Author name cluster

Shuo Zhang

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

40 papers

2 author rows

AAAI Conference 2026 Conference Paper

A Theoretical Analysis of Detecting Large Model-Generated Time Series

Junji Hou
Junzhou Zhao
Shuo Zhang
Pinghui Wang

Motivated by the increasing risks of data misuse and fabrication, we investigate the problem of identifying synthetic time series generated by Time-Series Large Models (TSLMs) in this work. While there are extensive researches on detecting model generated text, we find that these existing methods are not applicable to time series data due to the fundamental modality difference, as time series usually have lower information density and smoother probability distributions than text data, which limit the discriminative power of token-based detectors. To address this issue, we examine the subtle distributional differences between real and model-generated time series and propose the contraction hypothesis, which states that model-generated time series, unlike real ones, exhibit progressively decreasing uncertainty under recursive forecasting. We formally prove this hypothesis under theoretical assumptions on model behavior and time series structure. Model-generated time series exhibit progressively concentrated distributions under recursive forecasting, leading to uncertainty contraction. We provide empirical validation of the hypothesis across diverse datasets. Building on this insight, we introduce the Uncertainty Contraction Estimator (UCE), a white-box detector that aggregates uncertainty metrics over successive prefixes to identify TSLM‑generated time series. Extensive experiments on 32 datasets show that UCE consistently outperforms state-of-the-art baselines, offering a reliable and generalizable solution for detecting model-generated time series.

PDF Details DOI

EAAI Journal 2026 Journal Article

A unified open-world semi-supervised learning framework for industrial defect detection via contrastive embedding and dynamic attention

Xiaoqing Zheng
Lixiang Zhou
Anke Xue
Shuo Zhang
Zhangping Chen
Yaguang Kong

Automated surface inspection in industrial manufacturing increasingly encounters novel defect classes that are unobserved during training, resulting in an open-world learning scenario that challenges the assumptions of conventional closed-set semi-supervised learning methods. To address this challenge, we propose a robust and effective Open-World Semi-Supervised Learning (OWSSL) framework specifically tailored for industrial surface defect classification. The proposed framework integrates a dual supervised and self-supervised contrastive learning strategy with dynamic attention adjustment mechanisms, all built upon a Vision Transformer (ViT) backbone. The contrastive learning strategy enforces both intra-class compactness and inter-class separability, which is essential for novel class discovery in open-world settings. In addition, a Dynamic Attention Adjustment (DAA) module is introduced to adaptively re-weight spatial features, enhancing the model’s focus on defect regions based on image content. To further improve generalization, a Teacher–Student network architecture with Exponential Moving Average (EMA) is incorporated, enabling the generation of consistent and reliable pseudo-labels for unlabeled data. Comprehensive experiments on public defect datasets, a private Copper Clad Laminate (CCL) dataset, and the Canadian Institute for Advanced Research ten-class dataset (CIFAR-10) benchmark demonstrate clear gains over strong baselines. The proposed method achieves 92. 46% accuracy on Copper Clad Laminate and 97. 10% on Marble, surpassing the strongest robust semi-supervised baseline by 2. 23 and 0. 73 points, respectively; and 97. 7% on unseen classes of CIFAR-10, improving over a strong open-world semi-supervised baseline by 8. 5 points.

EAAI Journal 2026 Journal Article

Contrast-enhanced heterogeneous multi-view graph for session-based recommendation via subsequence units

Fan Yang
Li Ji
Shuo Zhang
Dunlu Peng
Yiming Xu
Nan Chen

Session-based recommendation aims to capture user’s short-term dynamic preferences based on the dependencies between items within a session, and then predicts the next item that the user is most likely to interact with. Currently, session sequences are typically modeled as single-view structures, which focus on learning the interaction between individual items. However, these methods lack rich contextual information and are difficult to understand the user’s intent from a higher dimensional perspective. To better leverage the associations between contexts, this work proposes a Contrast-enhanced Heterogeneous Multi-view Graph via Subsequence Units (CHMGSU) for session-based recommendation. The sequences are modeled as both single-view and heterogeneous multi-view structures, where the single-view graph is constructed at the level of individual items to learn information transfer between items, while the heterogeneous multi-view graph is built using multiple consecutive items to better grasp the user’s high-dimensional intent. A hybrid readout function extracts the intent of subsequences, and captures relationships with contextual relevance. Next, single-view graph attention networks and heterogeneous multi-view graph neural networks are employed to generate item-level and subsequence-level embeddings. By fusing these two types of information, a session-level embedding with information from different perspectives is formed. The prediction results are optimized using the sample-adaptive loss function and the contrastive control gate. In addition, CHMGSU introduces Tmall, Gowalla, Diginetica and Nowplaying datasets to verify the effectiveness of the model on different types of datasets, and experimental results demonstrate that CHMGSU achieves consistent improvements over state-of-the-art baselines, thereby highlighting the incremental yet meaningful advancements achieved.

AAAI Conference 2026 Conference Paper

GitTaskBench: A Benchmark for Code Agents Solving Real-World Tasks Through Code Repository Leveraging

Ziyi Ni
Huacan Wang
Shuo Zhang
Shuo Lu
Ziyang He
WangYou
Zhenheng Tang
Sen Hu

Beyond scratch coding, exploiting large-scale code repositories (e.g., GitHub) for practical tasks is vital in real-world software development, yet current benchmarks rarely evaluate code agents in such authentic, workflow-driven scenarios. To bridge this gap, we introduce GitTaskBench, a benchmark designed to systematically assess this capability via 54 realistic tasks across 7 modalities and 7 domains. Each task pairs a relevant repository with an automated, human-curated evaluation harness specifying practical success criteria. Beyond measuring execution and task success, we also propose the alpha-value metric to quantify the economic benefit of agent performance, which integrates task success rates, token cost, and average developer salaries. Experiments across three state-of-the-art agent frameworks with multiple advanced LLMs show that leveraging code repositories for complex task solving remains challenging: even the best-performing system, OpenHands+Claude 3.7, solves only 48.15% of tasks. Error analysis attributes over half of failures to seemingly mundane yet critical steps like environment setup and dependency resolution, highlighting the need for more robust workflow management and increased timeout preparedness. By releasing GitTaskBench, we aim to drive progress and attention toward repository-aware code reasoning, execution, and deployment---moving agents closer to solving complex, end-to-end real-world tasks.

