real-world data further demonstrate the superiority of our new ADMMs in practice. Finally, we release a toolbox at https://github.com/canyilu/LibADMM that implements efficient ADMMs for many problems in compressed sensing.

preprint2016arXiv

Attentive Contexts for Object Detection

Modern deep neural network based object detection methods typically classify candidate proposals using their interior features. However, global and local surrounding contexts that are believed to be valuable for object detection are not fully exploited by existing methods yet. In this work, we take a step towards understanding what is a robust practice to extract and utilize contextual information to facilitate object detection in practice. Specifically, we consider the following two questions: "how to identify useful global contextual information for detecting a certain object?" and "how to exploit local context surrounding a proposal for better inferring its contents?". We provide preliminary answers to these questions through developing a novel Attention to Context Convolution Neural Network (AC-CNN) based object detection model. AC-CNN effectively incorporates global and local contextual information into the region-based CNN (e.g. Fast RCNN) detection model and provides better object detection performance. It consists of one attention-based global contextualized (AGC) sub-network and one multi-scale local contextualized (MLC) sub-network. To capture global context, the AGC sub-network recurrently generates an attention map for an input image to highlight useful global contextual locations, through multiple stacked Long Short-Term Memory (LSTM) layers. For capturing surrounding local context, the MLC sub-network exploits both the inside and outside contextual information of each specific proposal at multiple scales. The global and local context are then fused together for making the final decision for detection. Extensive experiments on PASCAL VOC 2007 and VOC 2012 well demonstrate the superiority of the proposed AC-CNN over well-established baselines. In particular, AC-CNN outperforms the popular Fast-RCNN by 2.0% and 2.2% on VOC 2007 and VOC 2012 in terms of mAP, respectively.

preprint2016arXiv

Collaborative Layer-wise Discriminative Learning in Deep Neural Networks

Intermediate features at different layers of a deep neural network are known to be discriminative for visual patterns of different complexities. However, most existing works ignore such cross-layer heterogeneities when classifying samples of different complexities. For example, if a training sample has already been correctly classified at a specific layer with high confidence, we argue that it is unnecessary to enforce rest layers to classify this sample correctly and a better strategy is to encourage those layers to focus on other samples. In this paper, we propose a layer-wise discriminative learning method to enhance the discriminative capability of a deep network by allowing its layers to work collaboratively for classification. Towards this target, we introduce multiple classifiers on top of multiple layers. Each classifier not only tries to correctly classify the features from its input layer, but also coordinates with other classifiers to jointly maximize the final classification performance. Guided by the other companion classifiers, each classifier learns to concentrate on certain training examples and boosts the overall performance. Allowing for end-to-end training, our method can be conveniently embedded into state-of-the-art deep networks. Experiments with multiple popular deep networks, including Network in Network, GoogLeNet and VGGNet, on scale-various object classification benchmarks, including CIFAR100, MNIST and ImageNet, and scene classification benchmarks, including MIT67, SUN397 and Places205, demonstrate the effectiveness of our method. In addition, we also analyze the relationship between the proposed method and classical conditional random fields models.

preprint2016arXiv

Deep Recurrent Regression for Facial Landmark Detection

We propose a novel end-to-end deep architecture for face landmark detection, based on a deep convolutional and deconvolutional network followed by carefully designed recurrent network structures. The pipeline of this architecture consists of three parts. Through the first part, we encode an input face image to resolution-preserved deconvolutional feature maps via a deep network with stacked convolutional and deconvolutional layers. Then, in the second part, we estimate the initial coordinates of the facial key points by an additional convolutional layer on top of these deconvolutional feature maps. In the last part, by using the deconvolutional feature maps and the initial facial key points as input, we refine the coordinates of the facial key points by a recurrent network that consists of multiple Long-Short Term Memory (LSTM) components. Extensive evaluations on several benchmark datasets show that the proposed deep architecture has superior performance against the state-of-the-art methods.

preprint2016arXiv

Hybrid CNN and Dictionary-Based Models for Scene Recognition and Domain Adaptation

