Yancheng Wang

Yancheng Wang, Ziyan Jiang, Zheng Chen et al. · amazon-science

While the recommendation system (RS) has advanced significantly through deep learning, current RS approaches usually train and fine-tune models on task-specific datasets, limiting their generalizability to new recommendation tasks and their ability to leverage external knowledge due to model scale and data size constraints. Thus, we designed an LLM-powered autonomous recommender agent, RecMind, which is capable of leveraging external knowledge, utilizing tools with careful planning to provide zero-shot personalized recommendations. We propose a Self-Inspiring algorithm to improve the planning ability. At each intermediate step, the LLM self-inspires to consider all previously explored states to plan for the next step. This mechanism greatly improves the model's ability to comprehend and utilize historical information in planning for recommendation. We evaluate RecMind's performance in various recommendation scenarios. Our experiment shows that RecMind outperforms existing zero/few-shot LLM-based recommendation baseline methods in various tasks and achieves comparable performance to a fully trained recommendation model P5.

Yancheng Wang, Ning Xu, Yingzhen Yang

Non-local attention module has been proven to be crucial for image restoration. Conventional non-local attention processes features of each layer separately, so it risks missing correlation between features among different layers. To address this problem, we aim to design attention modules that aggregate information from different layers. Instead of finding correlated key pixels within the same layer, each query pixel is encouraged to attend to key pixels at multiple previous layers of the network. In order to efficiently embed such attention design into neural network backbones, we propose a novel Adaptive Cross-Layer Attention (ACLA) module. Two adaptive designs are proposed for ACLA: (1) adaptively selecting the keys for non-local attention at each layer; (2) automatically searching for the insertion locations for ACLA modules. By these two adaptive designs, ACLA dynamically selects a flexible number of keys to be aggregated for non-local attention at previous layer while maintaining a compact neural network with compelling performance. Extensive experiments on image restoration tasks, including single image super-resolution, image denoising, image demosaicing, and image compression artifacts reduction, validate the effectiveness and efficiency of ACLA. The code of ACLA is available at \url{https://github.com/SDL-ASU/ACLA}.

Yancheng Wang, Yingzhen Yang

Self-attention and transformers have been widely used in deep learning. Recent efforts have been devoted to incorporating transformer blocks into different neural architectures, including those with convolutions, leading to various visual transformers for computer vision tasks. In this paper, we propose a novel and compact transformer block, Transformer with Learnable Token Merging (LTM), or LTM-Transformer. LTM-Transformer performs token merging in a learnable scheme. LTM-Transformer is compatible with many popular and compact transformer networks, and it reduces the FLOPs and the inference time of the visual transformers while maintaining or even improving the prediction accuracy. In the experiments, we replace all the transformer blocks in popular visual transformers, including MobileViT, EfficientViT, ViT, and Swin, with LTM-Transformer blocks, leading to LTM-Transformer networks with different backbones. The LTM-Transformer is motivated by reduction of Information Bottleneck, and a novel and separable variational upper bound for the IB loss is derived. The architecture of the mask module in our LTM blocks, which generates the token merging mask, is designed to reduce the derived upper bound for the IB loss. Extensive results on computer vision tasks evidence that LTM-Transformer renders compact and efficient visual transformers with comparable or much better prediction accuracy than the original visual transformers. The code of the LTM-Transformer is available at https://github.com/Statistical-Deep-Learning/LTM}

Yancheng Wang, Yingzhen Yang

Graph Neural Networks (GNNs) have been widely used to learn node representations and with outstanding performance on various tasks such as node classification. However, noise, which inevitably exists in real-world graph data, would considerably degrade the performance of GNNs as the noise is easily propagated via the graph structure. In this work, we propose a novel and robust method, Bayesian Robust Graph Contrastive Learning (BRGCL), which trains a GNN encoder to learn robust node representations. The BRGCL encoder is a completely unsupervised encoder. Two steps are iteratively executed at each epoch of training the BRGCL encoder: (1) estimating confident nodes and computing robust cluster prototypes of node representations through a novel Bayesian nonparametric method; (2) prototypical contrastive learning between the node representations and the robust cluster prototypes. Experiments on public and large-scale benchmarks demonstrate the superior performance of BRGCL and the robustness of the learned node representations. The code of BRGCL is available at \url{https://github.com/BRGCL-code/BRGCL-code}.

Yu Cao, Yancheng Wang, Yifei Xue et al.