PDF Details DOI

EAAI Journal 2026 Journal Article

Multi-scale spatio-temporal hierarchical collaborative network and powergrid construction video action dataset for action recognition in power construction scenarios

Shuo Zhang
Yaguang Kong
Xiaoqing Zheng
Chuangxun Zhang
Weibin Zhang

In power engineering, construction accidents are often caused by unsafe actions of workers. Traditional manual inspections are inefficient and provide limited coverage, making it challenging to meet safety monitoring requirements of complex sites. Recently, video understanding and computer vision techniques have been widely applied to the development and deployment of artificial intelligence (AI) systems for unsafe behavior recognition. However, power construction scenarios pose challenges such as large variations in target scale, complex occlusions, and imbalanced class distributions, which negatively affect recognition performance. To address these issues, this paper implements a Multi-Scale Spatio-Temporal Hierarchical Collaborative Network (MSST-HCNet) and applies it to safety monitoring in power construction sites. MSST-HCNet achieves deep modeling of action representations through adaptive fusion of multi-scale hierarchical features. Its Adaptive Spatio-Temporal Hierarchical Module (ASHM) selectively integrates features from different stages of the slow pathway, capturing continuous action trajectories, enhancing recognition of small and occluded targets, and enabling multi-level feature learning. To evaluate the system, we constructed the PowerGrid Construction Video Action (PCVA) dataset, comprising 8300 video clips covering 30 typical construction behaviors. Experimental results show that MSST-HCNet achieves a mean Average Precision (mAP) of 38. 16%, outperforming the best comparative method by 7. 54 percentage points and significantly surpassing existing algorithms. This study demonstrates both the implementation of an effective AI model and its practical application for safety monitoring in power construction sites, reducing accident risk and inspection time while improving operational safety management. The code is publicly available at https: //github. com/ys-sy/MSST.

EAAI Journal 2026 Journal Article

Rethinking the local constraints: Geometric continuity regularization for image alignment

Yinqi Chen
Yangting Zheng
Peiwen Li
Weijian Luo
Shuo Kang
Xiang Gao
Chao Liu
Shuo Zhang

In image alignment, existing studies frequently neglect the modeling of featureless areas where reliable features are inherently absent. While indirect strategies, such as adding more geometric features, have been used to reduce such regions, they are limited by the natural variability of scenes. Instead, directly modeling these areas allows local consistency constraints to propagate transformations from feature-rich to featureless regions. However, existing local consistency constraints rely solely on parametric continuity (C1), which can cause excessive smoothness and distortion due to the excessive constraints on parameters. In contrast, geometric continuity (G1) relaxes parameter constraints and ensures visual accuracy, leading to results with lower distorted energy. Thus, this paper, for the first time, rigorously examines the rationale of local constraints, validates their capacity for featureless-region modeling, and theoretically demonstrates that G1 continuity effectively minimizes distortion. Building on these analyses, we introduce G1 continuity regularization; to enforce this property, the regularization term directly penalizes deviations from collinearity at mesh vertices or within network-learned transformations. Compared with existing approaches, our method achieves markedly superior performance.

AAAI Conference 2026 Conference Paper

RetouchAgent: Towards Interactive and Explainable Image Retouching with MLLM Agents

Shuo Zhang
Xinyu Yang

Although deep learning-based image retouching has made significant progress, its inherent subjectivity renders current black-box methods limited in interactivity and explainability. Among existing efforts, parameter-controlled methods aim to improve interactivity, but often suffer from ambiguous semantics and lack support for natural language control. Reinforcement learning–based explainability methods are constrained by low-dimensional and limited action spaces, which result in suboptimal performance. To address the above issues, we propose RetouchAgent, a novel framework that leverages collaboration among multiple MLLM agents for image retouching. Our method consists of the following key steps: (1) Retrieval: By constructing a multimodal retouching database, we enable an ICL sample retrieval mechanism guided by retouching intent. (2) Engine: Leveraging the vision-language understanding capabilities of MLLM, a carefully designed prompting strategy, and a dedicated operation library, we enable precise and controllable image retouching. (3) Reflection: We evaluate each retouching interaction and optimize the retouching process for progressive result refinement. Finally, through multiple rounds of collaboration among MLLM agents, RetouchAgent achieves state-of-the-art performance in quantitative and qualitative evaluations.

PDF Details DOI

AAAI Conference 2026 Conference Paper

SimpleDiffusion: A Lightweight and Efficient Conditional Diffusion Model for Multi-Modal Salient Object Detection

Shuo Zhang
Jiaming Huang
Wenbing Tang
Jing Liu
LI HAN
Jiandun Li
Hongchun Yuan
Zizhu Fan

Multi-modal salient object detection (MSOD), which integrates complementary modalities such as depth or thermal data, primarily faces two challenges: accurately preserving salient object details and effectively aligning cross-modal features. Recent advances in using Stable Diffusion to generate images with fine edge details have inspired researchers to reformulate MSOD as a conditional mask generation process guided by salient features, which has achieved excellent visual results. However, these approaches often overlook the high computational cost and large-scale architecture of Stable Diffusion, both of which render it unsuitable for real-world MSOD applications. Therefore, we propose SimpleDiffusion, the first lightweight and efficient conditional diffusion model for MSOD that does not rely on Stable Diffusion. Specifically, we propose an Adaptive Cross-Modal Fusion Conditional Network and a Latent Denoising Network to reduce the complexity of diffusion models. Furthermore, we design a Multi-modal Feature Rectification and Fusion Module to enhance the representational capacity of cross-modal salient features. Customized training and sampling strategies are also developed to improve inference efficiency and reduce erroneous object segmentations. Experiments on multiple MSOD datasets demonstrate that SimpleDiffusion reduces model size by over tenfold and improves inference speed by more than fivefold compared to other diffusion-based methods, while maintaining comparable or superior performance.

PDF Details DOI

AAAI Conference 2026 Conference Paper

Synergizing Multigrid Algorithms with Vision Transformer: A Novel Approach to Enhance the Seismic Foundation Model

Huiwen Wu
Shuo Zhang
Yi Liu
Hongbin Ye

Due to the rapid advancement and homogenization of Artificial Intelligence (AI) technology development, transformer-based foundation models have revolutionized scientific applications, such as drug discovery, materials research, and astronomy. However, seismic data presents unique characteristics that require specialized processing techniques for pretraining foundation models in seismic contexts with high- and low-frequency features playing crucial roles. Existing Vision Transformer (ViT) with sequential image tokenization fails to efficiently and effectively capture both high- and low-frequency seismic information because they ignore the intrinsic structural patterns of seismograms. This work introduces ADATG, a novel adaptive two-grid training strategy with Hilbert encoding, explicitly tailored for seismogram data and leveraging the hierarchical structures inherent in seismic data. Specifically, our approach employs spectrum decomposition to separate high- and low-frequency components, and hierarchical Hilbert encoding to represent the data effectively. Moreover, inspired by the frequency principle, we propose an adaptive training strategy that initially emphasizes coarse-level information and then progressively refines the model's focus on fine-level features. Extensive experiments demonstrate the effectiveness and efficiency of our method. This research highlights the importance of data encoding and training strategies informed by the distinct characteristics of high- and low-frequency features in seismic images, ultimately enhancing the pretraining of visual seismic foundation models.

