Author name cluster

Peng Tang

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.

17 papers

2 author rows

AAAI Conference 2026 Conference Paper

FedTopo: Topology-Informed Representation Alignment in Federated Learning Under Non-I.I.D. Conditions

Ke Hu
Liyao Xiang
Peng Tang
Weidong Qiu

Current federated-learning models deteriorate under heterogeneous (non-I.I.D.) client data, as their feature representations diverge and pixel- or patch-level objectives fail to capture the global topology which is essential for high-dimensional visual tasks. We propose FedTopo, a framework that integrates Topological-Guided Block Screening (TGBS) and Topological Embedding (TE) to leverage topological information, yielding coherently aligned cross-client representations by Topological Alignment Loss (TAL). First, Topology-Guided Block Screening (TGBS) automatically selects the most topology-informative block, i.e., the one with maximal topological separability, whose persistence-based signatures best distinguish within- versus between-class pairs, ensuring that subsequent analysis focuses on topology-rich features. Next, this block yields a compact Topological Embedding, which quantifies the topological information for each client. Finally, a Topological Alignment Loss (TAL) guides clients to maintain topological consistency with the global model during optimization, reducing representation drift across rounds. Experiments on Fashion-MNIST, CIFAR-10, and CIFAR-100 under four non-I.I.D. partitions show that FedTopo accelerates convergence and improves accuracy over strong baselines.

PDF Details DOI

AAAI Conference 2026 Conference Paper

Learning to Explore: Policy-Guided Outlier Synthesis for Graph Out-of-Distribution Detection

Li Sun
Lanxu Yang
Jiayu Tian
Bowen Fang
Xiaoyan Yu
Junda Ye
Peng Tang
Hao Peng

Detecting Out-of-Distribution (OOD) graphs—those are drawn from a different distribution from the training data-is a critical task for ensuring the safety and reliability of Graph Neural Networks. The main challenge in unsupervised graph-level Out-of-Distribution detection lies in its common reliance on purely in-distribution (ID) data. This ID-only training paradigm leads to an incomplete characterization of the feature space, resulting in decision boundaries that lack the robustness needed to effectively separate ID from OOD samples. While incorporating synthesized outliers into the training process is a promising direction, existing generation methods are limited by their dependence on pre-defined, non-adaptive sampling heuristics (e.g., distance- or density-based). Such fixed strategies lack the flexibility to systematically explore the most informative OOD regions for refining decision boundaries. To overcome this limitation, we propose a novel Policy-Guided Outlier Synthesis (PGOS) framework that replaces static heuristics with a learned, adaptive exploration policy. PGOS trains a reinforcement learning agent to autonomously navigate low-density regions within a structured latent space, sampling representations that are maximally effective for regularizing the OOD decision boundary. These sampled points are then decoded into high-quality pseudo-OOD graphs to enhance the detector's robustness. Extensive experiments demonstrate the strong performance of our method, state-of-the-art results on multiple graph OOD and anomaly detection benchmarks.

PDF Details DOI

JBHI Journal 2025 Journal Article

Robust R-Peak Detection in Noisy ECGs via Strip-Attention YOLO with Multilead Fusion

Wang Peng
Peng Tang
Hao Wang
Yuhang Liu
Qiang Li
Peng Zhang

The R-peak in electrocardiogram (ECG) signals is a critical physiological marker for the diagnosis of cardiovascular diseases. Although various R-peak detection methods have been proposed, their performance is often hindered by noise, especially in dynamic ECG monitoring. Furthermore, the potential of harnessing complementary information from 12-lead ECG signals has not been fully exploited. To address these challenges, this study conceptualized 12-lead ECG data as two-dimensional images and employed YOLOv5 as the model's backbone for R-peak detection, effectively transforming a signal segmentation task into an object detection task in images. Specifically, considering the characteristics of consistent R-peak positions across different leads, we proposed a strip attention mechanism to treat horizontal or vertical strips as tokens for computing inter- and intra-strip attention, enhancing the model's ability to capture R-peak positional information and likelihood. Additionally, a one-dimensional Manhattan distance-based NMS algorithm was used to minimize redundant detection frames, thereby enhancing model performance. The proposed model was rigorously evaluated on two publicly available datasets, INCART and LUDB, under varying noise conditions. On the INCART dataset, the model achieved F1 scores of 99. 97%, 99. 86%, 99. 63%, and 98. 00% at noise levels of Original, SNR = 10 dB, SNR = 5 dB, and SNR = 0 dB, respectively. Similarly, on the LUDB dataset, the F1 scores were 99. 89%, 100%, 100%, and 99. 86% for the corresponding noise levels. Extensive testing across multiple datasets and noise scenarios demonstrated that the proposed model outperformed existing state-of-the-art methods in terms of accuracy, noise robustness, and generalization capability.