Convolutional neural network (CNN) has achieved state-of-the-art performance in many different visual tasks. Learned from a large-scale training dataset, CNN features are much more discriminative and accurate than the hand-crafted features. Moreover, CNN features are also transferable among different domains. On the other hand, traditional dictionarybased features (such as BoW and SPM) contain much more local discriminative and structural information, which is implicitly embedded in the images. To further improve the performance, in this paper, we propose to combine CNN with dictionarybased models for scene recognition and visual domain adaptation. Specifically, based on the well-tuned CNN models (e.g., AlexNet and VGG Net), two dictionary-based representations are further constructed, namely mid-level local representation (MLR) and convolutional Fisher vector representation (CFV). In MLR, an efficient two-stage clustering method, i.e., weighted spatial and feature space spectral clustering on the parts of a single image followed by clustering all representative parts of all images, is used to generate a class-mixture or a classspecific part dictionary. After that, the part dictionary is used to operate with the multi-scale image inputs for generating midlevel representation. In CFV, a multi-scale and scale-proportional GMM training strategy is utilized to generate Fisher vectors based on the last convolutional layer of CNN. By integrating the complementary information of MLR, CFV and the CNN features of the fully connected layer, the state-of-the-art performance can be achieved on scene recognition and domain adaptation problems. An interested finding is that our proposed hybrid representation (from VGG net trained on ImageNet) is also complementary with GoogLeNet and/or VGG-11 (trained on Place205) greatly.

preprint2016arXiv

Instance-Aware Hashing for Multi-Label Image Retrieval

Similarity-preserving hashing is a commonly used method for nearest neighbour search in large-scale image retrieval. For image retrieval, deep-networks-based hashing methods are appealing since they can simultaneously learn effective image representations and compact hash codes. This paper focuses on deep-networks-based hashing for multi-label images, each of which may contain objects of multiple categories. In most existing hashing methods, each image is represented by one piece of hash code, which is referred to as semantic hashing. This setting may be suboptimal for multi-label image retrieval. To solve this problem, we propose a deep architecture that learns \textbf{instance-aware} image representations for multi-label image data, which are organized in multiple groups, with each group containing the features for one category. The instance-aware representations not only bring advantages to semantic hashing, but also can be used in category-aware hashing, in which an image is represented by multiple pieces of hash codes and each piece of code corresponds to a category. Extensive evaluations conducted on several benchmark datasets demonstrate that, for both semantic hashing and category-aware hashing, the proposed method shows substantial improvement over the state-of-the-art supervised and unsupervised hashing methods.

preprint2016arXiv

Manipulated Object Proposal: A Discriminative Object Extraction and Feature Fusion Framework for First-Person Daily Activity Recognition

Detecting and recognizing objects interacting with humans lie in the center of first-person (egocentric) daily activity recognition. However, due to noisy camera motion and frequent changes in viewpoint and scale, most of the previous egocentric action recognition methods fail to capture and model highly discriminative object features. In this work, we propose a novel pipeline for first-person daily activity recognition, aiming at more discriminative object feature representation and object-motion feature fusion. Our object feature extraction and representation pipeline is inspired by the recent success of object hypotheses and deep convolutional neural network based detection frameworks. Our key contribution is a simple yet effective manipulated object proposal generation scheme. This scheme leverages motion cues such as motion boundary and motion magnitude (in contrast, camera motion is usually considered as "noise" for most previous methods) to generate a more compact and discriminative set of object proposals, which are more closely related to the objects which are being manipulated. Then, we learn more discriminative object detectors from these manipulated object proposals based on region-based convolutional neural network (R-CNN). Meanwhile, we develop a network based feature fusion scheme which better combines object and motion features. We show in experiments that the proposed framework significantly outperforms the state-of-the-art recognition performance on a challenging first-person daily activity benchmark.