PDF Details DOI

AIIM Journal 2025 Journal Article

A novel memory interaction neural network for multi-label drug–drug interaction prediction with neighbor importance sampling

Jing Wang
Runzhi Li
Shuo Zhang
YunLi Xing
Siyu Yan
Lihong Ma

Co-administration of multiple drugs can frequently cause drug–drug interactions (DDIs), including adverse drug reactions (ADRs) that may increase the likelihood of morbidity and mortality. Identifying potential DDIs presents a significant challenge, due to the complexity of pharmacology. Recent advances in knowledge graphs have contributed to DDI prediction by providing a robust framework for representing various relationships between drugs and other entities, such as proteins, diseases, and drug attributes. However, current network-based models often fail to uncover interaction information among DDI triplets, as they typically encode triplets independently. Additionally, uniform sampling methods may overlook differences in neighboring node properties. In this work, we propose a novel memory interaction neural network for DDI prediction, which integrates drug molecular sequences with semantic information from the drug knowledge graph. Specifically, we introduce a neighbor importance sampling strategy that selectively samples highly connected neighbors, improving computational efficiency and reducing noise. We also design a memory interaction module that utilizes multi-head attention mechanisms and deep neural networks to capture interactions among DDI triplets. Experimental evaluation on KEGG and OGB-biokg datasets demonstrates the superiority of our model compared to classical and state-of-the-art methods in predicting DDIs. Datasets and code for this proposed DDIs prediction model are freely accessible at https: //github. com/wj1108114106/Multi-label-DDIs.

AAAI Conference 2025 Conference Paper

AI-Powered Algorithm-Centric Quantum Processor Topology Design

Tian Li
Xiao-Yue Xu
Chen Ding
Tian-Ci Tian
Wei-You Liao
Shuo Zhang
He-Liang Huang

Quantum computing promises to revolutionize various fields, yet the execution of quantum programs necessitates an effective compilation process. This involves strategically mapping quantum circuits onto the physical qubits of a quantum processor. The qubits' arrangement, or topology, is pivotal to the circuit's performance, a factor that often defies traditional heuristic or manual optimization methods due to its complexity. In this study, we introduce a novel approach leveraging reinforcement learning to dynamically tailor qubit topologies to the unique specifications of individual quantum circuits, guiding algorithm-driven quantum processor topology design for reducing the depth of mapped circuit, which is particularly critical for the output accuracy on noisy quantum processors. Our method marks a significant departure from previous methods that have been constrained to mapping circuits onto a fixed processor topology. Experiments demonstrate that we have achieved notable enhancements in circuit performance, with a minimum of 20% reduction in circuit depth in 60% of the cases examined, and a maximum enhancement of up to 46%. Furthermore, the pronounced benefits of our approach in reducing circuit depth become increasingly evident as the scale of the quantum circuits increases, exhibiting the scalability of our method in terms of problem size. This work advances the co-design of quantum processor architecture and algorithm mapping, offering a promising avenue for future research and development in the field.

PDF Details DOI

IROS Conference 2025 Conference Paper

Crouch Gait Recognition of Children with Cerebral Palsy Based on CNN-LSTM Hybrid Model

Junhang Liu
Mingxiang Luo
Shuo Zhang
Wujing Cao
Xinyu Wu 0001

Crouch gait is one of the key characteristics of children with cerebral palsy, and early detection of gait changes is crucial for subsequent exoskeleton-assisted therapy. This study uses the Vicon 3D motion capture system to collect experimental data on four gait phases of children with cerebral palsy and introduces a CNN-LSTM hybrid model. The model combines the spatial feature extraction strengths of CNN with the temporal sequence modeling capabilities of LSTM, enabling it to effectively identify the complex dynamic changes in gait specific to children with cerebral palsy. By integrating these two components, the model not only accurately extracts key gait features but also captures the temporal dependencies within the gait cycle, allowing for precise recognition of crouch gait. Experimental results demonstrate that the proposed model exhibits good robustness and achieves high accuracy in both overall gait recognition and distinguishing the four individual gait phases. It significantly outperforms traditional machine learning architectures.

AAAI Conference 2025 Conference Paper

DiMSOD: A Diffusion-Based Framework for Multi-Modal Salient Object Detection

Shuo Zhang
Jiaming Huang
Wenbing Tang
Yan Wu
Terrence Hu
Xiaogang Xu
Jing Liu

Multi-modal salient object detection (SOD) through the integration of additional data such as depth or thermal information has become a significant task in computer vision during recent years. Traditionally, the challenges of identifying salient objects in RGB, RGB-D (Depth), and RGB-T (Thermal) images are tackled separately. However, without intricate cross-modal fusion strategies, such approaches struggle to effectively integrate multi-modal information, often resulting in poorly defined object edges or overconfident inaccurate predictions. Recent studies have shown that designing a unified end-to-end framework to handle all three types of SOD tasks simultaneously is both necessary and difficult. To address this need, we propose a novel approach that treats multi-modal SOD as a conditional mask generation task utilizing diffusion models. We introduce DiMSOD, which enables the concurrent use of local (depth maps, thermal maps) and global controls (original images) within a unified model for progressive denoising and refined prediction. DiMSOD is efficient, only requiring fine-tuning of our newly introduced modules on the existing stable diffusion, which not only reduces the fine-tuning cost, making it more viable for practical use, but also enhances the integration of multi-modal conditional controls. Specifically, we have developed modules including SOD-ControlNet, Feature Adaptive Network (FAN), and Feature Injection Attention Network (FIAN) to enhance the model's performance. Extensive experiments demonstrate that DiMSOD efficiently detects salient objects across RGB, RGB-D, and RGB-T datasets, achieving superior performance compared to previous well-established methods.

PDF Details DOI

NeurIPS Conference 2025 Conference Paper

EAReranker: Efficient Embedding Adequacy Assessment for Retrieval Augmented Generation

Dongyang Zeng
Yaping Liu
Wei Zhang
Shuo Zhang
Xinwang Liu
Binxing Fang

With the increasing adoption of Retrieval-Augmented Generation (RAG) systems for knowledge-intensive tasks, ensuring the adequacy of retrieved documents has become critically important for generation quality. Traditional reranking approaches face three significant challenges: substantial computational overhead that scales with document length, dependency on plain text that limits application in sensitive scenarios, and insufficient assessment of document value beyond simple relevance metrics. We propose EAReranker, an efficient embedding-based adequacy assessment framework that evaluates document utility for RAG systems without requiring access to original text content. The framework quantifies document adequacy through a comprehensive scoring methodology considering verifiability, coverage, completeness and structural aspects, providing interpretable adequacy classifications for downstream applications. EAReranker employs a Decoder-Only Transformer architecture that introduces embedding dimension expansion method and bin-aware weighted loss, designed specifically to predict adequacy directly from embedding vectors. Our comprehensive evaluation across four public benchmarks demonstrates that EAReranker achieves competitive performance with state-of-the-art plaintext rerankers while maintaining constant memory usage ($\sim$550MB) regardless of input length and processing 2-3x faster than traditional approaches. The semantic bin adequacy prediction accuracy of 92. 85\% LACC@10 and 86. 12\% LACC@25 demonstrates its capability to effectively filter out inadequate documents that could potentially mislead or adversely impact RAG system performance, thereby ensuring only high-utility information serves as generation context. These results establish EAReranker as an efficient and practical solution for enhancing RAG system performance through improved context selection while addressing the computational and privacy challenges of existing methods.