Details DOI

EAAI Journal 2025 Journal Article

Self-supervised visual assessment of railway track curvature via homography learning based on projective curvilinear geometry model

Peng Tang
Zhibin Yu

Railway track curvature monitoring is crucial for ensuring operational safety and passenger comfort. As a robust complement to the single-point physical sensor approaches, vision-based methods have recently gained increasing adoption. However, existing approaches frequently neglect the systematic exploitation of railway ego-vision geometry in two critical aspects: (1) the rail-camera kinematic coupling that relates the rail appearance in camera’s view and the track curvature during curvilinear motion, and (2) the potential of self-supervised learning to overcome annotation scarcity in this domain. This geometric oversight limits their accuracy in real-world dynamic scenarios. To address these gaps, this study proposes a novel vision-based framework that systematically exploits the railway ego vision geometry. Our methodology comprises two key innovations: First, a projective curvilinear geometry model that mathematically relates the ground-planes-induced homography to actual track curvature, thereby establishing a mapping from curvature and its variation to rail imaging curves. Second, a self-supervised curvature prediction network trained using automatically generated labels from our geometric model, eliminating the need for manual curve annotations. The self-supervision is achieved through a cyclic consistency mechanism between predicted curvatures and reprojected image features. Experimental validation using real-world railway footage demonstrates significant improvements: Our method reduces the average root mean squared error by 23. 31% compared to state-of-the-art vision-based curvature estimation methods. These results underscore the effectiveness of geometry-aware computer vision for railway geometry monitoring

Details DOI

EAAI Journal 2025 Journal Article

Variational learning to rank for Test Case Prioritization via prioritizing metric inspired differentiable loss

Peng Tang
Junfeng Wang
Mingxing Liu

The Test Case Prioritization (TCP) plays a vital role in software testing by strategically ordering test cases to expose faults at the earliest feasible stages. Recent TCP studies have embraced learning-to-rank techniques to tackle the challenges of complexity and scalability in test ranking. However, the measurements of TCP, such as the Average Percentage of Fault Detection (APFD), are non-differentiable and incompatible with gradient-based optimizations. Moreover, learning models are susceptible to overfitting due to their complex natures of embedding capabilities. The aim of this study was to investigate differentiable APFD-inspired objectives and examine the consistency in priority embedding space, to jointly improve the model generalization. Therefore, we proposed a novel variational learning based TCP model, named APFD-Net. The model was constructed upon a simplified Set-Transformer and optimized by combining three losses, namely, the variational likelihood loss weighted by APFD difference, the differentiable soft APFD loss and the prior loss, thus making APFD applicable to direct optimizations. Industry datasets were used to verify the validity of our approach. Experiments results demonstrate that the APFD-Net achieves best effectiveness and relatively improved at least 4. 8% in APFD compared to state-of-art methods, indicating its potential for industrial applications.

Details DOI

AAAI Conference 2024 Conference Paper

DocFormerv2: Local Features for Document Understanding

Srikar Appalaraju
Peng Tang
Qi Dong
Nishant Sankaran
Yichu Zhou
R. Manmatha

We propose DocFormerv2, a multi-modal transformer for Visual Document Understanding (VDU). The VDU domain entails understanding documents (beyond mere OCR predictions) e.g., extracting information from a form, VQA for documents and other tasks. VDU is challenging as it needs a model to make sense of multiple modalities (visual, language and spatial) to make a prediction. Our approach, termed DocFormerv2 is an encoder-decoder transformer which takes as input - vision, language and spatial features. DocFormerv2 is pre-trained with unsupervised tasks employed asymmetrically i.e., two novel document tasks on encoder and one on the auto-regressive decoder. The unsupervised tasks have been carefully designed to ensure that the pre-training encourages local-feature alignment between multiple modalities. DocFormerv2 when evaluated on nine challenging datasets shows state-of-the-art performance on all over strong baselines - On TabFact (+4.3%), InfoVQA (+1.4%), FUNSD (+1.0%). Furthermore, to show generalization capabilities, on three VQA tasks involving scene-text, DocFormerv2 outperforms previous comparably-sized models and even does better than much larger models (such as GIT2, PaLI and Flamingo) on these tasks. Extensive ablations show that due to its novel pre-training tasks, DocFormerv2 understands multiple modalities better than prior-art in VDU.

