检索结果-内蒙古大学图书馆

Uncovering the Impact of Chain-of-Thought Reasoning for Direct Preference Optimization: Lessons from Text-to-SQL

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Liu, Hanbing Li, Haoyang Zhang, Xiaokang Chen, Ruotong Xu, Haiyong Tian, Tian Qi, Qi Zhang, Jing Gaoling School of Artificial Intelligence Renmin University of China Beijing China School of Information Renmin University of China Beijing China Key Laboratory of Data Engineering and Knowledge Engineering Beijing China Engineering Research Center of Database and Business Intelligence Beijing China China Mobile Information Technology Center China

Direct Preference Optimization (DPO) has proven effective in complex reasoning tasks like math word problems and code generation. However, when applied to Text-to-SQL datasets, it often fails to improve performance and can even degrade it. Our investigation reveals the root cause: unlike math and code tasks, which naturally integrate Chain-of-Thought (CoT) reasoning with DPO, Text-to-SQL datasets typically include only final answers (gold SQL queries) without detailed CoT solutions. By augmenting Text-to-SQL datasets with synthetic CoT solutions, we achieve, for the first time, consistent and significant performance improvements using DPO. Our analysis shows that CoT reasoning is crucial for unlocking DPO’s potential, as it mitigates reward hacking, strengthens discriminative capabilities, and improves scalability. These findings offer valuable insights for building more robust Text-to-SQL models. To support further research, we publicly release the code and CoT-enhanced datasets © 2025, CC BY.

关键词： Chains

Causal-Inspired Multitask Learning for Video-Based Human Pose Estimation

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Chen, Haipeng Wu, Sifan Wang, Zhigang Yin, Yifang Jiao, Yingying Lyu, Yingda Liu, Zhenguang College of Computer Science and Technology Jilin University China Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin University China College of Computer Science and Technology Zhejiang Gongshang University China A*STAR Singapore Public Computer Education and Research Center Jilin University China The State Key Laboratory of Blockchain and Data Security Zhejiang University China Institute of Blockchain and Data Security China

Video-based human pose estimation has long been a fundamental yet challenging problem in computer vision. Previous studies focus on spatio-temporal modeling through the enhancement of architecture design and optimization strategies. However, they overlook the causal relationships in the joints, leading to models that may be overly tailored and thus estimate poorly to challenging scenes. Therefore, adequate causal reasoning capability, coupled with good interpretability of model, are both indispensable and prerequisite for achieving reliable results. In this paper, we pioneer a causal perspective on pose estimation and introduce a causal-inspired multitask learning framework, consisting of two stages. In the first stage, we try to endow the model with causal spatio-temporal modeling ability by introducing two self-supervision auxiliary tasks. Specifically, these auxiliary tasks enable the network to infer challenging keypoints based on observed keypoint information, thereby imbuing causal reasoning capabilities into the model and making it robust to challenging scenes. In the second stage, we argue that not all feature tokens contribute equally to pose estimation. Prioritizing causal (keypoint-relevant) tokens is crucial to achieve reliable results, which could improve the interpretability of the model. To this end, we propose a Token Causal Importance Selection module to identify the causal tokens and non-causal tokens (e.g., background and objects). Additionally, non-causal tokens could provide potentially beneficial cues but may be redundant. We further introduce a non-causal tokens clustering module to merge the similar non-causal tokens. Extensive experiments show that our method outperforms state-of-the-art methods on three large-scale benchmark datasets. Copyright © 2025, The Authors. All rights reserved.

关键词： Self-supervised learning

UAV-assisted Joint Mobile Edge Computing and data Collection via Matching-enabled Deep Reinforcement Learning

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Wang, Boxiong Kang, Hui Li, Jiahui Sun, Geng Sun, Zemin Wang, Jiacheng Niyato, Dusit College of Computer Science and Technology Jilin University Changchun130012 China Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education Jilin University Changchun130012 China College of Computing and Data Science Nanyang Technological University 639798 Singapore College of Computing and Data Science Nanyang Technological University Singapore