AAAI Conference 2025 Conference Paper

Exploring Intrinsic Alignments Within Text Corpus

Zi Liang
Pinghui Wang
Ruofei Zhang
Haibo Hu
Shuo Zhang
Qingqing Ye
Nuo Xu
Yaxin Xiao

Recent years have witnessed rapid advancements in the safety alignments of large language models (LLMs). Methods such as supervised instruction fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) have thus emerged as vital components in constructing LLMs. While these methods achieve robust and fine-grained alignment to human values, their practical application is still hindered by high annotation costs and incomplete human alignments. Besides, the intrinsic human values within training corpora have not been fully exploited. To address these issues, we propose ISAAC (Intrinsically Supervised Alignments by Assessing Corpus), a primary and coarse-grained safety alignment strategy for LLMs. ISAAC only relies on a prior assumption about the text corpus, and does not require preferences in RLHF or human responses selection in SFT. Specifically, it assumes a long-tail distribution of text corpus and employs a specialized sampling strategy to automatically sample high-quality responses. Theoretically, we prove that this strategy can improve the safety of LLMs under our assumptions. Empirically, our evaluations on mainstream LLMs show that ISAAC achieves a safety score comparable to current SFT solutions. Moreover, we conduct experiments on ISAAC for some RLHF-based LLMs, where we find that ISAAC can even improve the safety of these models under specific safety domains. These findings demonstrate that ISAAC can provide preliminary alignment to LLMs, thereby reducing the construction costs of existing human-feedback-based methods.

PDF Details DOI

JBHI Journal 2025 Journal Article

Generalized Cross-Domain Framework for Gesture Recognition via Wrist-Worn Sensing

Shuo Zhang
Jin Qi
Duidi Wu
Qianyou Zhao
Jie Hu

Wearable sensing technology offers a natural and convenient means of human-computer interaction, particularly for gesture recognition, yet domain shifts in wrist-worn single-site sensing pose significant challenges for cross-domain gesture recognition. To address this, we proposed a generalized cross-domain framework for fine-grained gesture recognition using wrist-worn single-site sensing. Concretely, we presented a Multi-Branch Network, which combines feature-level multimodal fusion with enhanced inter-modal interaction to effectively capture fine-grained gestures. To this end, we constructed a multimodal dataset, which comprises fifteen static and eighteen dynamic gestures. Furthermore, we developed five fine-tuning strategies and evaluated them across the paradigms of cross-session, cross-subject, cross-gesture, and cross-modality. Through comprehensive analyses, this study provides valuable insights into the selection of optimal fine-tuning strategies and elucidates the internal mechanisms underlying multiple cross-domain paradigms. To investigate the intricate trade-off between recognition accuracy and computational cost, we applied nonlinear least squares to construct the Accuracy-Cost trade-off functions. Experimental findings indicated that the optimal transfer learning ratios for these cross-domain paradigms ranged from 6. 1% to 9. 0%, with most clustering around 9. 0%, offering a valuable reference for determining optimal transfer learning ratios within diverse cross-domain scenarios. Additionally, we implemented a real-time online gesture recognition system, validating the feasibility of our approach through preliminary tests in real-world scenarios. In conclusion, this study serves as a preliminary investigation into the application of wrist-worn single-site sensing for fine-grained gesture recognition.

AAAI Conference 2025 Conference Paper

MTVHunter: Smart Contracts Vulnerability Detection Based on Multi-Teacher Knowledge Translation

Guokai Sun
Yuan Zhuang
Shuo Zhang
Xiaoyu Feng
Zhenguang Liu
Liguo Zhang

Smart contracts, closely intertwined with cryptocurrency transactions, have sparked widespread concerns about considerable financial losses of security issues. To counteract this, a variety of tools have been developed to identify vulnerability in smart contract. However, they fail to overcome two challenges at the same time when faced with smart contract bytecode: (i) strong interference caused by enormous non-relevant instructions; (ii) missing semantics of bytecode due to incomplete data and control flow dependencies. In this paper, we propose a multi-teacher based bytecode vulnerability detection method, namely Multi-Teacher Vulnerability Hunter (MTVHunter), which delivers effective denoising and missing semantic to bytecode under multi-teacher guidance. Specifically, we first propose an instruction denoising teacher to eliminate noise interference by abstract vulnerability pattern and further reflect in contract embeddings. Secondly, we design a novel semantic complementary teacher with neuron distillation, which effectively extracts necessary semantic from source code to replenish the bytecode. Particularly, the proposed neuron distillation accelerate this semantic filling by turning the knowledge transition into a regression task. We conduct experiments on 229,178 real-world smart contracts that concerns four types of common vulnerabilities. Extensive experiments show MTVHunter achieves significantly performance gains over state-of-the-art approaches.

PDF Details DOI

NeurIPS Conference 2025 Conference Paper

REFED: A Subject Real-time Dynamic Labeled EEG-fNIRS Synchronized Recorded Emotion Dataset

Xiaojun Ning
Jing Wang
Zhiyang Feng
Tianzuo Xin
Shuo Zhang
Shaoqi Zhang
Zheng Lian
Yi Ding

Affective brain-computer interfaces (aBCIs) play a crucial role in personalized human–computer interaction and neurofeedback modulation. To develop practical and effective aBCI paradigms and to investigate the spatial-temporal dynamics of brain activity under emotional inducement, portable electroencephalography (EEG) signals have been widely adopted. To further enhance spatial-temporal perception, functional near-infrared spectroscopy (fNIRS) has attracted increasing interest in the aBCI field and has been explored in combination with EEG. However, existing datasets typically provide only static fixation labels, overlooking the dynamic changes in subjects' emotions. Notably, some studies have attempted to collect continuously annotated emotional data, but they have recorded only peripheral physiological signals without directly observing brain activity, limiting insight into underlying neural states under different emotions. To address these challenges, we present the Real-time labeled EEG-fNIRS Dataset (REFED). To the best of our knowledge, this is the first EEG-fNIRS dataset with real-time dynamic emotional annotations. REFED simultaneously records brain signals from both EEG and fNIRS modalities while providing continuous, real-time annotations of valence and arousal. The results of the data analysis demonstrate the effectiveness of emotion inducement and the reliability of real-time annotation. This dataset offers the possibility for studying the neurovascular coupling mechanism under emotional evolution and for developing dynamic, robust affective BCIs.