PDF Details DOI

IJCAI Conference 2024 Conference Paper

Feature Norm Regularized Federated Learning: Utilizing Data Disparities for Model Performance Gains

Ke Hu
Liyao Xiang
Peng Tang
Weidong Qiu

Federated learning (FL) is a machine learning paradigm that aggregates knowledge and utilizes computational power from multiple participants to train a global model. However, a commonplace challenge—non-independent and identically distributed (non-i. i. d. ) data across participants—can lead to significant divergence in model updates, thus diminishing training efficacy. In this paper, we propose the Feature Norm Regularized Federated Learning (FNR-FL) algorithm to tackle the non-i. i. d challenge. FNR-FL incorporates class average feature norms into the loss function by a straightforward yet effective regularization strategy. The core idea of FNR-FL is to penalize the deviations in the update directions of local models caused by the non-i. i. d data. Theoretically, we provide convergence guarantees for FNR-FL when training under non-i. i. d scenarios. Practically, our comprehensive experimental evaluations demonstrate that FNR-FL significantly outperforms existing FL algorithms in terms of test accuracy, and maintains a competitive convergence rate with lower communication overhead and shorter duration. Compared to FedAvg, FNR-FL exhibits a 66. 24% improvement in accuracy and an 11. 40% reduction in training time, underscoring its enhanced effectiveness and efficiency. The code is available on GitHub at: https: //github. com/LonelyMoonDesert/FNR-FL.

PDF Details DOI

AAMAS Conference 2024 Conference Paper

Fuzzy Clustered Federated Learning Under Mixed Data Distributions

Peng Tang
Lifan Wang
Weidong Qiu
Zheng Huang
Qiangmin Wang

Federated learning (FL) is deeply troubled by non-independent and identically distributed (non-IID) data, leading to suboptimal training results. Clustered FL partitions clients’ unique data into different clusters to reduce the heterogeneity among clients. Current approaches are unable to eliminate the impact of data heterogeneity and provide personalized models to client devices. By assuming the clients’ data can be divided into different data distributions, we propose a novel fuzzy clustered FL method. We partition the client’s data and generate a personalized model for each client. The experiments demonstrated that our method achieved excellent results. In the case of 𝑁 clusters, our method achieved a communication cost reduction of 1/𝑁 compared to the SOTA methods, while improving performance by 10. 4% on CIFAR-10.

PDF

AAAI Conference 2024 Conference Paper

No Head Left Behind – Multi-Head Alignment Distillation for Transformers

Tianyang Zhao
Kunwar Yashraj Singh
Srikar Appalaraju
Peng Tang
Vijay Mahadevan
R. Manmatha
Ying Nian Wu

Knowledge distillation aims at reducing model size without compromising much performance. Recent work has applied it to large vision-language (VL) Transformers, and has shown that attention maps in the multi-head attention modules of vision-language Transformers contain extensive intra-modal and cross-modal co-reference relations to be distilled. The standard approach is to apply a one-to-one attention map distillation loss, i.e. the Teacher's first attention head instructs the Student's first head, the second teaches the second, and so forth, but this only works when the numbers of attention heads in the Teacher and Student are the same. To remove this constraint, we propose a new Attention Map Alignment Distillation (AMAD) method for Transformers with multi-head attention, which works for a Teacher and a Student with different numbers of attention heads. Specifically, we soft-align different heads in Teacher and Student attention maps using a cosine similarity weighting. The Teacher head contributes more to the Student heads for which it has a higher similarity weight. Each Teacher head contributes to all the Student heads by minimizing the divergence between the attention activation distributions for the soft-aligned heads. No head is left behind. This distillation approach operates like cross-attention. We experiment on distilling VL-T5 and BLIP, and apply AMAD loss on their T5, BERT, and ViT sub-modules. We show, under vision-language setting, that AMAD outperforms conventional distillation methods on VQA-2.0, COCO captioning, and Multi30K translation datasets. We further show that even without VL pre-training, the distilled VL-T5 models outperform corresponding VL pre-trained VL-T5 models that are further fine-tuned by ground-truth signals, and that fine-tuning distillation can also compensate to some degree for the absence of VL pre-training for BLIP models.

