检索结果-内蒙古大学图书馆

Are Large Language Models Really Good Logical Reasoners? A Comprehensive Evaluation and Beyond

IEEE TRANSACTIONS ON knowledge AND data engineering 2025年第4期37卷 1620-1634页

作者： Xu, Fangzhi Lin, Qika Han, Jiawei Zhao, Tianzhe Liu, Jun Cambria, Erik Xi An Jiao Tong Univ Sch Comp Sci & Technol Key Lab Intelligent Networks & Net work Secur Minist Educ Xian 710049 Shaanxi Peoples R China Shaanxi Prov Key Lab Big Data Knowledge Engn Xian 710049 Shaanxi Peoples R China Nanyang Technol Univ Coll Comp & Data Sci Singapore 639798 Singapore

Logical reasoning consistently plays a fundamental and significant role in the domains of knowledge engineering and artificial intelligence. Recently, Large Language Models (LLMs) have emerged as a noteworthy innovation in natural language processing (NLP). However, the question of whether LLMs can effectively address the task of logical reasoning, which requires gradual cognitive inference similar to human intelligence, remains unanswered. To this end, we aim to bridge this gap and provide comprehensive evaluations in this paper. First, to offer systematic evaluations, we select fifteen typical logical reasoning datasets and organize them into deductive, inductive, abductive and mixed-form reasoning settings. Considering the comprehensiveness of evaluations, we include 3 early-era representative LLMs and 4 trending LLMs. Second, different from previous evaluations relying only on simple metrics (e.g., accuracy), we propose fine-level evaluations in objective and subjective manners, covering both answers and explanations, including answer correctness, explain correctness, explain completeness and explain redundancy. Additionally, to uncover the logical flaws of LLMs, problematic cases will be attributed to five error types from two dimensions, i.e., evidence selection process and reasoning process. Third, to avoid the influences of knowledge bias and concentrate purely on benchmarking the logical reasoning capability of LLMs, we propose a new dataset with neutral content. Based on the in-depth evaluations, this paper finally forms a general evaluation scheme of logical reasoning capability from six dimensions (i.e., Correct, Rigorous, Self-aware, Active, Oriented and No hallucination). It reflects the pros and cons of LLMs and gives guiding directions for future works.

关键词： Cognition Benchmark testing Measurement Large language models Self-aware Systematics Redundancy knowledge engineering Chatbots Accuracy Logical reasoning large language model deductive reasoning inductive reasoning abductive reasoning

来源：评论

学校读者我要写书评

暂无评论

Lightweight expression recognition combined attention fusion network with hybrid knowledge distillation for occluded e-learner facial images

引用

NEUROCOMPUTING 2025年 628卷

作者： Chen, Yan Li, Kexuan Tian, Feng Wei, Ganglin Seberi, Morteza Xi An Jiao Tong Univ Sch Comp Sci & Technol Xian 710049 Peoples R China Xi An Jiao Tong Univ Key Lab Intelligent Networks & Network Secur Minist Educ Xian 710049 Peoples R China Xi An Jiao Tong Univ Shaanxi Prov Key Lab Big Data Knowledge Engn Xian 710049 Peoples R China Univ Technol Sydney Sch Comp Sci Sydney Australia

In online education, it is crucial to monitor the students' learning status timely and accurately. Facial expression recognition serves as the main tool for assessing their engagement levels. Existing algorithms still have some issues when directly applied in online learning. These issues primarily involve the loss of facial features due to occlusion, which directly affects the accuracy of expression recognition and at the same time, the expression recognition models require a large number of parameters and significant computational power, making them difficult to deploy and apply effectively on mobile devices with limited hardware resources. We aims to address the two issues by a two-stage framework: training process of an occluded facial expression recognition model, and the compression process of the occlusion facial expression recognition. To be more specific, in the first stage, we propose an occlusion facial expression recognition model based on attention fusion (AFNet), which adopts a multi-branch spatial attention network, extracts local facial features, automatically perceives facial occluded regions, reduces the weight of the occluded areas, and enhance robustness to occlusion by combining with a randomly masked channel network. Meanwhile, a feature pyramid network is introduced to extract global multi-scale features. In the second stage, we propose a hybrid model compression algorithm based on multi-layer knowledge distillation (MKD). We introduce a spatial attention network to focus on the important knowledge, reducing the information loss during knowledge distillation. Experimental results on five datasets show that the AFNet and MKD outperformed the baseline.

