Information Processing & Management最新文献_第4页

FedDPKD: Federated learning with dual-phase knowledge distillation for label distribution skew FedDPKD：基于双阶段知识蒸馏的标签分布倾斜联邦学习

IF 6.9 1区管理学

Information Processing & Management Pub Date : 2026-07-01 Epub Date: 2026-02-03 DOI: 10.1016/j.ipm.2026.104657

Fanfan Shen , Wenzhang Su , Chao Xu , Zhiquan Liu , Jun Feng , Yanxiang He

{"title":"FedDPKD: Federated learning with dual-phase knowledge distillation for label distribution skew","authors":"Fanfan Shen , Wenzhang Su , Chao Xu , Zhiquan Liu , Jun Feng , Yanxiang He","doi":"10.1016/j.ipm.2026.104657","DOIUrl":"10.1016/j.ipm.2026.104657","url":null,"abstract":"<div><div>The non-independent and identically distributed (non-IID) nature of client data presents a major challenge in federated learning, particularly in the form of label distribution skew. This skew often exacerbates the tendency of models to forget minority and absent labels, resulting in classification outcomes biased toward majority labels. To address this issue, we propose a novel approach called Federated Learning with Dual-Phase Knowledge Distillation (FedDPKD). In this framework, client labels are dynamically categorized into majority, minority, and absent labels based on local label distribution. FedDPKD employs a dual-phase knowledge distillation strategy, consisting of Global-to-Client (G2C) and Client-to-Global (C2G) distillation stages, enhanced by a dynamic weight adjustment mechanism to improve training flexibility and efficiency. The proposed method enhances the global model’s classification accuracy and significantly alleviates the forgetting of minority and absent labels under label distribution skew. Extensive experiments on CIFAR-10, CIFAR-100, MNIST, and CINIC-10 demonstrate that FedDPKD consistently outperforms existing leading methods. Notably, under severe label distribution skew, FedDPKD achieves a 6.4% improvement in accuracy over the second-best baseline on the CIFAR-10 dataset.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104657"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190715","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Break fake frontiers: A triple-knowledge approach to multi-domain fake news detection 突破假新闻前沿：一种多领域假新闻检测的三重知识方法

IF 6.9 1区管理学

Information Processing & Management Pub Date : 2026-07-01 Epub Date: 2026-01-28 DOI: 10.1016/j.ipm.2026.104649

Xinnan Liu , Chenxi Liu , Anran Yu , Bin Tang , Zhenyang Cao , Zhengxiong Long , Runqi Su , Heng-Yang Lu

{"title":"Break fake frontiers: A triple-knowledge approach to multi-domain fake news detection","authors":"Xinnan Liu , Chenxi Liu , Anran Yu , Bin Tang , Zhenyang Cao , Zhengxiong Long , Runqi Su , Heng-Yang Lu","doi":"10.1016/j.ipm.2026.104649","DOIUrl":"10.1016/j.ipm.2026.104649","url":null,"abstract":"<div><div>Nowdays, news events increasingly exhibit cross-domain topicality, making multi-domain fake news detection a critical yet challenging task. A key challenge is enhancing model performance for fake news detection while mitigating domain bias caused by unbalanced dataset. To address this, we propose the Knowledge-assistance Knowledge-mining Knowledge-debiased Multi-domain Fake News Detection Framework (K<sup>3</sup>MDFEND) which introduces external knowledge and proposes multi-domain contrastive learning. In particular, we integrate Large Language Models (LLMs) through a novel argumentation-based prompt engineering framework to obtain reliable external knowledge. We design Quality-Aware Attention Fusion module that dynamically weights evidence credibility to handle reviews of varying quality while combining with our deep learning framework. To further distill key insights from the comments while preserving their inherent semantic integrity, we leverage feature alignment techniques on the comment features. To further mitigate domain bias, we propose multi-domain contrastive learning and successfully combine the spurious correlations between domains and news authenticity. Extensive experiments on Chinese and English datasets demonstrate that K<sup>3</sup>MDFEND achieves state-of-the-art performance in both detection performance and bias metric reduction. On Chinese and English datasets, F1 scores, increase 92.89% ∼ <strong>95.13%</strong> and 83.59% ∼ <strong>85.28%</strong>, bias metric, decrease 0.8522 ∼ <strong>0.5612</strong> and 0.2698 ∼ <strong>0.1931</strong>.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104649"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146081579","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

