Information Processing & Management最新文献

{"title":"Causal inference for alleviating confounding bias in multi-criteria rating recommendation","authors":"Zhihao Guo ,&nbsp;Peng Song ,&nbsp;Chenjiao Feng ,&nbsp;Kaixuan Yao ,&nbsp;Jiye Liang","doi":"10.1016/j.ipm.2025.104364","DOIUrl":"10.1016/j.ipm.2025.104364","url":null,"abstract":"<div><div>Integrating multi-criteria (MC) ratings into recommender systems can enhance the service quality of online platforms. MC ratings depict more fine-grained user preferences from multiple dimensions, such as a hotel system, including ratings for overall, location, cleanliness, etc. The existing MC methods focus on mining the correlation from historical interactions through the data-driven paradigm. However, the traditional methods may capture spurious association in biased observations due to various confounders, which can reduce prediction accuracy. So far, research on how to alleviate confounding bias in MC rating recommendation scenarios remains unexplored. To fill this research gap, we propose a novel <em>Deconfounding Multi-Criteria Recommendation</em> (DMCR) framework, which is used to mitigate the harmful impact triggered by confounders. Specifically, we block the back-door paths that cause bias through the front-door adjustment and estimate the causal effect between user-item pair and overall rating. In the inference phase, the DMCR approximates the outcome after intervention by conditional probabilities on the observational MC data. Moreover, we leverage graph neural network to model underlying higher-order dependencies in MC ratings. This modeling scheme helps to develop the heterogeneity of user MC behavioral preferences. Experimental results on six real datasets demonstrate that the DMCR outperforms the existing baselines.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 1","pages":"Article 104364"},"PeriodicalIF":6.9,"publicationDate":"2025-08-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144893477","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}

{"title":"CLIP-based knowledge projector for image–text matching","authors":"Xinfeng Dong ,&nbsp;Dingwen Zhang ,&nbsp;Longfei Han ,&nbsp;Huaxiang Zhang ,&nbsp;Li Liu ,&nbsp;Junwei Han","doi":"10.1016/j.ipm.2025.104357","DOIUrl":"10.1016/j.ipm.2025.104357","url":null,"abstract":"<div><div>Image–text matching is an essential research area within multimedia research. However, images often contain richer information than text, and representing an image with only one vector can be limited to fully capture its semantics, leading to suboptimal performance in cross-modal matching tasks. To address this limitation, we propose a CLIP-based knowledge projector network that encodes an image into a set of embeddings. These embeddings capture different semantics of an image, guided by prior knowledge from the large vision-language pretrained model CLIP(Contrastive Language-Image Pre-Training). To ensure that the generated slot features stay aligned with global semantics, we design an adaptive weighted fusion module that incorporates global features into slot representations. During the test phase, we present an effective and explainable similarity calculation method compared with existing fine-grained image–text matching methods. The proposed framework’s effectiveness is evidenced by the experimental results, with performance improvements of at least 7% in R@1 on image retrieval tasks compared to CLIP on the MSCOCO and Flickr30K datasets.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 1","pages":"Article 104357"},"PeriodicalIF":6.9,"publicationDate":"2025-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144890115","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}

{"title":"A dual-branch multi-path propagation reasoning network for rumor detection integrating neural symbolic commonsense reasoning mechanism","authors":"Weiming Yin ,&nbsp;Jinzhong Ning ,&nbsp;Mingyu Lu ,&nbsp;Hongfei Lin ,&nbsp;Yijia Zhang","doi":"10.1016/j.ipm.2025.104362","DOIUrl":"10.1016/j.ipm.2025.104362","url":null,"abstract":"<div><div>With the widespread dissemination of rumors on social media platforms, achieving automated rumor detection in the early stage has become an important challenge. To this end, we propose a <strong>D</strong>ual-branch <strong>M</strong>ulti-path <strong>P</strong>ropagation <strong>R</strong>easoning <strong>N</strong>etwork (DMPRN) for rumor detection. For branch 1: to simulate various human thinking chains, we calculate the centrality of the nodes in the propagation graph and use pruning methods to construct propagation graphs of different paths. Then, we use the Graph Convolutional Network to capture the rumor propagation structure. For branch 2: to simulate human logical reasoning based on common sense, we design a Neural-Symbolic Commonsense Reasoning Mechanism. First, we use the Transformer network and the commonsense knowledge graph to dynamically reason about the commonsense knowledge related to tweets. Then, we use neural-symbolic learning to denoise the knowledge and tweets. Finally, we use logic operators <span><math><mo>∧</mo></math></span> and <span><math><mo>∨</mo></math></span> to integrate the knowledge with the rumor content. The model achieves accuracies of 89.7%, 91.4%, and 78.6% on three publicly available datasets. Compared to state-of-the-art baseline methods, our approach improves accuracy by up to 3% across all three datasets. Moreover, experiments demonstrate that the proposed method is effective for early rumor detection.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 1","pages":"Article 104362"},"PeriodicalIF":6.9,"publicationDate":"2025-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144892134","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}