Unmanned aerial vehicle (UAV)-assisted mobile edge computing (MEC) and data collection (DC) have been popular research issues. Different from existing works that consider MEC and DC scenarios separately, this paper investigates a multi-UAV-assisted joint MEC-DC system. Specifically, we formulate a joint optimization problem to minimize the MEC latency and maximize the collected data volume. This problem can be classified as a non-convex mixed integer programming problem that exhibits long-term optimization and dynamics. Thus, we propose a deep reinforcement learning-based approach that jointly optimizes the UAV movement, user transmit power, and user association in real time to solve the problem efficiently. Specifically, we reformulate the optimization problem into an action space-reduced Markov decision process (MDP) and optimize the user association by using a two-phase matching-based association (TMA) strategy. Subsequently, we propose a soft actor-critic (SAC)-based approach that integrates the proposed TMA strategy (SAC-TMA) to solve the formulated joint optimization problem collaboratively. Simulation results demonstrate that the proposed SAC-TMA is able to coordinate the two subsystems and can effectively reduce the system latency and improve the data collection volume compared with other benchmark algorithms. Copyright © 2025, The Authors. All rights reserved.

关键词： Unmanned aerial vehicles (UAV)

xJailbreak: Representation Space Guided Reinforcement Learning for Interpretable LLM Jailbreaking

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Lee, Sunbowen Ni, Shiwen Wei, Chi Li, Shuaimin Fan, Liyang Argha, Ahmadreza Alinejad-Rokny, Hamid Xu, Ruifeng Gong, Yicheng Yang, Min Shenzhen Key Laboratory for High Performance Data Mining Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences China Shenzhen University of Advanced Technology China WUST United States HITSZ China School of Biomedical Engineering UNSW Sydney Australia

Safety alignment mechanism are essential for preventing large language models (LLMs) from generating harmful information or unethical content. However, cleverly crafted prompts can bypass these safety measures without accessing the model’s internal parameters, a phenomenon known as black-box jailbreak. Existing heuristic black-box attack methods, such as genetic algorithms, suffer from limited effectiveness due to their inherent randomness, while recent reinforcement learning (RL) based methods often lack robust and informative reward signals. To address these challenges, we propose a novel black-box jailbreak method leveraging RL, which optimizes prompt generation by analyzing the embedding proximity between benign and malicious prompts. This approach ensures that the rewritten prompts closely align with the intent of the original prompts while enhancing the attack’s effectiveness. Furthermore, we introduce a comprehensive jailbreak evaluation framework incorporating keywords, intent matching, and answer validation to provide a more rigorous and holistic assessment of jailbreak success. Experimental results show the superiority of our approach, achieving state-of-the-art (SOTA) performance on several prominent open and closed-source LLMs, including Qwen2.5-7B-Instruct, Llama3.1-8B-Instruct, and GPT-4o-0806. Our method sets a new benchmark in jailbreak attack effectiveness, highlighting potential vulnerabilities in LLMs. The codebase for this work is available at https://***/Aegis1863/xJailbreak. Copyright © 2025, The Authors. All rights reserved.

关键词： Reinforcement learning

Out-of-Distribution Generalization on Graphs via Progressive Inference

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Xu, Yiming Shi, Bin Peng, Zhen Liu, Huixiang Dong, Bo Chen, Chen School of Computer Science and Technology Xi’an Jiaotong University China Shaanxi Provincial Key Laboratory of Big Data Knowledge Engineering Xi’an Jiaotong University China School of Distance Education Xi’an Jiaotong University China University of Virginia CharlottesvilleVA United States

The development and evaluation of graph neural networks (GNNs) generally follow the independent and identically distributed (i.i.d.) assumption. Yet this assumption is often untenable in practice due to the uncontrollable data generation mechanism. In particular, when the data distribution shows a significant shift, most GNNs would fail to produce reliable predictions and may even make decisions randomly. One of the most promising solutions to improve the model generalization is to pick out causal invariant parts in the input graph. Nonetheless, we observe a significant distribution gap between the causal parts learned by existing methods and the ground-truth, leading to undesirable performance. In response to the above issues, this paper presents GPro, a model that learns graph causal invariance with progressive inference. Specifically, the complicated graph causal invariant learning is decomposed into multiple intermediate inference steps from easy to hard, and the perception of GPro is continuously strengthened through a progressive inference process to extract causal features that are stable to distribution shifts. We also enlarge the training distribution by creating counterfactual samples to enhance the capability of the GPro in capturing the causal invariant parts. Extensive experiments demonstrate that our proposed GPro outperforms the state-of-the-art methods by 4.91% on average. For datasets with more severe distribution shifts, the performance improvement can be up to 6.86%. 1 © 2025, CC BY.