DUIC: User-descriptive intention guided clustering for personalized and understandable document partitions DUIC：用户描述意图引导的用于个性化和可理解文档分区的聚类

IF 6.9 1区管理学

Information Processing & Management Pub Date : 2026-07-01 Epub Date: 2026-02-03 DOI: 10.1016/j.ipm.2026.104648

Jingjing Xue , Ruizhang Huang , Ruina Bai , Yongbin Qin , Yanping Chen , Chuan Lin

{"title":"DUIC: User-descriptive intention guided clustering for personalized and understandable document partitions","authors":"Jingjing Xue , Ruizhang Huang , Ruina Bai , Yongbin Qin , Yanping Chen , Chuan Lin","doi":"10.1016/j.ipm.2026.104648","DOIUrl":"10.1016/j.ipm.2026.104648","url":null,"abstract":"<div><div>Personalized document clustering often fails to directly utilize user-descriptive intentions expressed in natural language and produces clusters that are difficult to understand. We propose Descriptive-Intention-Guided Understandable Clustering (DUIC), an end-to-end model that integrates a Personalized Intention-Guided Clustering (PIGC) module with a User-Aligned Cluster Explanation (UACE) module. PIGC parses and leverages user-descriptive intentions to guide clustering, while UACE generates human-understandable explanations aligned with these intentions. DUIC was evaluated on six benchmark datasets (737–6,000 text documents) including MIND-6K, BBCSport, and ArXiv variants, against ten state-of-the-art baselines. It achieved up to 99.05% accuracy, 92.80% NMI, and 96.24% ARI. Compared with the strongest baseline on each dataset, DUIC achieved maximum improvements of 16.28% in ACC and 10.93% in NMI. Results show DUIC not only effectively guides clustering results to match user intention but also helps users understand what each cluster represents, bridging the gap between user-descriptive intention and personalized, understandable clustering results.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104648"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190674","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Instance and prototype contrastive learning for multi-view 3D model retrieval and classification 多视图三维模型检索与分类的实例与原型对比学习

IF 6.9 1区管理学

Information Processing & Management Pub Date : 2026-07-01 Epub Date: 2026-01-28 DOI: 10.1016/j.ipm.2026.104655

Yaqian Zhou , Zhenghao Fang , Zhibin Gu , Song Yang , Yan Wang

{"title":"Instance and prototype contrastive learning for multi-view 3D model retrieval and classification","authors":"Yaqian Zhou , Zhenghao Fang , Zhibin Gu , Song Yang , Yan Wang","doi":"10.1016/j.ipm.2026.104655","DOIUrl":"10.1016/j.ipm.2026.104655","url":null,"abstract":"<div><div>Unsupervised 3D model retrieval and classification have obtained much attention due to widespread applications. However, existing methods focus only on global representations while ignoring local saliency learning, leading to redundant distraction and insufficient complementarity. In addition, they neglect intra- and inter-class contextual relevance during representation learning, leading to inaccurate embedding space partitioning and missing representative prototypes. To address these challenges, we present Instance and Prototype Contrastive Learning (IPCL), an unsupervised dual-network framework that simultaneously captures view-level local features and model-level semantic information. Specifically, we treat each view as an instance and employ inter-instance contrastive learning to extract discriminative local salient features, mitigating redundancy and enhancing cross-view complementarity. For global semantic modeling, we establish class prototypes for 3D models and propagate the semantic information to global features via a prototype-aware contrastive loss, strengthening class-level discriminability. Innovatively, we employ a bottom-up adaptive clustering algorithm called voting clustering, which mines deeper semantic correlations to refine prototype selection and embedding space structure. Comprehensive evaluations demonstrate the superiority of IPCL, e.g., IPCL outperforms most unsupervised methods, achieving classification accuracy improvements by 0.1% to 18.1% on ModelNet40 and 0.1% to 16.7% on ShapeNet55. IPCL achieves average retrieval gains of 15.2% in NN on ModelNet40 and 16.7% in mAP under the micro setting on ShapeNet55.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104655"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146081582","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Invariant learning improves out-of-distribution generalization for IP geolocation 不变学习改进了IP地理定位的分布外泛化