NeurIPS Conference 2025 Conference Paper

RepoMaster: Autonomous Exploration and Understanding of GitHub Repositories for Complex Task Solving

Huacan Wang
Ziyi Ni
Shuo Zhang
Shuo Lu
Sen Hu
Ziyang He
Chen Hu
Jiaye Lin

The ultimate goal of code agents is to solve complex tasks autonomously. Although large language models (LLMs) have made substantial progress in code generation, real-world tasks typically demand full-fledged code repositories rather than simple scripts. Building such repositories from scratch remains a major challenge. Fortunately, GitHub hosts a vast, evolving collection of open-source repositories, which developers frequently reuse as modular components for complex tasks. Yet, existing frameworks like OpenHands and SWE-Agent still struggle to effectively leverage these valuable resources. Relying solely on README files provides insufficient guidance, and deeper exploration reveals two core obstacles: overwhelming information and tangled dependencies of repositories, both constrained by the limited context windows of current LLMs. To tackle these issues, we propose RepoMaster, an autonomous agent framework designed to explore and reuse GitHub repositories for solving complex tasks. For efficient understanding, RepoMaster constructs function-call graphs, module-dependency graphs, and hierarchical code trees to identify essential components, providing only identified core elements to the LLMs rather than the entire repository. During autonomous execution, it progressively explores related components using our exploration tools and prunes information to optimize context usage. Evaluated on the adjusted MLE-bench, RepoMaster achieves a 110\% relative boost in valid submissions over the strongest baseline OpenHands. On our newly released GitTaskBench, RepoMaster lifts the task-pass rate from 40. 7% to 62. 9% while reducing token usage by 95%. Our code and demonstration materials are publicly available at https: //github. com/QuantaAlpha/RepoMaster.

AAAI Conference 2025 Conference Paper

Unlocking the Potential of Reverse Distillation for Anomaly Detection

Xinyue Liu
Jianyuan Wang
Biao Leng
Shuo Zhang

Knowledge Distillation (KD) is a promising approach for unsupervised Anomaly Detection (AD). However, the student network's over-generalization often diminishes the crucial representation differences between teacher and student in anomalous regions, leading to detection failures. To address this problem, the widely accepted Reverse Distillation (RD) paradigm designs the asymmetry teacher and student network, using an encoder as teacher and a decoder as student. Yet, the design of RD does not ensure that the teacher encoder effectively distinguishes between normal and abnormal features or that the student decoder generates anomaly-free features. Additionally, the absence of skip connections results in a loss of fine details during feature reconstruction. To address these issues, we propose RD with Expert, which introduces a novel Expert-Teacher-Student network for simultaneous distillation of both the teacher encoder and student decoder. The added expert network enhances the student's ability to generate normal features and optimizes the teacher's differentiation between normal and abnormal features, reducing missed detections. Additionally, Guided Information Injection is designed to filter and transfer features from teacher to student, improving detail reconstruction and minimizing false positives. Experiments on several benchmarks prove that our method outperforms existing unsupervised AD methods under RD paradigm, fully unlocking RD’s potential.

PDF Details DOI

AAAI Conference 2024 Conference Paper

Data Shunt: Collaboration of Small and Large Models for Lower Costs and Better Performance

Dong Chen
Yueting Zhuang
Shuo Zhang
Jinfeng Liu
Su Dong
Siliang Tang

Pretrained large models, particularly large language models, have garnered increasing attention, as they have demonstrated remarkable abilities through contextual learning. Pretrained large models are increasingly recognized as fundamental tools for solving various tasks. However, the substantial computational demands of large models have dissuaded most product teams and individuals from running them. In such scenarios, to leverage the exceptional performance of large models, one must solely depend on costly APIs, further burdening product teams and individuals. On the other hand, despite the overall inferior performance of small models compared to large models, there are certain distributions where small models can achieve comparable or even superior results. For instance, during training, small models may become trapped in a local optimum that is unique to certain distributions, leading to superior performance. Hence, we propose Data Shunt (DS), a general paradigm for collaboration of small and large models. DS not only substantially reduces the cost associated with deploying large models but also effectively enhances overall performance. Specifically, DS determines the shunting direction by evaluating the confidence level of small models. When the confidence level falls below a specific threshold, the input data is forwarded to large models. To further leverage the advantages of the small and large models, we introduce Prompt Pruning (PP) and 2-Stage Confidence Distillation (2CD), which facilitate mutual collaboration, leading to better results and less cost. The remarkable performance across diverse modalities and tasks demonstrates the superiority of the proposed DS over large models. For instance, ChatGPT achieves an accuracy of 94.43% on Amazon Product sentiment analysis, and DS achieves an accuracy of 95.64%, while the cost has been reduced to only 31.18%. The code for the proposed method are provided for research purposes https://github.com/Anfeather/Data-Shunt.

PDF Details DOI

TCS Journal 2024 Journal Article

Improved deterministic algorithms for non-monotone submodular maximization

Xiaoming Sun
Jialin Zhang
Shuo Zhang
Zhijie Zhang

Submodular maximization is one of the central topics in combinatorial optimization. It has found numerous applications in the real world. In the past decades, a series of algorithms have been proposed for this problem. However, most of the state-of-the-art algorithms are randomized. There remain non-negligible gaps with respect to approximation ratios between deterministic and randomized algorithms in submodular maximization. In this paper, we propose deterministic algorithms with improved approximation ratios for non-monotone submodular maximization. Specifically, for the matroid constraint, we provide a deterministic 0. 283 − o ( 1 ) approximation algorithm, while the previous best deterministic algorithm only achieves a 1/4 approximation ratio. For the knapsack constraint, we provide a deterministic 1/4 approximation algorithm, while the previous best deterministic algorithm only achieves a 1/6 approximation ratio. For the linear packing constraints with large widths, we provide a deterministic 1 / 6 − ϵ approximation algorithm. To the best of our knowledge, there is currently no deterministic approximation algorithm for the constraints.

AAAI Conference 2024 Conference Paper

Learning with Noisy Labels Using Hyperspherical Margin Weighting

Shuo Zhang
Yuwen Li
Zhongyu Wang
Jianqing Li
Chengyu Liu

Datasets often include noisy labels, but learning from them is difficult. Since mislabeled examples usually have larger loss values in training, the small-loss trick is regarded as a standard metric to identify the clean example from the training set for better performance. Nonetheless, this proposal ignores that some clean but hard-to-learn examples also generate large losses. They could be misidentified by this criterion. In this paper, we propose a new metric called the Integrated Area Margin (IAM), which is superior to the traditional small-loss trick, particularly in recognizing the clean but hard-to-learn examples. According to the IAM, we further offer the Hyperspherical Margin Weighting (HMW) approach. It is a new sample weighting strategy that restructures the importance of each example. It should be highlighted that our approach is universal and can strengthen various methods in this field. Experiments on both benchmark and real-world datasets indicate that our HMW outperforms many state-of-the-art approaches in learning with noisy label tasks. Codes are available at https://github.com/Zhangshuojackpot/HMW.