关键词： Occlusion facial expression recognition Attention mechanism knowledge distillation Model compression

来源：评论

学校读者我要写书评

暂无评论

Hierarchy-Based Diagram-Sentence Matching on Dual-Modal Graphs

引用

PATTERN RECOGNITION 2025年 165卷

作者： Wu, Wenjun Zhang, Lingling Liu, Jun Ren, Ming Hu, Xin Wang, Jiaxin Wang, Qianying Xi'an Jiaotong Univ Sch Comp Sci & Technol Xian 710049 Peoples R China Xi An Jiao Tong Univ Minist Educ Key Lab Intelligent Networks & Network Secur Xian 710049 Peoples R China Changan Univ Sch Data Sci & Artificial Intelligence Xian 710049 Peoples R China Lenovo Res Beijing Peoples R China Xi An Jiao Tong Univ Shaanxi Prov Key Lab Big Data Knowledge Engn Xian 710049 Peoples R China

Diagram is a special kind of image drawn by domain experts, which mainly consists of graphic symbols and abstract drawings. It is essential to combine the diagrams with other modalities (e.g. textual descriptions and subtitles in teaching videos) for in-depth understanding of the knowledge concepts. Diagram-sentence matching, a novel task proposed to bridge abstract diagram representation and explicit natural language, is significant to textbook question answering (TQA) and diagram understanding but remains challenging. Existing vision-language matching works mainly focus on the field of natural images and are not applicable to diagrams due to the following two characteristics: (1) the relation in diagrams has diversified representation forms;(2) the knowledge concepts conveyed in diagrams are key to fine-grained diagram-sentence matching. In this paper, we propose the Hierarchy-Based Diagram-Sentence Matching (HBDSM) model and transfer this problem into a cross-modal knowledge concept matching task at multiple levels. To achieve this, the HBDSM first encodes the diagram and sentence as symmetrical dual-modal graphs. For diagram, a novel Visual Relation Structure Learning (VRSL) method is designed to explore the structural relations between objects, which constitute the edges. For sentence, words are fused into object and relation chunks as nodes, associated by edges according to their semantic dependencies. Motivated by the human cognitive process, the fine-grained correspondence between diagram and sentence is modeled based on the hierarchy of dual-modal graphs progressively, using from low-order to high-order information. Node-level matching establishes alignment of object nodes, based on which structure-level matching compares the internal structures of both graphs. Further, concept-level matching includes relation semantics to match the cross-modal concepts based on structure alignment. Extensive experiments demonstrate the effectiveness of HBDSM in diagram-sen

关键词： Diagram-sentence matching Diagram understanding Image-text matching

来源：评论

学校读者我要写书评

暂无评论

Dual-Aspect Noise-Based Regularization for Multi-Modal Relation Extraction in Media Posts

IEEE Transactions on Audio, Speech and Language Processing

引用

IEEE Transactions on Audio, Speech and Language Processing 2025年 33卷 1324-1336页

作者： Kai Sun Bin Shi Samuel Mensah Wenjian Liu Bo Dong School of Computer Science and Technology and Shaanxi Provincial Key Laboratory of Big Data Knowledge Engineering Xi'an Jiaotong University Xi'an China School of Computer Science and Technology The University of Sheffield Sheffield U.K. Faculty of Data Science City University of Macau Macau China School of Continuing Education and Shaanxi Provincial Key Laboratory of Big Data Knowledge Engineering Xi'an Jiaotong University Xi'an China

Multi-Modal Relation Extraction (MMRE) plays a key role in various multimedia applications including, recommendation and information retrieval systems. MMRE aims to extract the semantic relation between entities by leveraging context from a text-image pair. By utilizing context from images, the challenge of learning from noisy images in MMRE emerges as a research problem by itself. For instance, subtle variations in similar images can act as noise and potentially impact the predictions made by MMRE models. To tackle this problem, current work utilizes attention mechanisms to fuse relevant text and image features or devise data augmentation techniques (e.g., via generative models) to improve generalization. However, the current performance still remains unsatisfactory. In an effort to improve upon the performance, we propose a Dual-Aspect Noise-based Regularization framework that encompasses two techniques: 1) noise removal through an adaptive gating mechanism, 2) fighting noise with noise to improve feature stability in the learning process. We find that combining these techniques encourages the model to focus on more relevant image features for MMRE. We carry out extensive experiments and demonstrate that our proposed model is further enhanced by exploring data augmentation techniques. This additional improvement leads the model to achieve state-of-the-art performance on the widely-used Multi-modal Neural Relation Extraction (MNRE) dataset, and show its effectiveness and generalizability on the Multi-Modal Named Entity Recognition task.