IF 6.9 1区管理学

Information Processing & Management Pub Date : 2026-07-01 Epub Date: 2026-01-29 DOI: 10.1016/j.ipm.2026.104641

Xiaohan Wang , Xueting Liu , Wenxin Tai , Joojo Walker , Yong Wang , Kai Chen , Fan Zhou

{"title":"Invariant learning improves out-of-distribution generalization for IP geolocation","authors":"Xiaohan Wang , Xueting Liu , Wenxin Tai , Joojo Walker , Yong Wang , Kai Chen , Fan Zhou","doi":"10.1016/j.ipm.2026.104641","DOIUrl":"10.1016/j.ipm.2026.104641","url":null,"abstract":"<div><div>Accurate IP geolocation plays a critical role in a wide range of location-aware applications, from cybersecurity to content delivery. While recent advances in deep learning have led to substantial improvements in geolocation accuracy, existing methods often fail to generalize under out-of-distribution (OOD) scenarios caused by distribution shifts. To address this challenge, we propose a novel framework-Graph Invariant Learning (GIL)-for IP geolocation, referred to as GILGeo. Our approach is designed to identify invariant structural patterns in IP graphs across diverse environments, thereby enhancing model generalizability. By dynamically recombining invariant and spurious features, GILGeo simulates a variety of environmental conditions during training. This promotes the learning of domain-invariant representations and leads to significantly improved performance in unseen OOD settings. Extensive experiments on three real-world datasets show that GILGeo outperforms state-of-the-art baselines, establishing a new benchmark for IP geolocation under distributional shift. Our anonymized code and datasets are publicly available at: <span><span>https://github.com/xiaohanwang01/GILGeo</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104641"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146081580","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Human cognitive process aligned rumor detection with small language models enhanced large language models 人类认知过程将谣言检测与小语言模型增强的大语言模型相一致

IF 6.9 1区管理学

Information Processing & Management Pub Date : 2026-07-01 Epub Date: 2026-02-05 DOI: 10.1016/j.ipm.2026.104659

Yirong Zeng, Xiao Ding, Bibo Cai, Bing Qin, Ting Liu

{"title":"Human cognitive process aligned rumor detection with small language models enhanced large language models","authors":"Yirong Zeng, Xiao Ding, Bibo Cai, Bing Qin, Ting Liu","doi":"10.1016/j.ipm.2026.104659","DOIUrl":"10.1016/j.ipm.2026.104659","url":null,"abstract":"<div><div>The rapid spread of rumors underscores the critical need for effective detection methods. Recent studies have begun leveraging large language models (LLMs) for rumor detection in prompt-based scenarios Unfortunately, LLMs lack specialized skills without proper training in rumor detection compare to fine-tuned small language models (SLMs). To address these challenges, we introduce a human cognitive process into rumor detection to efficiently utilize data fragments. Further, we propose CoRuD, a multi-stage rumor detection method that integrates LLMs and SLMs, aligned with the human cognitive process. This approach integrates the strengths of SLMs in fine-tuned scenario and LLMs in prompt scenario. Specifically, it structures the rumor detection process into multiple stages aligned with the human cognitive process, generates expert insights through fine-tuned SLMs at each stage, and incorporates these insights into the prompts of foundation models to enhance their rumor detection ability. Our method achieves significant improvements across three public rumor detection datasets (Liar, Politifact, and Weibo, with over 14K test examples). It outperforms LLM-based approaches (e.g., surpassing GPT-3.5 by 17%) and various fine-tuned SLMs by over 5%, while also addressing the interpretability limitations of SLMs.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104659"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190716","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

SageJavon: A scalable AI tutor for personalized programming learning SageJavon：一个可扩展的AI导师，用于个性化编程学习

IF 6.9 1区管理学

Information Processing & Management Pub Date : 2026-07-01 Epub Date: 2026-02-03 DOI: 10.1016/j.ipm.2025.104605