PDF Details DOI

AAAI Conference 2024 Conference Paper

MERGE: Fast Private Text Generation

Zi Liang
Pinghui Wang
Ruofei Zhang
Nuo Xu
Shuo Zhang
Lifeng Xing
Haitao Bai
Ziyang Zhou

The drastic increase in language models' parameters has led to a new trend of deploying models in cloud servers, raising growing concerns about private inference for Transformer-based models. Existing two-party privacy-preserving techniques, however, only take into account natural language understanding (NLU) scenarios. Private inference in natural language generation (NLG), crucial for applications like translation and code completion, remains underexplored. In addition, previous privacy-preserving techniques suffer from convergence issues during model training and exhibit poor inference speed when used with NLG models due to the neglect of time-consuming operations in auto-regressive generations. To address these issues, we propose a fast private text generation framework for Transformer-based language models, namely MERGE. MERGE reuses the output hidden state as the word embedding to bypass the embedding computation and reorganize the linear operations in the Transformer module to accelerate the forward procedure. Extensive experiments show that MERGE achieves a 26.5x speedup to the vanilla encrypted model under the sequence length 512, and reduces 80% communication cost, with an up to 10x speedup to state-of-the-art approximated models.

PDF Details DOI

JBHI Journal 2023 Journal Article

CCS-Net: Cascade Detection Network With the Convolution Kernel Switch Block and Statistics Optimal Anchors Block in Hypopharyngeal Cancer MRI

Shuo Zhang
Yang Miao
Jun Chen
Xiwei Zhang
Lei Han
Zehao Huang
Ning Pei
Haibin Liu

Magnetic resonance imaging (MRI) is a common diagnostic method for hypopharyngeal cancer (HPC). It is a challenge to automatically detect HPC tumors and swollen lymph nodes (HPC risk areas) from MRI slices because of the small size and irregular shape of HPC risk areas. Herein, we propose a cascade detection network with Convolution Kernel Switch (CKS) Block and Statistics Optimal Anchors (SOA) Block in HPC MRI (CCS-Net). CKS Block can adaptively switch standard convolution to deformable convolution in some appropriate layers to detect irregular objects more efficiently without taking up too much computing resources. SOA Block can automatically generate the optimal anchors based on the size distribution of objects. Compared with other methods, our method achieves splendid detection performance and outperforms other methods on the HPC dataset (more than 1800 T2 MRI slices), achieving the highest AP 50 of 78. 90%. Experiments show that the proposed network can be the basis of a computer aided diagnosis utility that helps achieve faster and more accurate diagnostic decisions for HPC.

EAAI Journal 2023 Journal Article

Event-triggered impulsive control for stability of stochastic delayed complex networks under deception attacks

Ni Yang
Shuo Zhang
Huan Su

In this paper, we consider the stability of stochastic complex networks with time-varying delays (SCNTD) under deception attacks, adopting the event-triggered impulsive control (ETIC). A new event-triggered mechanism (ETM) which was used to determine the impulsive instance is given using the topology of networks and the Lyapunov functions of subsystems. For stochastic complex networks, to avoid the Zeno behavior, a positive minimum inter-event time (MIET) is guaranteed. Combining the graph theory and the Lyapunov–Razumikhin method, several criteria for achieving p th moment exponential stability ( p MES) of SCNTD under deception attacks which are related to the topology of networks, the event-triggered parameters and the signal sent by the enemy are obtained. Finally, the theoretical results apply to the stochastic oscillator systems, single-link robot arms systems, and Chua’s circuit systems. The stability of these systems under deception attacks is considered separately, and to show the validity of our theories, numerical simulations are provided.

AAAI Conference 2023 Conference Paper

Multi-Action Dialog Policy Learning from Logged User Feedback

Shuo Zhang
Junzhou Zhao
Pinghui Wang
Tianxiang Wang
Zi Liang
Jing Tao
Yi Huang
Junlan Feng

Multi-action dialog policy (MADP), which generates multiple atomic dialog actions per turn, has been widely applied in task-oriented dialog systems to provide expressive and efficient system responses. Existing MADP models usually imitate action combinations from the labeled multi-action dialog samples. Due to data limitations, they generalize poorly toward unseen dialog flows. While reinforcement learning-based methods are proposed to incorporate the service ratings from real users and user simulators as external supervision signals, they suffer from sparse and less credible dialog-level rewards. To cope with this problem, we explore to improve MADPL with explicit and implicit turn-level user feedback received for historical predictions (i.e., logged user feedback) that are cost-efficient to collect and faithful to real-world scenarios. The task is challenging since the logged user feedback provides only partial label feedback limited to the particular historical dialog actions predicted by the agent. To fully exploit such feedback information, we propose BanditMatch, which addresses the task from a feedback-enhanced semi-supervised learning perspective with a hybrid learning objective of SSL and bandit learning. BanditMatch integrates pseudo-labeling methods to better explore the action space through constructing full label feedback. Extensive experiments show that our BanditMatch improves MADPL over the state-of-the-art methods by generating more concise and informative responses. The source code and the appendix of this paper can be obtained from https://github.com/ShuoZhangXJTU/BanditMatch.

PDF Details DOI

AAMAS Conference 2023 Conference Paper

Think Twice: A Human-like Two-stage Conversational Agent for Emotional Response Generation

Yushan Qian
Bo Wang
Shangzhao Ma
Wu Bin
Shuo Zhang
Dongming Zhao
Kun Huang
Yuexian Hou

Towards human-like dialogue systems, current emotional dialogue approaches jointly model emotion and semantics with a unified neural network. This strategy tends to generate safe responses due to the mutual restriction between emotion and semantics, and requires the rare large-scale emotion-annotated dialogue corpus. Inspired by the "think twice" behavior in human intelligent dialogue, we propose a two-stage conversational agent for the generation of emotional dialogue. Firstly, a dialogue model trained without the emotion-annotated dialogue corpus generates a prototype response that meets the contextual semantics. Secondly, the first-stage prototype is modified by a controllable emotion refiner with the empathy hypothesis. Experimental results on the DailyDialog and EmpatheticDialogues datasets demonstrate that the proposed conversational agent outperforms the compared models in the emotion generation and maintains the semantic performance in the automatic and human evaluations.