{"title":"A quasi-experimental analysis of capabilities and limitations of generative AI in academic content evaluation in social sciences","authors":"Yu Zhu ,&nbsp;Yongrong Lu ,&nbsp;Huan Xie ,&nbsp;Jiyuan Ye ,&nbsp;Ming Chen","doi":"10.1016/j.ipm.2025.104365","DOIUrl":"10.1016/j.ipm.2025.104365","url":null,"abstract":"<div><div>The complexity of social sciences research and the limitations of traditional evaluation methods highlight the need to explore the capabilities and application potential of generative AI in academic evaluation. Previous research in fields such as biomedical and other natural sciences has demonstrated the potential of generative AI to estimate the quality of research articles. This study adopts a quasi-experimental approach, 100 volunteers produced 600 social sciences academic texts across 6 types of topics, which were evaluated by 8 mainstream generative AI models. Statistical and sentiment analysis was conducted to compare the evaluation results using zero-shot and few-shot prompting strategies. The results show that AI-generated total scores are unreliable (precision = 66.35 %), and the actual total scores differ moderately from the human benchmark (average Cohen's <em>d</em> = 0.425). Few-shot prompt exhibited weaker differentiation capabilities across dimensions (average correlation = 5.25), while zero-shot prompt performed better (e.g., correlation_{Clarity, Significance} = 0.13), particularly in writing quality (average standard deviation = 5.38). Significant score differences were observed across the eight models (all <em>p</em> &lt; 0.001), indicating inconsistency among models. Additionally, AI-generated comments across dimensions were generally positive, with different models exhibiting strengths across various dimensions and tasks. This study provides empirical evidence for scholars, peer reviewers, and research evaluation professionals interested in integrating generative AI into social sciences’ evaluation workflows. Overall, generative AI shows potential for enhancing evaluation efficiency and reducing favoritism in the peer review of social sciences, especially in large-scale or preliminary evaluations. However, when evaluating the novelty and significance, its dependency on domain knowledge and the interpretability of the results still requires prudent consideration and refinement.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 1","pages":"Article 104365"},"PeriodicalIF":6.9,"publicationDate":"2025-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144890114","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}

{"title":"Retrieval-enhanced, Adaptively Collaborative, and Temporal-aware user behavior comprehension for LLM-based sequential recommendation","authors":"Zheng Hu ,&nbsp;Yongsen Pan ,&nbsp;Zetao Li ,&nbsp;Jiaming Huang ,&nbsp;Satoshi Nakagawa ,&nbsp;Jiawen Deng ,&nbsp;Shimin Cai ,&nbsp;Fuji Ren","doi":"10.1016/j.ipm.2025.104354","DOIUrl":"10.1016/j.ipm.2025.104354","url":null,"abstract":"<div><div>The rapid advancement of large language models (LLMs) presents new opportunities for recommender systems. However, LLM-based sequential recommenders often struggle to extract effective user interest signals from long and complex behavior sequences, leading to the <em>sequential behavior incomprehension problem</em>. To address this, we propose <strong>Re</strong>trieval-enhanced <strong>A</strong>daptive <strong>C</strong>ollaborative- and <strong>T</strong>emporal-aware user behavior comprehension (ReACT), a retrieval-augmented framework that empowers LLMs to better model user interests. ReACT introduces: (i) Temporal Pointwise Mutual Information (TPMI), which integrates temporal and collaborative signals to retrieve relevant historical behaviors; and (ii) Adaptive User Behavior Retrieval (AUBR), which dynamically selects the most informative user behaviors for each recommendation. Extensive experiments on three real-world datasets (MovieLens-1M, Amazon-Book, and MovieLens-100K) demonstrate that ReACT achieves competitive performance while utilizing only approximately 20% of the average user sequence and 5% of the training data. An LLM-as-judger evaluation across three datasets demonstrates that our method achieves the highest selection ratios (78%, 64%, and 64%, respectively), showing that the user behaviors retrieved by ReACT are the most informative and interpretable for LLM-based user behavior comprehension.<span><span>¹</span></span></div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 1","pages":"Article 104354"},"PeriodicalIF":6.9,"publicationDate":"2025-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144886761","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}