Segmentation from point cloud data is essential in many applications such as remote sensing, mobile robots, or autonomous cars. However, the point clouds captured by the 3D range sensor are commonly sparse and unstructured, challenging efficient segmentation. In this paper, we present a fast solution to point cloud instance segmentation with small computational demands. To this end, we propose a novel fast Euclidean clustering (FEC) algorithm which applies a pointwise scheme over the clusterwise scheme used in existing works. Our approach is conceptually simple, easy to implement (40 lines in C++), and achieves two orders of magnitudes faster against the classical segmentation methods while producing high-quality results.

Haoyang Xie, Xinyuan Wang, Yancheng Wang et al.

Large language model (LLM) agents are increasingly used to operate browsers, files, code and tools, making personal assistants a natural deployment target. Yet personal agents face a privacy-cost-capability tension: cloud models execute multi-step workflows well but expose sensitive intermediate context to external APIs, while local models preserve privacy but remain less reliable. Both settings also pay repeatedly for long skill prompts and growing histories. We propose constant-context skill learning, a context-to-weights framework for recurring agent workflows: reusable procedures are learned in lightweight task-family modules, while inference conditions only on the current observation and a compact state block. A deterministic tracker renders this state block from task progress and supplies aligned subgoal rewards, so each module can be trained with step-level SFT and refined through online RL. Across ALFWorld, WebShop, and SciWorld, our agents achieve strong performance across Qwen3-4B, Qwen3-8B and Llama-3.1-8B. With Qwen3-8B, SFT+RL reaches 89.6\% unseen success on ALFWorld, 76.8\% success on WebShop, and 66.4\% unseen success on SciWorld. They match or exceed strong published agent-training results while reducing prompt tokens per turn by 2--7$\times$ relative to controlled ReAct prompting baselines, showing that procedural context can be moved from prompts into weights.

Hong Guan, Yancheng Wang, Lulu Xie et al.

Effective fraud detection and analysis of government-issued identity documents, such as passports, driver's licenses, and identity cards, are essential in thwarting identity theft and bolstering security on online platforms. The training of accurate fraud detection and analysis tools depends on the availability of extensive identity document datasets. However, current publicly available benchmark datasets for identity document analysis, including MIDV-500, MIDV-2020, and FMIDV, fall short in several respects: they offer a limited number of samples, cover insufficient varieties of fraud patterns, and seldom include alterations in critical personal identifying fields like portrait images, limiting their utility in training models capable of detecting realistic frauds while preserving privacy. In response to these shortcomings, our research introduces a new benchmark dataset, IDNet, designed to advance privacy-preserving fraud detection efforts. The IDNet dataset comprises 837,060 images of synthetically generated identity documents, totaling approximately 490 gigabytes, categorized into 20 types from $10$ U.S. states and 10 European countries. We evaluate the utility and present use cases of the dataset, illustrating how it can aid in training privacy-preserving fraud detection methods, facilitating the generation of camera and video capturing of identity documents, and testing schema unification and other identity document management functionalities.

Yancheng Wang, Osama Hanna, Ruiming Xie et al.

Emotion recognition in speech presents a complex multimodal challenge, requiring comprehension of both linguistic content and vocal expressivity, particularly prosodic features such as fundamental frequency, intensity, and temporal dynamics. Although large language models (LLMs) have shown promise in reasoning over textual transcriptions for emotion recognition, they typically neglect fine-grained prosodic information, limiting their effectiveness and interpretability. In this work, we propose VowelPrompt, a linguistically grounded framework that augments LLM-based emotion recognition with interpretable, fine-grained vowel-level prosodic cues. Drawing on phonetic evidence that vowels serve as primary carriers of affective prosody, VowelPrompt extracts pitch-, energy-, and duration-based descriptors from time-aligned vowel segments, and converts these features into natural language descriptions for better interpretability. Such a design enables LLMs to jointly reason over lexical semantics and fine-grained prosodic variation. Moreover, we adopt a two-stage adaptation procedure comprising supervised fine-tuning (SFT) followed by Reinforcement Learning with Verifiable Reward (RLVR), implemented via Group Relative Policy Optimization (GRPO), to enhance reasoning capability, enforce structured output adherence, and improve generalization across domains and speaker variations. Extensive evaluations across diverse benchmark datasets demonstrate that VowelPrompt consistently outperforms state-of-the-art emotion recognition methods under zero-shot, fine-tuned, cross-domain, and cross-linguistic conditions, while enabling the generation of interpretable explanations that are jointly grounded in contextual semantics and fine-grained prosodic structure.