AAAI Conference 2022 Conference Paper

DuMLP-Pin: A Dual-MLP-Dot-Product Permutation-Invariant Network for Set Feature Extraction

Jiajun Fei
Ziyu Zhu
Wenlei Liu
Zhidong Deng
Mingyang Li
Huanjun Deng
Shuo Zhang

Existing permutation-invariant methods can be divided into two categories according to the aggregation scope, i. e. global aggregation and local one. Although the global aggregation methods, e. g. , PointNet and Deep Sets, get involved in simpler structures, their performance is poorer than the local aggregation ones like PointNet++ and Point Transformer. It remains an open problem whether there exists a global aggregation method with a simple structure, competitive performance, and even much fewer parameters. In this paper, we propose a novel global aggregation permutation-invariant network based on dual MLP dot-product, called DuMLP- Pin, which is capable of being employed to extract features for set inputs, including unordered or unstructured pixel, attribute, and point cloud data sets. We strictly prove that any permutation-invariant function implemented by DuMLP-Pin can be decomposed into two or more permutation-equivariant ones in a dot-product way as the cardinality of the given input set is greater than a threshold. We also show that the DuMLP- Pin can be viewed as Deep Sets with strong constraints under certain conditions. The performance of DuMLP-Pin is evaluated on several different tasks with diverse data sets. The experimental results demonstrate that our DuMLP-Pin achieves the best results on the two classification problems for pixel sets and attribute sets. On both the point cloud classification and the part segmentation, the accuracy of DuMLP-Pin is very close to the so-far best-performing local aggregation method with only a 1-2% difference, while the number of required parameters is significantly reduced by more than 85% in classification and 69% in segmentation, respectively. The code is publicly available on https: //github. com/JaronTHU/ DuMLP-Pin.

EAAI Journal 2022 Journal Article

Improved double TQWT sparse representation using the MQGA algorithm and new norm for aviation bearing compound fault detection

Shuo Zhang
Zhiwen Liu
Sihai He
Jinglin Wang
Lufeng Chen

The double tunable wavelet transform sparse representation realizes signal decomposition by constructing a basis function dictionary that match various characteristic waveforms of compound fault signal. However, the quality factor describing the resonance characteristic of the wavelet basis function can only be determined from practical experience, which is often subjective, and can significantly affect the matching degree between the wavelet basis function and the fault signal. To solve this problem, a new sparse representation method and a new norm are proposed. First, the multi-population quantum genetic algorithm (MQGA) is used to optimize the selected quality factor parameter combinations. The cross-correlated kurtosis of the periodic impact signal is established as the new norm and used to evaluate the optimized parameters. Then, according to the principle of energy entropy dominance, main sub-bands of the low resonance component are reconstructed to reduce noise interference and enhance the impact characteristics of the signal. Finally, Hilbert envelope demodulation analysis is performed on the reconstructed signal to obtain the instantaneous fault characteristic frequency. The proposed method was applied to diagnose compound faults of aviation bearings. The results show that the proposed method can effectively separate and extract the compound fault signal of a bearing in an aero-engine testbed. Furthermore, the compound fault of a damaged bearing in a helicopter transmission system was successfully decoupled, which verified the effectiveness and practicability of the proposed method.

JBHI Journal 2022 Journal Article

Interpretability Analysis of One-Year Mortality Prediction for Stroke Patients Based on Deep Neural Network

Shuo Zhang
Jing Wang
Lulu Pei
Kai Liu
Yuan Gao
Hui Fang
Rui Zhang
Lu Zhao

Clinically, physicians collect the benchmark medical data to establish archives for a stroke patient and then add the follow up data regularly. It has great significance on prognosis prediction for stroke patients. In this paper, we present an interpretable deep learning model to predict the one-year mortality risk on stroke. We design sub-modules to reconstruct features from original clinical data that highlight the dissimilarity and temporality of different variables. The model consists of Bidirectional Long Short-Term Memory (Bi-LSTM), in which a novel correlation attention module is proposed that takes the correlation of variables into consideration. In experiments, datasets are collected clinically from the department of neurology in a local AAA hospital. It consists of 2, 275 stroke patients hospitalized in the department of neurology from 2014 to 2016. Our model achieves a precision of 0. 9414, a recall of 0. 9502 and an F1-score of 0. 9415. In addition, we provide the analysis of the interpretability by visualizations with reference to clinical professional guidelines.

IJCAI Conference 2022 Conference Paper

“Think Before You Speak”: Improving Multi-Action Dialog Policy by Planning Single-Action Dialogs

Shuo Zhang
Junzhou Zhao
Pinghui Wang
Yu Li
Yi Huang
Junlan Feng

Multi-action dialog policy (MADP), which generates multiple atomic dialog actions per turn, has been widely applied in task-oriented dialog systems to provide expressive and efficient system responses. Existing MADP models usually imitate action combinations from the labeled multi-action dialog samples. Due to data limitations, they generalize poorly toward unseen dialog flows. While interactive learning and reinforcement learning algorithms can be applied to incorporate external data sources of real users and user simulators, they take significant manual effort to build and suffer from instability. To address these issues, we propose Planning Enhanced Dialog Policy (PEDP), a novel multi-task learning framework that learns single-action dialog dynamics to enhance multi-action prediction. Our PEDP method employs model-based planning for conceiving what to express before deciding the current response through simulating single-action dialogs. Experimental results on the MultiWOZ dataset demonstrate that our fully supervised learning-based method achieves a solid task success rate of 90. 6%, improving 3% compared to the state-of-the-art methods. The source code and the appendix of this paper can be obtained from https: //github. com/ShuoZhangXJTU/PEDP.

PDF Details DOI

AAAI Conference 2021 Conference Paper

Attention-based Multi-Level Fusion Network for Light Field Depth Estimation

Jiaxin Chen
Shuo Zhang
Youfang Lin

Depth estimation from Light Field (LF) images is a crucial basis for LF related applications. Since multiple views with abundant information are available, how to effectively fuse features of these views is a key point for accurate LF depth estimation. In this paper, we propose a novel attention-based multi-level fusion network. Combining with the four-branch structure, we design intra-branch fusion strategy and interbranch fusion strategy to hierarchically fuse effective features from different views. By introducing the attention mechanism, features of views with less occlusions and richer textures are selected inside and between these branches to provide more effective information for depth estimation. The depth maps are finally estimated after further aggregation. Experimental results show the proposed method achieves state-of-the-art performance in both quantitative and qualitative evaluation, which also ranks first in the commonly used HCI 4D Light Field Benchmark.

AAAI Conference 2021 Conference Paper

Inferring Camouflaged Objects by Texture-Aware Interactive Guidance Network

Jinchao Zhu
Xiaoyu Zhang
Shuo Zhang
Junnan Liu

Camouflaged objects, similar to the background, show indefinable boundaries and deceptive textures, which increases the difficulty of detection task and makes the model rely on features with more information. Herein, we design a texture label to facilitate our network for accurate camouflaged object segmentation. Motivated by the complementary relationship between texture labels and camouflaged object labels, we propose an interactive guidance framework named TINet, which focuses on finding the indefinable boundary and the texture difference by progressive interactive guidance. It maximizes the guidance effect of refined multi-level texture cues on segmentation. Specifically, texture perception decoder (TPD) makes a comprehensive analysis of texture information in multiple scales. Feature interaction guidance decoder (FGD) interactively refines multi-level features of camouflaged object detection and texture detection level by level. Holistic perception decoder (HPD) enhances FGD results by multi-level holistic perception. In addition, we propose a boundary weight map to help the loss function pay more attention to the object boundary. Sufficient experiments conducted on COD and SOD datasets demonstrate that the proposed method performs favorably against 23 state-of-the-art methods.