preprint2016arXiv

Multi-Path Feedback Recurrent Neural Network for Scene Parsing

In this paper, we consider the scene parsing problem and propose a novel Multi-Path Feedback recurrent neural network (MPF-RNN) for parsing scene images. MPF-RNN can enhance the capability of RNNs in modeling long-range context information at multiple levels and better distinguish pixels that are easy to confuse. Different from feedforward CNNs and RNNs with only single feedback, MPF-RNN propagates the contextual features learned at top layer through \textit{multiple} weighted recurrent connections to learn bottom features. For better training MPF-RNN, we propose a new strategy that considers accumulative loss at multiple recurrent steps to improve performance of the MPF-RNN on parsing small objects. With these two novel components, MPF-RNN has achieved significant improvement over strong baselines (VGG16 and Res101) on five challenging scene parsing benchmarks, including traditional SiftFlow, Barcelona, CamVid, Stanford Background as well as the recently released large-scale ADE20K.

preprint2016arXiv

Multi-stage Object Detection with Group Recursive Learning

Most of existing detection pipelines treat object proposals independently and predict bounding box locations and classification scores over them separately. However, the important semantic and spatial layout correlations among proposals are often ignored, which are actually useful for more accurate object detection. In this work, we propose a new EM-like group recursive learning approach to iteratively refine object proposals by incorporating such context of surrounding proposals and provide an optimal spatial configuration of object detections. In addition, we propose to incorporate the weakly-supervised object segmentation cues and region-based object detection into a multi-stage architecture in order to fully exploit the learned segmentation features for better object detection in an end-to-end way. The proposed architecture consists of three cascaded networks which respectively learn to perform weakly-supervised object segmentation, object proposal generation and recursive detection refinement. Combining the group recursive learning and the multi-stage architecture provides competitive mAPs of 78.6% and 74.9% on the PASCAL VOC2007 and VOC2012 datasets respectively, which outperforms many well-established baselines [10] [20] significantly.

preprint2016arXiv

Robust LSTM-Autoencoders for Face De-Occlusion in the Wild

Face recognition techniques have been developed significantly in recent years. However, recognizing faces with partial occlusion is still challenging for existing face recognizers which is heavily desired in real-world applications concerning surveillance and security. Although much research effort has been devoted to developing face de-occlusion methods, most of them can only work well under constrained conditions, such as all the faces are from a pre-defined closed set. In this paper, we propose a robust LSTM-Autoencoders (RLA) model to effectively restore partially occluded faces even in the wild. The RLA model consists of two LSTM components, which aims at occlusion-robust face encoding and recurrent occlusion removal respectively. The first one, named multi-scale spatial LSTM encoder, reads facial patches of various scales sequentially to output a latent representation, and occlusion-robustness is achieved owing to the fact that the influence of occlusion is only upon some of the patches. Receiving the representation learned by the encoder, the LSTM decoder with a dual channel architecture reconstructs the overall face and detects occlusion simultaneously, and by feat of LSTM, the decoder breaks down the task of face de-occlusion into restoring the occluded part step by step. Moreover, to minimize identify information loss and guarantee face recognition accuracy over recovered faces, we introduce an identity-preserving adversarial training scheme to further improve RLA. Extensive experiments on both synthetic and real datasets of faces with occlusion clearly demonstrate the effectiveness of our proposed RLA in removing different types of facial occlusion at various locations. The proposed method also provides significantly larger performance gain than other de-occlusion methods in promoting recognition performance over partially-occluded faces.

preprint2016arXiv

Scale-aware Fast R-CNN for Pedestrian Detection

In this work, we consider the problem of pedestrian detection in natural scenes. Intuitively, instances of pedestrians with different spatial scales may exhibit dramatically different features. Thus, large variance in instance scales, which results in undesirable large intra-category variance in features, may severely hurt the performance of modern object instance detection methods. We argue that this issue can be substantially alleviated by the divide-and-conquer philosophy. Taking pedestrian detection as an example, we illustrate how we can leverage this philosophy to develop a Scale-Aware Fast R-CNN (SAF R-CNN) framework. The model introduces multiple built-in sub-networks which detect pedestrians with scales from disjoint ranges. Outputs from all the sub-networks are then adaptively combined to generate the final detection results that are shown to be robust to large variance in instance scales, via a gate function defined over the sizes of object proposals. Extensive evaluations on several challenging pedestrian detection datasets well demonstrate the effectiveness of the proposed SAF R-CNN. Particularly, our method achieves state-of-the-art performance on Caltech, INRIA, and ETH, and obtains competitive results on KITTI.