Utkarsh Nath, Yancheng Wang, Yingzhen Yang

Deep Neural Networks are vulnerable to adversarial attacks. Neural Architecture Search (NAS), one of the driving tools of deep neural networks, demonstrates superior performance in prediction accuracy in various machine learning applications. However, it is unclear how it performs against adversarial attacks. Given the presence of a robust teacher, it would be interesting to investigate if NAS would produce robust neural architecture by inheriting robustness from the teacher. In this paper, we propose Robust Neural Architecture Search by Cross-Layer Knowledge Distillation (RNAS-CL), a novel NAS algorithm that improves the robustness of NAS by learning from a robust teacher through cross-layer knowledge distillation. Unlike previous knowledge distillation methods that encourage close student/teacher output only in the last layer, RNAS-CL automatically searches for the best teacher layer to supervise each student layer. Experimental result evidences the effectiveness of RNAS-CL and shows that RNAS-CL produces small and robust neural architecture.

Yancheng Wang, Changyu Liu, Yingzhen Yang

Graph diffusion models have recently been proposed to synthesize entire graphs, such as molecule graphs. Although existing methods have shown great performance in generating entire graphs for graph-level learning tasks, no graph diffusion models have been developed to generate synthetic graph structures, that is, synthetic nodes and associated edges within a given graph, for node-level learning tasks. Inspired by the research in the computer vision literature using synthetic data for enhanced performance, we propose Diffusion on Graph (DoG), which generates synthetic graph structures to boost the performance of GNNs. The synthetic graph structures generated by DoG are combined with the original graph to form an augmented graph for the training of node-level learning tasks, such as node classification and graph contrastive learning (GCL). To improve the efficiency of the generation process, a Bi-Level Neighbor Map Decoder (BLND) is introduced in DoG. To mitigate the adverse effect of the noise introduced by the synthetic graph structures, a low-rank regularization method is proposed for the training of graph neural networks (GNNs) on the augmented graphs. Extensive experiments on various graph datasets for semi-supervised node classification and graph contrastive learning have been conducted to demonstrate the effectiveness of DoG with low-rank regularization. The code of DoG is available at https://github.com/Statistical-Deep-Learning/DoG.

Yancheng Wang, Yingzhen Yang

Graph Neural Networks (GNNs) have achieved remarkable success in learning node representations and have shown strong performance in tasks such as node classification. However, recent findings indicate that the presence of noise in real-world graph data can substantially impair the effectiveness of GNNs. To address this challenge, we introduce a robust and innovative node representation learning method named Graph Contrastive Learning with Low-Rank Regularization, or GCL-LRR, which follows a two-stage transductive learning framework for node classification. In the first stage, the GCL-LRR encoder is optimized through prototypical contrastive learning while incorporating a low-rank regularization objective. In the second stage, the representations generated by GCL-LRR are employed by a linear transductive classifier to predict the labels of unlabeled nodes within the graph. Our GCL-LRR is inspired by the Low Frequency Property (LFP) of the graph data and its labels, and it is also theoretically motivated by our sharp generalization bound for transductive learning. To the best of our knowledge, our theoretical result is among the first to theoretically demonstrate the advantage of low-rank regularization in transductive learning, which is also supported by strong empirical results. To further enhance the performance of GCL-LRR, we present an improved model named GCL-LR-Attention, which incorporates a novel LR-Attention layer into GCL-LRR. GCL-LR-Attention reduces the kernel complexity of GCL-LRR and contributes to a tighter generalization bound, leading to improved performance. Extensive evaluations on standard benchmark datasets evidence the effectiveness and robustness of both GCL-LRR and GCL-LR-Attention in learning meaningful node representations. The code is available at https://github.com/Statistical-Deep-Learning/GCL-LR-Attention.

Yancheng Wang, Nebojsa Jojic, Yingzhen Yang

In this paper, we propose a novel attention module termed the Differentiable Channel Selection Attention module, or the DCS-Attention module. In contrast with conventional self-attention, the DCS-Attention module features selection of informative channels in the computation of the attention weights. The selection of the feature channels is performed in a differentiable manner, enabling seamless integration with DNN training. Our DCS-Attention is compatible with either fixed neural network backbones or learnable backbones with Differentiable Neural Architecture Search (DNAS), leading to DCS with Fixed Backbone (DCS-FB) and DCS-DNAS, respectively. Importantly, our DCS-Attention is motivated by the principle of Information Bottleneck (IB), and a novel variational upper bound for the IB loss, which can be optimized by SGD, is derived and incorporated into the training loss of the networks with the DCS-Attention modules. In this manner, a neural network with DCS-Attention modules is capable of selecting the most informative channels for feature extraction so that it enjoys state-of-the-art performance for the Re-ID task. Extensive experiments on multiple person Re-ID benchmarks using both DCS-FB and DCS-DNAS show that DCS-Attention significantly enhances the prediction accuracy of DNNs for person Re-ID, which demonstrates the effectiveness of DCS-Attention in learning discriminative features critical to identifying person identities. The code of our work is available at https://github.com/Statistical-Deep-Learning/DCS-Attention.