AAAI Conference 2021 Conference Paper

Learning to Check Contract Inconsistencies

Shuo Zhang
Junzhou Zhao
Pinghui Wang
Nuo Xu
Yang Yang
Yiting Liu
Yi Huang
Junlan Feng

Contract consistency is important in ensuring the legal validity of the contract. In many scenarios, a contract is written by filling the blanks in a precompiled form. Due to carelessness, two blanks that should be filled with the same (or different) content may be incorrectly filled with different (or same) content. This will result in the issue of contract inconsistencies, which may severely impair the legal validity of the contract. Traditional methods to address this issue mainly rely on manual contract review, which is labor-intensive and costly. In this work, we formulate a novel Contract Inconsistency Checking (CIC) problem, and design an end-to-end framework, called Pair-wise Blank Resolution (PBR), to solve the CIC problem with high accuracy. Our PBR model contains a novel BlankCoder to address the challenge of modeling meaningless blanks. BlankCoder adopts a two-stage attention mechanism that adequately associates a meaningless blank with its relevant descriptions while avoiding the incorporation of irrelevant context words. Experiments conducted on real-world datasets show the promising performance of our method with a balanced accuracy of 94. 05% and an F1 score of 90. 90% in the CIC problem.

ICRA Conference 2021 Conference Paper

PSF-LO: Parameterized Semantic Features Based Lidar Odometry

Guibin Chen
Bosheng Wang
Xiaoliang Wang
Huanjun Deng
Bing Wang
Shuo Zhang

Lidar odometry (LO) is a key technology in numerous reliable and accurate localization and mapping systems of autonomous driving. The state-of-the-art LO methods generally leverage geometric information to perform point cloud registration. Furthermore, obtaining the point cloud semantic information describing the environment more abundantly will facilitate the registration. We present a novel semantic lidar odometry method based on self-designed parameterized seman-tic features (PSFs) to achieve low-drift ego-motion estimation for autonomous vehicle in real time. We first use a convolutional neural network-based algorithm to obtain point-wise semantics from the input laser point cloud, and then use semantic labels to separate road, building, traffic sign and pole-like point cloud and fit them separately to obtain corresponding PSFs. A fast PSF-based matching enables us to refine geometric features (GeFs) registration, thereby reducing the impact of blurred submap surface on the accuracy of GeFs matching. Besides, we design an efficient instance-level method to accurately recognize and remove the dynamic objects while retaining static ones in the semantic point cloud, which are beneficial to further improve the accuracy of LO. We evaluate our method, namely PSF-LO, on the public dataset KITTI Odometry Benchmark and rank #1 among semantic lidar methods with an average translational error of 0. 82% in the test dataset.

IJCAI Conference 2021 Conference Paper

Removing Foreground Occlusions in Light Field using Micro-lens Dynamic Filter

Shuo Zhang
Zeqi Shen
Youfang Lin

Foreground occlusion removal task aims to automatically detect and remove foreground occlusions and recover background objects. Since for Light Fields (LFs), background objects occluded in some views may be seen in other views, the foreground occlusion removal task for LFs is easy to achieve. In this paper, we propose a learning-based method combining ‘seeking’ and ‘generating’ to recover occluded background. Specifically, the micro-lens dynamic filters are proposed to ‘seek’ occluded background points in shifted micro-lens images and remove occlusions using angular information. The shifted images are then combined to further ‘generate’ background regions to supplement more background details using spatial information. By fully exploring the angular and spatial information in LFs, the dense and complex occlusions can be easily removed. Quantitative and qualitative experimental results show that our method outperforms other state-of-the-arts methods by a large margin.

PDF Details DOI

IJCAI Conference 2020 Conference Paper

Improving Attention Mechanism in Graph Neural Networks via Cardinality Preservation

Shuo Zhang
Lei Xie

Graph Neural Networks (GNNs) are powerful for the representation learning of graph-structured data. Most of the GNNs use a message-passing scheme, where the embedding of a node is iteratively updated by aggregating the information from its neighbors. To achieve a better expressive capability of node influences, attention mechanism has grown to be popular to assign trainable weights to the nodes in aggregation. Though the attention-based GNNs have achieved remarkable results in various tasks, a clear understanding of their discriminative capacities is missing. In this work, we present a theoretical analysis of the representational properties of the GNN that adopts the attention mechanism as an aggregator. Our analysis determines all cases when those attention-based GNNs can always fail to distinguish certain distinct structures. Those cases appear due to the ignorance of cardinality information in attention-based aggregation. To improve the performance of attention-based GNNs, we propose cardinality preserved attention (CPA) models that can be applied to any kind of attention mechanisms. Our experiments on node and graph classification confirm our theoretical analysis and show the competitive performance of our CPA models. The code is available online: https: //github. com/zetayue/CPA.

PDF Details DOI

IROS Conference 2020 Conference Paper

Vision Global Localization with Semantic Segmentation and Interest Feature Points

Kai Li
Xudong Zhang
Kun Li
Shuo Zhang

In this work, we present a vision-only global localization architecture for autonomous vehicle applications, and achieves centimeter-level accuracy and high robustness in various scenarios. We first apply pixel-wise segmentation to the front-view mono camera and extract the semantic features, e. g. pole-like objects, lane markings, and curbs, which are robust to illumination, viewing angles and seasonal changes. For the scenes without enough semantic information, we extract interest feature points on static backgrounds, such as ground surface and buildings, assisted by our semantic segmentation. We create the visual global map with semantic feature map layers extracted from LiDAR point-cloud semantic map and the point feature map layer built with a fixed-pose SFM. A lumped Levenberg-Marquardt optimization solver is then applied to minimize the cost from two types of observations. We further evaluate the accuracy and robustness of our method with road tests on Alibaba's autonomous delivery vehicles in multiple scenarios as well as a KAIST urban dataset.

TIST Journal 2020 Journal Article

Web Table Extraction, Retrieval, and Augmentation

Shuo Zhang
Krisztian Balog

Tables are powerful and popular tools for organizing and manipulating data. A vast number of tables can be found on the Web, which represent a valuable knowledge resource. The objective of this survey is to synthesize and present two decades of research on web tables. In particular, we organize existing literature into six main categories of information access tasks: table extraction, table interpretation, table search, question answering, knowledge base augmentation, and table augmentation. For each of these tasks, we identify and describe seminal approaches, present relevant resources, and point out interdependencies among the different tasks.