关键词： Feature extraction Noise data mining Noise measurement Social networking (online) Adaptation models Transformers Training Predictive models data models

来源：评论

学校读者我要写书评

暂无评论

Disentangled Noisy Correspondence Learning

引用

IEEE Transactions on Image Processing 2025年 34卷 2602-2615页

作者： Dang, Zhuohang Luo, Minnan Wang, Jihong Jia, Chengyou Han, Haochen Wan, Herun Dai, Guang Chang, Xiaojun Wang, Jingdong Xi’an Jiaotong University School of Computer Science and Technology Ministry of Education Key Laboratory of Intelligent Networks and Network Security Shaanxi Province Key Laboratory of Big Data Knowledge Engineering Shaanxi Xi’an710049 China SGIT AI Laboratory Xi’an710048 China State Grid Corporation of China State Grid Shaanxi Electric Power Company Ltd. Xi’an710048 China University of Science and Technology of China School of Information Science and Technology Hefei230026 China Abu Dhabi United Arab Emirates Baidu Inc. Beijing100085 China

Cross-modal retrieval is crucial in understanding latent correspondences across modalities. However, existing methods implicitly assume well-matched training data, which is impractical as real-world data inevitably involves imperfect alignments, i.e., noisy correspondences. Although some works explore similarity-based strategies to address such noise, they suffer from sub-optimal similarity predictions influenced by modality-exclusive information (MEI), e.g., background noise in images and abstract definitions in texts. This issue arises as MEI is not shared across modalities, thus aligning it in training can markedly mislead similarity predictions. Moreover, although intuitive, directly applying previous cross-modal disentanglement methods suffers from limited noise tolerance and disentanglement efficacy. Inspired by the robustness of information bottlenecks against noise, we introduce DisNCL, a novel information-theoretic framework for feature Disentanglement in Noisy Correspondence Learning, to adaptively balance the extraction of modality-invariant information (MII) and MEI with certifiable optimal cross-modal disentanglement efficacy. DisNCL then enhances similarity predictions in modality-invariant subspace, thereby greatly boosting similarity-based alleviation strategy for noisy correspondences. Furthermore, DisNCL introduces soft matching targets to model noisy many-to-many relationships inherent in multi-modal inputs for noise-robust and accurate cross-modal alignment. Extensive experiments confirm DisNCL’s efficacy by 2% average recall improvement. Mutual information estimation and visualization results show that DisNCL learns meaningful MII/MEI subspaces, validating our theoretical analyses. © 1992-2012 IEEE.

关键词： Information theory

来源：评论

学校读者我要写书评

暂无评论

引用

data MINING AND knowledge DISCOVERY 2025年第2期39卷 1-24页

作者： Xu, Yiming Shi, Bin Dong, Bo Wang, Jiaxiang Wei, Hua Zheng, Qinghua Xi An Jiao Tong Univ Sch Comp Sci & Technol Xian Peoples R China Xi An Jiao Tong Univ Sch Distance Educ Xian Peoples R China Arizona State Univ Sch Comp & Augmented Intelligence Phoenix AZ USA Xi An Jiao Tong Univ Key Lab Intelligent Networks & Network Secur Minist Educ Xian Peoples R China Xi An Jiao Tong Univ Shaanxi Prov Key Lab Big Data Knowledge Engn Xian Peoples R China

Tax evasion causes severe losses of government revenues and disturbs the economic order of fair competition. To help alleviate this problem, the latest tax evasion detection solutions utilize expert knowledge to extract features and then train classifiers to determine whether a company is suspected of tax evasion. However, existing solutions mainly focus on the statistical features of the company, but fail to exploit the rich interactive information in tax scenarios, which affect the detection performance. In this paper, we first model the tax scenario as a heterogeneous graph and study the tax evasion detection problem under the heterogeneous graph model. To improve the performance of tax evasion detection, a novel graph neural network model is proposed to extract the comprehensive information of heterogeneous graphs. Specifically, we use heterogeneous and complex related party transaction groups to filter low-level noise information. Moreover, a hierarchical attention mechanism is designed to capture the deeper structure and semantic information hidden in the related party transaction group. We apply our method to the real risk management system of the tax bureau, and evaluate it on two human-labeled real-world tax datasets. The results demonstrate that our method significantly outperforms the state-of-the-art in the tax evasion detection task. The code and data are available at: https://***/yimingxu24/TED.