Kaize Ding, Yancheng Wang, Yingzhen Yang et al.

Graph Contrastive Learning (GCL) has recently drawn much research interest for learning generalizable node representations in a self-supervised manner. In general, the contrastive learning process in GCL is performed on top of the representations learned by a graph neural network (GNN) backbone, which transforms and propagates the node contextual information based on its local neighborhoods. However, nodes sharing similar characteristics may not always be geographically close, which poses a great challenge for unsupervised GCL efforts due to their inherent limitations in capturing such global graph knowledge. In this work, we address their inherent limitations by proposing a simple yet effective framework -- Simple Neural Networks with Structural and Semantic Contrastive Learning} (S^3-CL). Notably, by virtue of the proposed structural and semantic contrastive learning algorithms, even a simple neural network can learn expressive node representations that preserve valuable global structural and semantic patterns. Our experiments demonstrate that the node representations learned by S^3-CL achieve superior performance on different downstream tasks compared with the state-of-the-art unsupervised GCL methods. Implementation and more experimental details are publicly available at \url{https://github.com/kaize0409/S-3-CL.}

Yancheng Wang, Yang Xiao, Fu Xiong et al.

To facilitate depth-based 3D action recognition, 3D dynamic voxel (3DV) is proposed as a novel 3D motion representation. With 3D space voxelization, the key idea of 3DV is to encode 3D motion information within depth video into a regular voxel set (i.e., 3DV) compactly, via temporal rank pooling. Each available 3DV voxel intrinsically involves 3D spatial and motion feature jointly. 3DV is then abstracted as a point set and input into PointNet++ for 3D action recognition, in the end-to-end learning way. The intuition for transferring 3DV into the point set form is that, PointNet++ is lightweight and effective for deep feature learning towards point set. Since 3DV may lose appearance clue, a multi-stream 3D action recognition manner is also proposed to learn motion and appearance feature jointly. To extract richer temporal order information of actions, we also divide the depth video into temporal splits and encode this procedure in 3DV integrally. The extensive experiments on 4 well-established benchmark datasets demonstrate the superiority of our proposition. Impressively, we acquire the accuracy of 82.4% and 93.5% on NTU RGB+D 120 [13] with the cross-subject and crosssetup test setting respectively. 3DV's code is available at https://github.com/3huo/3DV-Action.

Chengshuai Zhao, Zhen Tan, Pingchuan Ma et al.

Chain-of-Thought (CoT) prompting has been shown to improve Large Language Model (LLM) performance on various tasks. With this approach, LLMs appear to produce human-like reasoning steps before providing answers (a.k.a., CoT reasoning), which often leads to the perception that they engage in deliberate inferential processes. However, some initial findings suggest that CoT reasoning may be more superficial than it appears, motivating us to explore further. In this paper, we study CoT reasoning via a data distribution lens and investigate if CoT reasoning reflects a structured inductive bias learned from in-distribution data, allowing the model to conditionally generate reasoning paths that approximate those seen during training. Thus, its effectiveness is fundamentally bounded by the degree of distribution discrepancy between the training data and the test queries. With this lens, we dissect CoT reasoning via three dimensions: task, length, and format. To investigate each dimension, we design DataAlchemy, an isolated and controlled environment to train LLMs from scratch and systematically probe them under various distribution conditions. Our results reveal that CoT reasoning is a brittle mirage that vanishes when it is pushed beyond training distributions. This work offers a deeper understanding of why and when CoT reasoning fails, emphasizing the ongoing challenge of achieving genuine and generalizable reasoning.

Rajeev Goel, Utkarsh Nath, Yancheng Wang et al.

Deep neural networks, including Convolutional Neural Networks (CNNs) and Visual Transformers (ViT), have achieved stunning success in medical image domain. We study thorax disease classification in this paper. Effective extraction of features for the disease areas is crucial for disease classification on radiographic images. While various neural architectures and training techniques, such as self-supervised learning with contrastive/restorative learning, have been employed for disease classification on radiographic images, there are no principled methods which can effectively reduce the adverse effect of noise and background, or non-disease areas, on the radiographic images for disease classification. To address this challenge, we propose a novel Low-Rank Feature Learning (LRFL) method in this paper, which is universally applicable to the training of all neural networks. The LRFL method is both empirically motivated by the low frequency property observed on all the medical datasets in this paper, and theoretically motivated by our sharp generalization bound for neural networks with low-rank features. In the empirical study, using a neural network such as a ViT or a CNN pre-trained on unlabeled chest X-rays by Masked Autoencoders (MAE), our novel LRFL method is applied on the pre-trained neural network and demonstrate better classification results in terms of both multiclass area under the receiver operating curve (mAUC) and classification accuracy.