preprint2016arXiv

Scale-aware Pixel-wise Object Proposal Networks

Object proposal is essential for current state-of-the-art object detection pipelines. However, the existing proposal methods generally fail in producing results with satisfying localization accuracy. The case is even worse for small objects which however are quite common in practice. In this paper we propose a novel Scale-aware Pixel-wise Object Proposal (SPOP) network to tackle the challenges. The SPOP network can generate proposals with high recall rate and average best overlap (ABO), even for small objects. In particular, in order to improve the localization accuracy, a fully convolutional network is employed which predicts locations of object proposals for each pixel. The produced ensemble of pixel-wise object proposals enhances the chance of hitting the object significantly without incurring heavy extra computational cost. To solve the challenge of localizing objects at small scale, two localization networks which are specialized for localizing objects with different scales are introduced, following the divide-and-conquer philosophy. Location outputs of these two networks are then adaptively combined to generate the final proposals by a large-/small-size weighting network. Extensive evaluations on PASCAL VOC 2007 show the SPOP network is superior over the state-of-the-art models. The high-quality proposals from SPOP network also significantly improve the mean average precision (mAP) of object detection with Fast-RCNN framework. Finally, the SPOP network (trained on PASCAL VOC) shows great generalization performance when testing it on ILSVRC 2013 validation set.

preprint2016arXiv

Semantic Object Parsing with Graph LSTM

By taking the semantic object parsing task as an exemplar application scenario, we propose the Graph Long Short-Term Memory (Graph LSTM) network, which is the generalization of LSTM from sequential data or multi-dimensional data to general graph-structured data. Particularly, instead of evenly and fixedly dividing an image to pixels or patches in existing multi-dimensional LSTM structures (e.g., Row, Grid and Diagonal LSTMs), we take each arbitrary-shaped superpixel as a semantically consistent node, and adaptively construct an undirected graph for each image, where the spatial relations of the superpixels are naturally used as edges. Constructed on such an adaptive graph topology, the Graph LSTM is more naturally aligned with the visual patterns in the image (e.g., object boundaries or appearance similarities) and provides a more economical information propagation route. Furthermore, for each optimization step over Graph LSTM, we propose to use a confidence-driven scheme to update the hidden and memory states of nodes progressively till all nodes are updated. In addition, for each node, the forgets gates are adaptively learned to capture different degrees of semantic correlation with neighboring nodes. Comprehensive evaluations on four diverse semantic object parsing datasets well demonstrate the significant superiority of our Graph LSTM over other state-of-the-art solutions.

preprint2016arXiv

Seq-NMS for Video Object Detection

Video object detection is challenging because objects that are easily detected in one frame may be difficult to detect in another frame within the same clip. Recently, there have been major advances for doing object detection in a single image. These methods typically contain three phases: (i) object proposal generation (ii) object classification and (iii) post-processing. We propose a modification of the post-processing phase that uses high-scoring object detections from nearby frames to boost scores of weaker detections within the same clip. We show that our method obtains superior results to state-of-the-art single image object detection techniques. Our method placed 3rd in the video object detection (VID) task of the ImageNet Large Scale Visual Recognition Challenge 2015 (ILSVRC2015).