{"title":"LLM-EPSP: Large language model empowered early prediction of student performance","authors":"Huawei Zhou ,&nbsp;Shuanghong Shen ,&nbsp;Yu Su ,&nbsp;Yongchun Miao ,&nbsp;Qi Liu ,&nbsp;Linbo Zhu ,&nbsp;Junyu Lu ,&nbsp;Zhenya Huang","doi":"10.1016/j.ipm.2025.104351","DOIUrl":"10.1016/j.ipm.2025.104351","url":null,"abstract":"<div><div>Early prediction of student performance (EPSP) has garnered significant attention due to its educational value, especially its importance in academic early warning systems. State-of-the-art data mining methods have achieved remarkable success by optimizing feature selection and model enhancements. However, these methods often face challenges, including the cold-start problem, limited exploration of the intrinsic relationships among features, and poor generalization. In this work, we explore the utilization of Large Language Models (LLMs) as information integrators to address these challenges and propose a novel model called Large Language Model Empowered Early Prediction of Student Performance (LLM-EPSP). Specifically, for the cold-start problem, LLM-EPSP benefits from the inherent advantages of LLMs, which stem from their extensive pretraining on diverse datasets. This enables the model to make informed predictions even with limited initial data. For exploring intrinsic relationships among features, LLM-EPSP employs feature fusion techniques to uncover underlying connections between various features, ensuring a comprehensive and robust analysis. To enhance the generalization capabilities of LLM-EPSP, we develop predefined templates that facilitate its adaptation to a wide range of educational contexts. We evaluate our method on two real-world datasets: (1) OULAD, which includes data on 22 courses and 32,593 students, and (2) the UCI Machine Learning Repository, which contains 23 types of features from 649 students. Extensive validation demonstrates that LLM-EPSP considerably outperforms baseline approaches across diverse scenarios. Further analysis results also demonstrate the robustness and versatility of LLM-EPSP, suggesting its enormous potential in practical applications.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 1","pages":"Article 104351"},"PeriodicalIF":6.9,"publicationDate":"2025-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144886755","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}

{"title":"Orchestrating mechanics, perception and control: Enabling embodied intelligence in humanoid robots","authors":"Jiahang Huang,&nbsp;Junyao Gao,&nbsp;Zhangguo Yu","doi":"10.1016/j.ipm.2025.104363","DOIUrl":"10.1016/j.ipm.2025.104363","url":null,"abstract":"<div><div>Humanoid robotics has evolved from early bipedal locomotion studies to modern systems integrating neuromorphic intelligence. This review systematically examines nearly 300 research studies, identifying key advancements in biomechanical optimization, multimodal perception, motion intelligence, and intelligent interaction. Recent progress in biomechanical optimization through material-structure co-design has led to lighter, more adaptive robotic frameworks, improving energy efficiency, compliance, and mechanical robustness. Meanwhile, multimodal perception has significantly enhanced environmental understanding by integrating vision, force, and proprioceptive sensing, enabling robust scene interpretation and adaptive interaction. However, challenges remain in real-time sensor fusion and uncertainty handling, limiting performance in dynamic and unstructured environments. Advancements in motion intelligence are increasingly driven by frameworks that integrate model-based control with learning-driven adaptation, allowing humanoid robots to achieve greater efficiency, agility, and generalizability in motion planning and execution. At the same time, intelligent interaction has evolved with approaches such as imitation learning, shared control, brain-computer interfaces, teleoperation, and large models, strengthening the link between perception and action for seamless human-robot collaboration. While these innovations enhance adaptability and interaction efficiency, robustness in intent-driven decision-making and real-world deployment remains a key challenge. Commercialization efforts have accelerated the transition from laboratory prototypes to practical applications, particularly in industrial automation and assistive robotics. However, scalability, autonomy, and safety remain critical concerns, requiring further advancements in hardware efficiency, neuromorphic computing, and AI-driven architectures. By synthesizing theoretical insights with recent technological developments, this review provides a structured roadmap for advancing humanoid robotics toward real-world implementation.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 1","pages":"Article 104363"},"PeriodicalIF":6.9,"publicationDate":"2025-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144886754","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}