关键词： Graph neural networks

Multi-Quadrotor Cooperative Encirclement and Capture Approach in Obstacle Environments

学校读者我要写书评

暂无评论

Multi-Quadrotor Cooperative Encirclement and Capture Approac...

International Conference on Mechatronics (ICM)

作者： Zhenyu Li Ruixin Wu Jin Wang Chuanqi Hu Zhi Zheng Jun Meng Guodong Lu School of Mechanical Engineering Zhejiang University Hangzhou China State Key Laboratory of Fluid Power and Mechatronic Systems Zhejiang University Hangzhou China Zhejiang Key Laboratory of Industrial Big Data and Robot Intelligent Systems Zhejiang University Hangzhou China Robotics Research Center of Yuyao City Ningbo China Polytechnic Institute Zhejiang University Hangzhou China College of Electrical Engineering Zhejiang University Hangzhou China Center for Data Mining and Systems Biology College of Electrical Engineering Zhejiang University Hangzhou China

ISBN: (数字)9798331533892

ISBN: (纸本)9798331533908

This article proposes an algorithm designed for autonomous quadrotors to capture target quadrotors in complex obstacle environments, enabling multiple pursuers to perform distributed capture while simultaneously avoiding obstacles and collisions. We first introduce a 3D obstacle space representation method based on safety-priority reachable region (SPRR) and implement real-time construction for general point cloud maps. Subsequently, leveraging SPRR, we employ an area minimization approach to ensure the effective capture of the evader within a limited time. The outcomes are then integrated into a MINCO-based trajectory planning framework to ensure their practical execution by real quadrotors. Relying on SPRR, the optimization problem is simplified and the solution efficiency is improved. In the scenario of simulating the dynamics of real quadrotors, we compare the proposed algorithm with state-of-the-art evasion and pursuit methods, validating its effectiveness and advancements. This work presents a practical solution for multi-quadrotor collaborative capture.

关键词： Point cloud compression Visualization Three-dimensional displays Trajectory planning Heuristic algorithms Collaboration Real-time systems Safety Optimization Quadrotors

Using Depth-Enhanced Spatial Transformation for Student Gaze Target Estimation in Dual-View Classroom Images

学校读者我要写书评

暂无评论

Using Depth-Enhanced Spatial Transformation for Student Gaze...

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Haonan Miao Peizheng Zhao Yuqi Sun Fang Nan Xiaolong Zhang Yaqiang Wu Feng Tian School of Computer Science and Technology Xi’an Jiaotong University Xi’an China Ministry of Education Key Laboratory of Intelligent Networks and Network Security Xi’an Jiaotong University Xi’an China School of Advanced Technology Xi’an Jiaotong-Liverpool University Suzhou China Shaanxi Province Key Laboratory of Big Data Knowledge Engineering Xi’an Jiaotong University Xi’an China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Dual-view gaze target estimation in classroom environments has not been thoroughly explored. Existing methods lack consideration of depth information, primarily focusing on 2D image information and neglecting the latent 3D spatial context, which could lead to suboptimal transformation and cause the gaze cone to intersect with an incorrect object. This paper introduces a novel dual-view gaze target estimation method tailored for classroom settings, leveraging depth-enhanced spatial transformations. By formulating a depth-enhanced 2D space, our method uses depth-enhanced spatial transformation to accurately project students’ gaze cones to the teacher-oriented image. Additionally, we collected a dataset named DVSGE, specifically for student gaze target estimation in dual-view classroom images. Experimental results demonstrate significant performance improvements of 9.8% in AUC and 19.9% in L2-Distance for our method, surpassing existing methods.