preprint2016arXiv

Training Skinny Deep Neural Networks with Iterative Hard Thresholding Methods

Deep neural networks have achieved remarkable success in a wide range of practical problems. However, due to the inherent large parameter space, deep models are notoriously prone to overfitting and difficult to be deployed in portable devices with limited memory. In this paper, we propose an iterative hard thresholding (IHT) approach to train Skinny Deep Neural Networks (SDNNs). An SDNN has much fewer parameters yet can achieve competitive or even better performance than its full CNN counterpart. More concretely, the IHT approach trains an SDNN through following two alternative phases: (I) perform hard thresholding to drop connections with small activations and fine-tune the other significant filters; (II)~re-activate the frozen connections and train the entire network to improve its overall discriminative capability. We verify the superiority of SDNNs in terms of efficiency and classification performance on four benchmark object recognition datasets, including CIFAR-10, CIFAR-100, MNIST and ImageNet. Experimental results clearly demonstrate that IHT can be applied for training SDNN based on various CNN architectures such as NIN and AlexNet.

preprint2016arXiv

Video Scene Parsing with Predictive Feature Learning

In this work, we address the challenging video scene parsing problem by developing effective representation learning methods given limited parsing annotations. In particular, we contribute two novel methods that constitute a unified parsing framework. (1) \textbf{Predictive feature learning}} from nearly unlimited unlabeled video data. Different from existing methods learning features from single frame parsing, we learn spatiotemporal discriminative features by enforcing a parsing network to predict future frames and their parsing maps (if available) given only historical frames. In this way, the network can effectively learn to capture video dynamics and temporal context, which are critical clues for video scene parsing, without requiring extra manual annotations. (2) \textbf{Prediction steering parsing}} architecture that effectively adapts the learned spatiotemporal features to scene parsing tasks and provides strong guidance for any off-the-shelf parsing model to achieve better video scene parsing performance. Extensive experiments over two challenging datasets, Cityscapes and Camvid, have demonstrated the effectiveness of our methods by showing significant improvement over well-established baselines.

preprint2016arXiv

Visual Processing by a Unified Schatten-$p$ Norm and $\ell_q$ Norm Regularized Principal Component Pursuit

In this paper, we propose a non-convex formulation to recover the authentic structure from the corrupted real data. Typically, the specific structure is assumed to be low rank, which holds for a wide range of data, such as images and videos. Meanwhile, the corruption is assumed to be sparse. In the literature, such a problem is known as Robust Principal Component Analysis (RPCA), which usually recovers the low rank structure by approximating the rank function with a nuclear norm and penalizing the error by an $\ell_1$-norm. Although RPCA is a convex formulation and can be solved effectively, the introduced norms are not tight approximations, which may cause the solution to deviate from the authentic one. Therefore, we consider here a non-convex relaxation, consisting of a Schatten-$p$ norm and an $\ell_q$-norm that promote low rank and sparsity respectively. We derive a proximal iteratively reweighted algorithm (PIRA) to solve the problem. Our algorithm is based on an alternating direction method of multipliers, where in each iteration we linearize the underlying objective function that allows us to have a closed form solution. We demonstrate that solutions produced by the linearized approximation always converge and have a tighter approximation than the convex counterpart. Experimental results on benchmarks show encouraging results of our approach.

Shuicheng Yan

What is connected

Connect this record

See the researcher in context

Building this map preview

91 published item(s)

Anisotropic Modality Align

EvoRoute: Experience-Driven Self-Routing LLM Agent Systems

JavisGPT: A Unified Multi-modal LLM for Sounding-Video Comprehension and Generation

Mega-ASR: Towards In-the-wild^2 Speech Recognition via Scaling up Real-world Acoustic Simulation

Mem-W: Latent Memory-Native GUI Agents

Memory in the Age of AI Agents

Tracking the Truth: Object-Centric Spatio-Temporal Monitoring for Video Large Language Models

UniF$^2$ace: A Unified Fine-grained Face Understanding and Generation Model

WARD: Adversarially Robust Defense of Web Agents Against Prompt Injections

Spikformer V2: Join the High Accuracy Club on ImageNet with an SNN Ticket

$O(N^2)$ Universal Antisymmetry in Fermionic Neural Networks

Geometry-Guided Progressive NeRF for Generalizable and Efficient Neural Human Rendering