Yancheng Wang, Yingzhen Yang

Self-attention and transformer architectures have become foundational components in modern deep learning. Recent efforts have integrated transformer blocks into compact neural architectures for computer vision, giving rise to various efficient vision transformers. In this work, we introduce Transformer with Kernel Complexity Reduction, or KCR-Transformer, a compact transformer block equipped with differentiable channel selection, guided by a novel and sharp theoretical generalization bound. KCR-Transformer performs input/output channel selection in the MLP layers of transformer blocks to reduce the computational cost. Furthermore, we provide a rigorous theoretical analysis establishing a tight generalization bound for networks equipped with KCR-Transformer blocks. Leveraging such strong theoretical results, the channel pruning by KCR-Transformer is conducted in a generalization-aware manner, ensuring that the resulting network retains a provably small generalization error. Our KCR-Transformer is compatible with many popular and compact transformer networks, such as ViT and Swin, and it reduces the FLOPs of the vision transformers while maintaining or even improving the prediction accuracy. In the experiments, we replace all the transformer blocks in the vision transformers with KCR-Transformer blocks, leading to KCR-Transformer networks with different backbones. The resulting TCR-Transformers achieve superior performance on various computer vision tasks, achieving even better performance than the original models with even less FLOPs and parameters.

Hong Guan, Ansh Tiwari, Summer Gautier et al.

Detecting inference queries running over personal attributes and protecting such queries from leaking individual information requires tremendous effort from practitioners. To tackle this problem, we propose an end-to-end workflow for automating privacy-preserving inference queries including the detection of subqueries that involve AI/ML model inferences on sensitive attributes. Our proposed novel declarative privacy-preserving workflow allows users to specify "what private information to protect" rather than "how to protect". Under the hood, the system automatically chooses privacy-preserving plans and hyper-parameters.

Fu Xiong, Yang Xiao, Zhiguo Cao et al.

Face verification can be regarded as a 2-class fine-grained visual recognition problem. Enhancing the feature's discriminative power is one of the key problems to improve its performance. Metric learning technology is often applied to address this need, while achieving a good tradeoff between underfitting and overfitting plays the vital role in metric learning. Hence, we propose a novel ensemble cascade metric learning (ECML) mechanism. In particular, hierarchical metric learning is executed in the cascade way to alleviate underfitting. Meanwhile, at each learning level, the features are split into non-overlapping groups. Then, metric learning is executed among the feature groups in the ensemble manner to resist overfitting. Considering the feature distribution characteristics of faces, a robust Mahalanobis metric learning method (RMML) with closed-form solution is additionally proposed. It can avoid the computation failure issue on inverse matrix faced by some well-known metric learning approaches (e.g., KISSME). Embedding RMML into the proposed ECML mechanism, our metric learning paradigm (EC-RMML) can run in the one-pass learning manner. Experimental results demonstrate that EC-RMML is superior to state-of-the-art metric learning methods for face verification. And, the proposed ensemble cascade metric learning mechanism is also applicable to other metric learning approaches.

Yang Xiao, Jun Chen, Yancheng Wang et al.

Dynamic imaging is a recently proposed action description paradigm for simultaneously capturing motion and temporal evolution information, particularly in the context of deep convolutional neural networks (CNNs). Compared with optical flow for motion characterization, dynamic imaging exhibits superior efficiency and compactness. Inspired by the success of dynamic imaging in RGB video, this study extends it to the depth domain. To better exploit three-dimensional (3D) characteristics, multi-view dynamic images are proposed. In particular, the raw depth video is densely projected with respect to different virtual imaging viewpoints by rotating the virtual camera within the 3D space. Subsequently, dynamic images are extracted from the obtained multi-view depth videos and multi-view dynamic images are thus constructed from these images. Accordingly, more view-tolerant visual cues can be involved. A novel CNN model is then proposed to perform feature learning on multi-view dynamic images. Particularly, the dynamic images from different views share the same convolutional layers but correspond to different fully connected layers. This is aimed at enhancing the tuning effectiveness on shallow convolutional layers by alleviating the gradient vanishing problem. Moreover, as the spatial occurrence variation of the actions may impair the CNN, an action proposal approach is also put forth. In experiments, the proposed approach can achieve state-of-the-art performance on three challenging datasets.