Hanyu Zhao , Yuzhuo Wu , Zhufeng Lu , Xiaohua Yu , Weikai Miao , Liangyu Chen

{"title":"SageJavon: A scalable AI tutor for personalized programming learning","authors":"Hanyu Zhao , Yuzhuo Wu , Zhufeng Lu , Xiaohua Yu , Weikai Miao , Liangyu Chen","doi":"10.1016/j.ipm.2025.104605","DOIUrl":"10.1016/j.ipm.2025.104605","url":null,"abstract":"<div><div>Despite advancements in programming education, providing adaptive learning experiences while managing cognitive load remains a challenge. Inspired by Bloom’s Taxonomy, we propose the <strong>Learn-Practice-Evaluate-Support (LPES)</strong> framework, integrated into SageJavon, an AI-powered tutoring system based on Large Language Models (LLMs). SageJavon overcomes the limitations of traditional AI tutors, such as rigid content delivery and a lack of personalized feedback. It includes four innovations: (1) an enhanced Retrieval-Augmented Generation (RAG) model for personalized resource delivery, (2) heuristic and follow-up questions to guide problem-solving, (3) adaptive knowledge tracing and recommendation models, and (4) the FUPS-Score for automatic code assessment. Deployed in a 12-week course with 85 students, SageJavon facilitated 5306 interactions in the Knowledge Q&A module and 5144 in the Programming Mentor module, with LLM-based scoring yielding average scores of 87.39 and 84.19. Compared to traditional tools, SageJavon reduces cognitive load, showing improved scores in physical demand, temporal demand, and performance (all <em>p</em> < .001). It also improves constructivist learning outcomes, with higher scores in prior knowledge activation (5.01 vs. 4.78), knowledge transfer (5.09 vs. 4.82), and error correction (5.24 vs. 4.68). We compare our improved RAG with other implementations and find that it outperforms the others with a 2.93% average improvement. In exercise recommendations, we achieve over 80% user satisfaction. As the number of recommended exercises increases, both SageJavon scores (<span><math><mrow><mi>r</mi><mo>=</mo><mn>0.68</mn></mrow></math></span>) and final exam scores (<span><math><mrow><mi>r</mi><mo>=</mo><mn>0.62</mn></mrow></math></span>) improve, demonstrating the positive impact of personalized recommendations. These results highlight SageJavon as a scalable, plug-and-play solution for programming education.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104605"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190590","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Grouping-enhanced personalization for federated recommendation 用于联合推荐的分组增强个性化

IF 6.9 1区管理学

Information Processing & Management Pub Date : 2026-06-01 Epub Date: 2026-01-08 DOI: 10.1016/j.ipm.2025.104599

Linrui Shen , Anchen Li , Xueyan Liu , Riting Xia , Bo Yang

{"title":"Grouping-enhanced personalization for federated recommendation","authors":"Linrui Shen , Anchen Li , Xueyan Liu , Riting Xia , Bo Yang","doi":"10.1016/j.ipm.2025.104599","DOIUrl":"10.1016/j.ipm.2025.104599","url":null,"abstract":"<div><div>Federated recommendation applies federated learning to recommender systems. The key challenge is how to provide personalized recommendations to clients under privacy constraints and mitigate heterogeneity. Existing methods have several limitations: First, graph convolution tends to over-smooth as training rounds increase, reducing personalization. Second, their frameworks introduce extra parameters, causing high computational and communication costs. To this end, we propose the Grouping-enhanced Personalization for Federated Recommendation (GFRec) framework, which uses a lightweight user grouping aggregation to enable collaborative filtering among users with similar preferences and overcome the over-smoothing. Moreover, we propose SGFRec, which extends GFRec to signed recommendations that handle positive and negative user-item feedback. Extensive experiments show that both GFRec and SGFRec achieve strong performances. Computation and communication overhead experiment results show that GFRec is lightweight. The anonymous code is at <span><span>https://anonymous.4open.science/r/GFRec-D6BE</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 4","pages":"Article 104599"},"PeriodicalIF":6.9,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145927679","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

GRAIN: Gravity-resistance adaptive framework for identifying influential nodes using multi-order structural diversity GRAIN：利用多阶结构多样性识别影响节点的重力-阻力自适应框架