关键词： Three-dimensional displays Estimation Focusing Signal processing Acoustics Speech processing

Enhancing multi-task performance through associative adversarial learning based on selective attacks

学校读者我要写书评

暂无评论

Neurocomputing 2025年 640卷

作者： Wang, Zhiming Yang, Yuanglong Zhang, Bilang Wang, Runqi Jing, Liping Zhang, Baochang School of Electronics and Information Engineering Beihang University Beijing China School of Computer Science and Technology Beijing Jiaotong University Beijing China Beijing Key Laboratory of Traffic Data Mining and Embodied Intelligence Beijing China State Key Laboratory of Advanced Rail Autonomous Operation Beijing China Hangzhou Research Institute School of Artificial Intelligence Beihang University China Zhongguancun Laboratory Beijing China Nanchang Institute of Technology Nanchang China

Adversarial learning can enhance model robustness against adversarial attacks. However, traditional adversarial learning generates adversarial samples by adding perturbations to entire images, causing excessive differences between adversarial samples and clean samples. To address this issue, we introduce the Associative Adversarial Learning (AAL) method based on a selective attack that can add perturbations to the images according to spatial attention maps, which demonstrate smaller differences from clean samples compared to those generated by global attacks. Thereby, it prevents the degradation of the model performance in clean samples. To further improve the robustness of the model, we develop a backtracking algorithm to effectively learn associative attention, enabling us to decouple attention from adversarial perturbations and mitigate the impact of adversarial attacks on the attention mechanism. Associative attention helps identify the key regions of adversarial examples, thereby enhancing the robustness of the models. We also introduce a projection function for associative attention, which allows AAL to select different attacking regions across various tasks, thus enabling its applicability to a variety of scenarios. Our experimental results show that AAL can be seamlessly integrated into various models, significantly improving adversarial robustness against multiple attacks. For example, AAL can increase InfraRed Small Target Detection mAP on SIRST-v2 by 4.5% and the accuracy of few-shot recognition on miniImageNet by 1.63%. © 2025 Elsevier B.V.

关键词： Adversarial machine learning

Hierarchical Alignment-enhanced Adaptive Grounding Network for Generalized Referring Expression Comprehension

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Wang, Yaxian Ding, Henghui He, Shuting Jiang, Xudong Wei, Bifan Liu, Jun School of Computer Science and Technology Xi'an Jiaotong University China Ministry of Education Key Laboratory of Intelligent Networks and Network Security Xi'an Jiaotong University China Institute of Big Data Fudan University China Shanghai University of Finance and Economics China Nanyang Technological University Singapore School of Continuing Education Xi'an Jiaotong University China Shaanxi Province Key Laboratory of Big Data Knowledge Engineering Xi'an Jiaotong University China

In this work, we address the challenging task of Generalized Referring Expression Comprehension (GREC). Compared to the classic Referring Expression Comprehension (REC) that focuses on single-target expressions, GREC extends the scope to a more practical setting by further encompassing notarget and multi-target expressions. Existing REC methods face challenges in handling the complex cases encountered in GREC, primarily due to their fixed output and limitations in multi-modal representations. To address these issues, we propose a Hierarchical Alignment-enhanced Adaptive Grounding Network (HieA2G) for GREC, which can flexibly deal with various types of referring expressions. First, a Hierarchical Multi-modal Semantic Alignment (HMSA) module is proposed to incorporate three levels of alignments, including word-object, phrase-object, and text-image alignment. It enables hierarchical cross-modal interactions across multiple levels to achieve comprehensive and robust multi-modal understanding, greatly enhancing grounding ability for complex cases. Then, to address the varying number of target objects in GREC, we introduce an Adaptive Grounding Counter (AGC) to dynamically determine the number of output targets. Additionally, an auxiliary contrastive loss is employed in AGC to enhance object-counting ability by pulling in multi-modal features with the same counting and pushing away those with different counting. Extensive experimental results show that HieA2G achieves new state-of-the-art performance on the challenging GREC task and also the other 4 tasks, including REC, Phrase Grounding, Referring Expression Segmentation (RES), and Generalized Referring Expression Segmentation (GRES), demonstrating the remarkable superiority and generalizability of the proposed HieA2G. Copyright © 2025, The Authors. All rights reserved.

关键词： Electric grounding