关键词： Tax evasion detection Heterogeneous graph Related party transaction Graph neural network Graph attention

来源：评论

学校读者我要写书评

暂无评论

TDGI: Translation-Guided Double-Graph Inference for Document-Level Relation Extraction

引用

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 2025年第4期47卷 2647-2659页

作者： Zhang, Lingling Zhong, Yujie Zheng, Qinghua Liu, Jun Wang, Qianying Wang, Jiaxin Chang, Xiaojun Xi An Jiao Tong Univ Sch Comp Sci & Technol Key Lab Intelligent Networks Minist Educ Xian 710049 Peoples R China Xi An Jiao Tong Univ Sch Comp Sci & Technol Shaanxi Prov Key Lab Big Data Knowledge Engn Xian 710049 Peoples R China Lenovo Res Beijing 100094 Peoples R China Univ Sci & Technol China Sch Informat Sci & Technol Hefei 230026 Peoples R China MBZUAI Dept Comp Vis Abu Dhabi U Arab Emirates

Document-level relation extraction (DocRE) aims at predicting relations of all entity pairs in one document, which plays an important role in information extraction. DocRE is more challenging than previous sentence-level relation extraction, as it often requires coreference and logical reasoning across multiple sentences. Graph-based methods are the mainstream solution to this complex reasoning in DocRE. They generally construct the heterogeneous graphs with entities, mentions, and sentences as nodes, co-occurrence and co-reference relations as edges. Their performance is difficult to further break through because the semantics and direction of the relation are not jointly considered in graph inference process. To this end, we propose a novel translation-guided double-graph inference network named TDGI for DocRE. On one hand, TDGI includes two relation semantics-aware and direction-aware reasoning graphs, i.e., mention graph and entity graph, to mine relations among long-distance entities more explicitly. Each graph consists of three elements: vectorized nodes, edges, and direction weights. On the other hand, we devise an interesting translation-based graph updating strategy that guides the embeddings of mention/entity nodes, relation edges, and direction weights following the specific translation algebraic structure, thereby to enhance the reasoning skills of TDGI. In the training procedure of TDGI, we minimize the relation multi-classification loss and triple contrastive loss together to guarantee the model's stability and robustness. Comprehensive experiments on three widely-used datasets show that TDGI achieves outstanding performance comparing with state-of-the-art baselines.

关键词： Cognition Translation Semantics Transformers data mining Electronic mail Context modeling Vectors Training Robustness Document-level relation extraction double-graph inference translation rule relation semantics relation direction

来源：评论

学校读者我要写书评

暂无评论

OSE: On-Site State of Health Estimation for Li-ion Battery using Real-Time Field data

引用

IEEE Transactions on Transportation Electrification 2025年

作者： Chen, Yanwen Zhao, Cong Yang, Shusen Zhao, Peng Ren, Xuebin Han, Qing Yang, Yuqian Wu, Shuaijun Xi'an Jiaotong University National Engineering Laboratory for Big Data Analytics Shaanxi Xi'an710049 China Xi'an Jiaotong University National Engineering Laboratory for Big Data Analytics The Ministry of Education Key Lab for Intelligent Networks and Network Security Shaanxi Xi'an710049 China

On-site lithium-ion battery state of health (SoH) estimation is of crucial importance for reliable operations of electric vehicles (EVs). Yet, due to the low-quality of unlabeled real-time field data, diverse operating environments of in-service EVs, and limited computational capability of onboard devices, existing techniques established on data from well-controlled experimental environments are not practical for real world EVs' SoH estimation. Accurate and rapid SoH estimation based on field data of in-service EVs still remains quite challenging. To tackle this challenge, we present an on-site SoH estimation (OSE) method using in-service EV field data through a new knowledge embedded deep transfer learning (DTL) model. Initially, an universal data preprocessing approach integrating mechanism knowledge is designed to process low-quality data under diverse operating environments. Then, we develop a domain adaptive hybrid deep neural network (DAHDNN) model suitable for unlabeled field data, which can be deployed via edge cloud collaborative framework to meet actual computational capability. We demonstrate the superiority of our method across four real datasets, where, OSE's estimation error is decreased by up to 78.5% compared with the state-of-the-art methods. The results indicate that the proposed method has good generalizability and reliability for SoH estimation on real-time field data. © 2015 IEEE.