PDF Details DOI

AAAI Conference 2024 Conference Paper

Prior and Prediction Inverse Kernel Transformer for Single Image Defocus Deblurring

Peng Tang
Zhiqiang Xu
Chunlai Zhou
Pengfei Wei
Peng Han
Xin Cao
Tobias Lasser

Defocus blur, due to spatially-varying sizes and shapes, is hard to remove. Existing methods either are unable to effectively handle irregular defocus blur or fail to generalize well on other datasets. In this work, we propose a divide-and-conquer approach to tackling this issue, which gives rise to a novel end-to-end deep learning method, called prior-and-prediction inverse kernel transformer (P2IKT), for single image defocus deblurring. Since most defocus blur can be approximated as Gaussian blur or its variants, we construct an inverse Gaussian kernel module in our method to enhance its generalization ability. At the same time, an inverse kernel prediction module is introduced in order to flexibly address the irregular blur that cannot be approximated by Gaussian blur. We further design a scale recurrent transformer, which estimates mixing coefficients for adaptively combining the results from the two modules and runs the scale recurrent ``coarse-to-fine" procedure for progressive defocus deblurring. Extensive experimental results demonstrate that our P2IKT outperforms previous methods in terms of PSNR on multiple defocus deblurring datasets.

PDF Details DOI

JBHI Journal 2022 Journal Article

Efficient, Revocable, and Privacy-Preserving Fine-Grained Data Sharing With Keyword Search for the Cloud-Assisted Medical IoT System

Yangyang Bao
Weidong Qiu
Peng Tang
Xiaochun Cheng

The cloud-assisted medical Internet of Things (MIoT) has played a revolutionary role in promoting the quality of public medical services. However, the practical deployment of cloud-assisted MIoT in an open healthcare scenario raises the concern on data security and user’s privacy. Despite endeavors by academic and industrial community to eliminate this concern by cryptographic methods, resource-constrained devices in MIoT may be subject to the heavy computational overheads of cryptographic computations. To address this issue, this paper proposes an efficient, revocable, privacy-preserving fine-grained data sharing with keyword search (ERPF-DS-KS) scheme, which realizes the efficient and fine-grained access control and ciphertext keyword search, and enables the flexible indirect revocation to malicious data users. A pseudo identity-based signature mechanism is designed to provide the data authenticity. We analyze the security properties of our proposed scheme, and via the theoretical comparison and experimental results we demonstrate that for the resource-constrained devices in the patient and doctor side of MIoT, in comparison with other related schemes, ERPF-DS-KS just consumes the lightweight and constant size communication/storage as well as computational time cost. For the keyword search, compared with related schemes, the cloud can quickly check whether a ciphertext contains the specified keyword with slight computations in the online phase. This further demonstrates that ERPF-DS-KS is efficient and practical in the cloud-assisted MIoT scenario.

Details DOI

ICLR Conference 2021 Conference Paper

Shape and texture are two prominent and complementary cues for recognizing objects. Nonetheless, Convolutional Neural Networks are often biased towards either texture or shape, depending on the training dataset. Our ablation shows that such bias degenerates model performance. Motivated by this observation, we develop a simple algorithm for shape-texture debiased learning. To prevent models from exclusively attending on a single cue in representation learning, we augment training data with images with conflicting shape and texture information (eg, an image of chimpanzee shape but with lemon texture) and, most importantly, provide the corresponding supervisions from shape and texture simultaneously. Experiments show that our method successfully improves model performance on several image recognition benchmarks and adversarial robustness. For example, by training on ImageNet, it helps ResNet-152 achieve substantial improvements on ImageNet (+1.2%), ImageNet-A (+5.2%), ImageNet-C (+8.3%) and Stylized-ImageNet (+11.1%), and on defending against FGSM adversarial attacker on ImageNet (+14.4%). Our method also claims to be compatible with other advanced data augmentation strategies, eg, Mixup, and CutMix. The code is available here: https://github.com/LiYingwei/ShapeTextureDebiasedTraining.