IF 6.9 1区管理学

Information Processing & Management Pub Date : 2026-06-01 Epub Date: 2026-01-13 DOI: 10.1016/j.ipm.2026.104618

Yirun Ruan , Xinghua Qin , Sizheng Liu , Mengmeng Zhang , Jun Tang , Yanming Guo , Tianyuan Yu

{"title":"GRAIN: Gravity-resistance adaptive framework for identifying influential nodes using multi-order structural diversity","authors":"Yirun Ruan , Xinghua Qin , Sizheng Liu , Mengmeng Zhang , Jun Tang , Yanming Guo , Tianyuan Yu","doi":"10.1016/j.ipm.2026.104618","DOIUrl":"10.1016/j.ipm.2026.104618","url":null,"abstract":"<div><div>Identifying influential spreaders is critical for applications like disease control and information dissemination. Existing methods often face trade-offs among accuracy, resolution, and computational cost. To achieve a more balanced and accurate identification, we propose GRAIN (Gravity-Resistance Adaptive Influential Node identification framework), a novel approach that effectively combines multi-order structural diversity and topological cohesion for influence assessment. Specifically, GRAIN introduces two key innovations: (1) A node-adaptive propagation diversity metric based on k-shell entropy. This metric dynamically balances the contributions of a node's direct (1-hop) and indirect (2-hop) neighbors to its diversity through an adaptive weighting parameter <em>λ</em>(<em>v</em>). (2) Resistance distance Ω<em><sub>ij</sub></em> to quantify the topological cohesion between nodes <em>i</em> and <em>j</em>, capturing the \"multiple-route distance diminishment\" effect observed in networks. We integrate these components within a physics-inspired gravity model: A node's \"mass\" combines its core position (k-shell) and neighborhood propagation diversity, while \"distance\" uses Ω<em><sub>ij</sub></em> with exponential attenuation for farther neighbors. A node’s total influence is the sum of gravitational forces from its 1-hop and 2-hop neighbors. Evaluations on 20 real-world and 15 synthetic networks using the Susceptible-Infected-Recovered (SIR) model show that GRAIN significantly outperforms a range of benchmarks in the majority of cases, achieving higher kendall’s tau correlation across a wide range of transmission rates and network topologies.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 4","pages":"Article 104618"},"PeriodicalIF":6.9,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145978422","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

ChartReLA: A compact vision-language model for comprehensive chart reasoning via relationship modeling ChartReLA：一个紧凑的视觉语言模型，用于通过关系建模进行全面的图表推理

IF 6.9 1区管理学

Information Processing & Management Pub Date : 2026-06-01 Epub Date: 2026-01-10 DOI: 10.1016/j.ipm.2025.104608

Xuan-Quang Nguyen , Quang-Tan Nguyen , Long H.B. Nguyen , Dien Dinh

{"title":"ChartReLA: A compact vision-language model for comprehensive chart reasoning via relationship modeling","authors":"Xuan-Quang Nguyen , Quang-Tan Nguyen , Long H.B. Nguyen , Dien Dinh","doi":"10.1016/j.ipm.2025.104608","DOIUrl":"10.1016/j.ipm.2025.104608","url":null,"abstract":"<div><div>We introduce ChartReLA, a compact vision-language model designed for comprehensive chart understanding and reasoning. Unlike existing models that either lack instruction-awareness or ignore inter-element relationships in charts, ChartReLA combines instruction-guided visual feature extraction with a novel Relationship Modeling Adapter. To train the model, we develop a new Referring Question-Answering (RefQA) objective and curate a large-scale dataset of over 1.3 million multimodal samples, including 318,986 RefQA annotations. ChartReLA demonstrates strong performance on public benchmarks: it achieves a 6.84% improvement in RMS F1 over UniChart on the Chart-to-Table task and 2.43% higher accuracy on ChartQA. While maintaining only 326M parameters, it performs competitively in Chart Summarization, narrowing the BLEU gap with larger models to around 1.0 point. These results highlight ChartReLA’s efficiency and effectiveness for instruction-driven chart reasoning. The source code is available at <span><span>https://github.com/nxquang-al/ChartReLA</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 4","pages":"Article 104608"},"PeriodicalIF":6.9,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145978419","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0