关键词： State of charge

来源：评论

学校读者我要写书评

暂无评论

TEST: Temporal-spatial separated transformer for temporal action localization

引用

NEUROCOMPUTING 2025年 614卷

作者： Wan, Herun Luo, Minnan Li, Zhihui Wang, Yang Xi An Jiao Tong Univ Sch Comp Sci & Technol Xian 710049 Peoples R China Xi An Jiao Tong Univ Minist Educ Key Lab Intelligent Networks & Network Secur Xian 710049 Peoples R China Xi An Jiao Tong Univ Shaanxi Prov Key Lab Big Data Knowledge Engn Xian 710049 Peoples R China Univ Sci & Technol China Sch Informat Sci & Technol Hefei 230026 Peoples R China Xi An Jiao Tong Univ Sch Continuing Educ Xian 710049 Peoples R China

Temporal action localization is a fundamental task in video understanding. Existing methods fall into three categories: anchor-based, actionness-guided, and anchor-free. Anchor-based and actionness-guided models need huge computation resources to process redundant proposals or enumerate every possible proposal. Anchor- free models with lighter parameters become a more attractive option as temporal actions become more complex. However, they typically struggle to achieve high performance due to the need to aggregate global temporal-spatial features at every time step. To overcome this limitation, we design three efficient transformer- based architectures, bringing two advantages: (i) the global receptive field of transformers enables models to aggregate spatial and temporal at each time step, and (ii) the transformers could capture the moment-level feature, enhancing localization performance. Our designed architectures are adapted to any framework, thus we propose a simple but effective anchor-free framework named TEST. Compared to strong baselines, TEST achieves 0.96% to 3.20% improvement on two real-world datasets. Meanwhile, it improves time efficiency by 1.36 times and space efficiency by 1.08 times. Further experiments prove the effectiveness of TEST's modules. Implementation of our work is available at https://***/whr000001/TeST.

关键词： Video transformer Temporal action localization High efficiency

来源：评论

学校读者我要写书评

暂无评论

A Unified Optimal Transport Framework for Cross-Modal Retrieval With Noisy labels

引用

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2025年 PP卷 PP页

作者： Han, Haochen Luo, Minnan Liu, Huan Nan, Fang Liu, Jun Pengcheng Lab Shenzhen 518066 Peoples R China Xi An Jiao Tong Univ Sch Comp Sci & Technol Xian 710049 Peoples R China Xi An Jiao Tong Univ Minist Educ Key Lab Intelligent Networks & Network Secur Xian 710049 Peoples R China Xi An Jiao Tong Univ Fac Elect & Informat Engn Xian 710049 Peoples R China Xi An Jiao Tong Univ Shaanxi Prov Key Lab Big Data Knowledge Engn Xian 710049 Peoples R China

Cross-modal retrieval (CMR) aims to establish interaction between different modalities, among which supervised CMR is emerging due to its flexibility in learning semantic category discrimination. Despite the remarkable performance of previous supervised CMR methods, much of their success can be attributed to the well-annotated data. However, even for unimodal data, precise annotation is expensive and time-consuming, and it becomes more challenging with the multimodal scenario. In practice, massive multimodal data are collected from the Internet with coarse annotation, which inevitably introduces noisy labels. Training with such misleading labels would bring two key challenges-enforcing the multimodal samples to align incorrect semantics and widen the heterogeneous gap, resulting in poor retrieval performance. To tackle these challenges, this work proposes UOT-RCL, a unified framework based on optimal transport (OT) for robust CMR. First, we propose a semantic alignment based on partial OT to progressively correct the noisy labels, where a novel cross-modal consistent cost function is designed to blend different modalities and provide precise transport cost. Second, to narrow the discrepancy in multimodal data, an OT-based relation alignment is proposed to infer the semantic-level cross-modal matching. Both of these components leverage the inherent correlation among multimodal data to facilitate effective cost function. The experiments on three widely used CMR datasets demonstrate that our UOT-RCL surpasses the state-of-the-art approaches and significantly improves the robustness against noisy labels.

关键词： Noise measurement Semantics Training Noise Cross modal retrieval Costs Cost function Visualization Hands Annotations Noisy labels optimal transport (OT) supervised cross-modal retrieval (CMR)

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：