Details

JBHI Journal 2020 Journal Article

GP-CNN-DTEL: Global-Part CNN Model With Data-Transformed Ensemble Learning for Skin Lesion Classification

Peng Tang
Qiaokang Liang
Xintong Yan
Shao Xiang
Dan Zhang

Precise skin lesion classification is still challenging due to two problems, i. e. , (1) inter-class similarity and intra-class variation of skin lesion images, and (2) the weak generalization ability of single Deep Convolutional Neural Network trained with limited data. Therefore, we propose a Global-Part Convolutional Neural Network (GP-CNN) model, which treats the fine-grained local information and global context information with equal importance. The Global-Part model consists of a Global Convolutional Neural Network (G-CNN) and a Part Convolutional Neural Network (P-CNN). Specifically, the G-CNN is trained with downscaled dermoscopy images, and is used to extract the global-scale information of dermoscopy images and produce the Classification Activation Map (CAM). While the P-CNN is trained with the CAM guided cropped image patches and is used to capture local-scale information of skin lesion regions. Additionally, we present a data-transformed ensemble learning strategy, which can further boost the classification performance by integrating the different discriminant information from GP-CNNs that are trained with original images, color constancy transformed images, and feature saliency transformed images, respectively. The proposed method is evaluated on the ISIC 2016 and ISIC 2017 Skin Lesion Challenge (SLC) classification datasets. Experimental results indicate that the proposed method can achieve the state-of-the-art skin lesion classification performance (i. e. , an AP value of 0. 718 on the ISIC 2016 SLC dataset and an Average Auc value of 0. 926 on the ISIC 2017 SLC dataset) without any external data, compared with other current methods which need to use external data.

Details DOI

EAAI Journal 2015 Journal Article

A lattice-based approach for chemical structural retrieval

Peng Tang
Siu Cheung Hui
Alvis C.M. Fong

Searching for chemical structures with similar structural and functional information of organic chemicals is an important part of the drug discovery process. However, the current chemical structural retrieval methods have focused mainly on finding chemicals with similar structures to the input chemical structural query, and tend to ignore the functional features which are important for determining the chemical property and activity of the chemicals. In this paper, we propose a lattice-based approach for chemical structural retrieval. The proposed lattice-based approach is based on Formal Concept Analysis. It retrieves chemical structures that have functional groups and interactions between functional groups similar to the chemical structural query. The performance of the proposed lattice-based approach is evaluated and its promising performance results have shown that the proposed approach is effective for chemical structural retrieval.

Details DOI

IROS Conference 2007 Conference Paper

Automatic generation of contact state graphs between a polygon and a planar kinematic chain

Peng Tang
Jing Xiao 0001

Information of high-level, topological contact states is useful and sometimes even necessary for a wide range of robotic tasks involving interactions between a robot and its environment or objects of manipulation. While most of the existing research is focused on contact states between two rigid bodies, this paper presents a practical approach to represent concisely and generate automatically graphs of contact states between a polygonal object and an articulated planar object, i. e. , a planar kinematic chain. The approach effectively exploits topological and geometrical constraints associated with such contact states to ensure both correctness and efficiency, as demonstrated by the implementation and applied examples.

Details

ICRA Conference 2006 Conference Paper

Automatic Generation of Contact State Graphs based on Curvature Monotonic Segmentation

Peng Tang
Jing Xiao 0001

This paper addresses representation and automatic generation of topological contact states between two 3-D curved objects. Information of contact states is useful for a wide range of applications, from robotic tasks involving compliant motion to virtual prototyping and simulation. As contact states between 3D curved objects are more commonplace and yet more complex and less studied, this work is more necessary. The approach has been implemented, and the implemented examples demonstrate the effectiveness of the approach

Details

ICRA Conference 2005 Conference Paper

Automatic Generation of High-level Contact State Space between Planar Curved Objects

Peng Tang
Jing Xiao 0001

Information of high-level, topological contact states is useful and even necessary for a wide range of applications, including many robotics applications. While there is considerable research related to topological contact states between two polyhedral objects, little is studied about how to characterize, represent, and automatically generate topological contact states between curved objects. In this paper we extend the representation of topological contact states between polyhedral objects to general planar curved objects in terms of contacting curve elements, obtained from curvature monotonic segmentation [6]. We further introduce an approach to generate automatically graphs of such contact states between two planar curved objects, which represent not only valid contact states but also adjacency relations among those contact states. Implementation results of the related algorithms demonstrate the effectiveness of our approach. The approach can be naturally extended to generation of contact states between 3-D curved objects.

Details