Improving Vision Transformers by Revisiting High-frequency Components

Inception Transformer

MetaFormer Is Actually What You Need for Vision

Modern Augmented Reality: Applications, Trends, and Future Directions

Mugs: A Multi-Granular Self-Supervised Learning Framework

PnP-DETR: Towards Efficient Visual Analysis with Transformers

Robustness and Accuracy Could Be Reconcilable by (Proper) Definition

Self-Promoted Supervision for Few-Shot Transformer

Value-Consistent Representation Learning for Data-Efficient Reinforcement Learning

Video Graph Transformer for Video Question Answering

A Survey on Concept Factorization: From Shallow to Deep Representation Learning

ConvBERT: Improving BERT with Span-based Dynamic Convolution

ORDNet: Capturing Omni-Range Dependencies for Scene Parsing

AdversarialNAS: Adversarial Neural Architecture Search for GANs

Compressed DenseNet for Lightweight Character Recognition

Convolutional Dictionary Pair Learning Network for Image Representation Learning

Dual Adversarial Auto-Encoders for Clustering

Highly Efficient Salient Object Detection with 100K Parameters

Learning Target Domain Specific Classifier for Partial Domain Adaptation

Multilayer Collaborative Low-Rank Coding Network for Robust Deep Subspace Discovery

Recapture as You Want

RC-DARTS: Resource Constrained Differentiable Architecture Search

Very Long Natural Scenery Image Prediction by Outpainting

Peak-Piloted Deep Network for Facial Expression Recognition

A Focused Dynamic Attention Model for Visual Question Answering

A Unified Alternating Direction Method of Multipliers by Majorization Minimization

Attentive Contexts for Object Detection

Collaborative Layer-wise Discriminative Learning in Deep Neural Networks

Deep Recurrent Regression for Facial Landmark Detection

Hybrid CNN and Dictionary-Based Models for Scene Recognition and Domain Adaptation

Instance-Aware Hashing for Multi-Label Image Retrieval

Manipulated Object Proposal: A Discriminative Object Extraction and Feature Fusion Framework for First-Person Daily Activity Recognition

Multi-Path Feedback Recurrent Neural Network for Scene Parsing

Multi-stage Object Detection with Group Recursive Learning

Robust LSTM-Autoencoders for Face De-Occlusion in the Wild

Scale-aware Fast R-CNN for Pedestrian Detection

Scale-aware Pixel-wise Object Proposal Networks

Semantic Object Parsing with Graph LSTM

Seq-NMS for Video Object Detection

Training Skinny Deep Neural Networks with Iterative Hard Thresholding Methods

Video Scene Parsing with Predictive Feature Learning

Visual Processing by a Unified Schatten-$p$ Norm and $\ell_q$ Norm Regularized Principal Component Pursuit

Adaptive Nonparametric Image Parsing

Computational Baby Learning

Correlation Adaptive Subspace Segmentation by Trace Lasso

Correntropy Induced L2 Graph for Robust Subspace Clustering

Cross-domain Image Retrieval with a Dual Attribute-aware Ranking Network

Cross-pose Face Recognition by Canonical Correlation Analysis

Crowded Scene Analysis: A Survey

Deep Human Parsing with Active Template Regression

Fast Proximal Linearized Alternating Direction Method of Multiplier with Parallel Splitting

Group $K$-Means

Incorporating Structural Alternatives and Sharing into Hierarchy for Multiclass Object Recognition and Detection

Matching-CNN Meets KNN: Quasi-Parametric Human Parsing

Modality-dependent Cross-media Retrieval

Nonconvex Nonsmooth Low-Rank Minimization via Iteratively Reweighted Nuclear Norm

Personalized Age Progression with Aging Dictionary

Proposal-free Network for Instance-level Object Segmentation

Purine: A bi-graph based deep learning framework

Reversible Recursive Instance-level Object Segmentation

Semantic Object Parsing with Local-Global Long Short-Term Memory

CNN: Single-label to Multi-label