{"title":"Multi-resolution leak detection based on shared expert MoE forecasting for natural gas pipelines","authors":"Xuguang Li ,&nbsp;Zhonglin Zuo ,&nbsp;Zheng Dong ,&nbsp;Hongke Zhao ,&nbsp;Luanfei Wan ,&nbsp;Hongfang Cheng","doi":"10.1016/j.ipm.2025.104353","DOIUrl":"10.1016/j.ipm.2025.104353","url":null,"abstract":"<div><div>Natural gas is a critical strategic energy resource, predominantly transported through extensive pipeline networks monitored by Supervisory Control and Data Acquisition (SCADA) systems. Developing accurate deep-learning models for pipeline leak detection using SCADA data is crucial for safeguarding this vital infrastructure. Reliable and timely leak detection remains challenging due to two inherent limitations: (1) severe sample imbalance from rare leak occurrences and (2) complex multi-resolution hydraulic patterns complicating leak characterization. To address the challenges, we propose a novel Multi-Resolution Shared-Expert Mixture-of-Experts (MR-SEMoE) framework for leakage detection. The framework employs multivariate time series forecasting, where deviations between predicted and observed sensor values trigger leak alarms through statistical thresholding. Two key innovations synergistically enhance detection performance: (1) a shared-expert MoE architecture improving generalization through cross-experts knowledge transfer. (2) A multi-resolution analysis framework featuring parallel multi-head forecasters with resolution-specific feature extractors that enable hierarchical representation learning across different time resolutions. Comprehensive experimental evaluations on real-world natural gas pipeline datasets demonstrate that the proposed MR-SEMoE effectively identifies leaks under imbalanced data conditions. Compared to the previous state-of-the-art method, MR-SEMoE’s F1-score improved by 1.67%. The MR-SEMoE model outperforms contemporary state-of-the-art approaches, establishing the premier natural gas pipeline leak detection framework. To our knowledge, this work constitutes the first successful implementation of the MoE methodology in this domain, facilitating future deployment of large-scale models.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 1","pages":"Article 104353"},"PeriodicalIF":6.9,"publicationDate":"2025-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144864093","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}

{"title":"Digital orientation, knowledge acquisition, and digitization in non-digital native firms","authors":"Lipeng Pan ,&nbsp;Shuchun Liu ,&nbsp;Yongqing Li","doi":"10.1016/j.ipm.2025.104361","DOIUrl":"10.1016/j.ipm.2025.104361","url":null,"abstract":"<div><div>Understanding how non-digital native firms achieve successful digital transformation remains a significant challenge in information management. This study investigates the digital transformation process from a knowledge-based perspective, using a panel dataset comprising 13,882 firm-year observations from 3109 industrial firms in Zhejiang Province, China (2014–2022). Employing fixed-effects regression and mediation-moderation analyses, the study examines how digital orientation influences digital transformation across three core business processes: procurement, production, and sales. Key findings highlight distinct knowledge acquisition pathways: digital orientation directly enhances production digitization (β = 0.032, <em>p</em> &lt; 0.05), whereas its influence on procurement and sales digitization operates indirectly through tactical (β = 0.058, <em>p</em> &lt; 0.01) and strategic knowledge acquisition (β = 0.041, <em>p</em> &lt; 0.001). Further analysis reveals that absorptive capacity significantly strengthens the effect of tactical knowledge acquisition (β = 0.034, <em>p</em> &lt; 0.1) but does not affect strategic knowledge pathways. The study underscores the importance of targeted knowledge management practices and suggests firms optimize their digital investments and internal training resources to initiate and sustain digital transformation effectively. Moreover, clarifying strategic goals related to digitalization is essential for guiding firms toward consistent innovation. These insights provide practical guidelines for enhancing digital transformation strategies, ultimately assisting traditional firms in overcoming internal capability constraints and fostering sustainable digitization outcomes.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 1","pages":"Article 104361"},"PeriodicalIF":6.9,"publicationDate":"2025-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144864091","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}

{"title":"MSPF: A multi-semantic prompting fusion framework for emotion-cause pair extraction in conversations","authors":"Bo Xie ,&nbsp;Junhao Wang ,&nbsp;Haixia Guo ,&nbsp;Pengliang Chen ,&nbsp;Hua Zhang ,&nbsp;Bo Jiang ,&nbsp;Ye Wang ,&nbsp;Liwen Chen","doi":"10.1016/j.ipm.2025.104356","DOIUrl":"10.1016/j.ipm.2025.104356","url":null,"abstract":"<div><div>Emotion-cause pair extraction in conversations (ECPEC) has garnered increasing attention but struggles with modeling multi-turn utterance dependencies. While semantic prompting improves language understanding, its high computational cost hinders widespread ECPEC adoption. To overcome these constraints, we innovatively develop a multi-semantic prompting fusion (MSPF) framework by introducing the pair-oriented sampling strategy, focusing on candidate utterance pairs and transforming the ECPEC task into a pair verification issue. This shift enables us to incorporate three specialized semantic prompts, including tagging, synonym, and causal claim prompts, designed to enrich the semantics of emotion sentiment and emotion-cause relationships. We further present a knowledge attention module for the integration of tagging and synonym prompts, and a two-layer attention pooling module for merging tagging and dual causal claim prompts. Experimental results demonstrate that our proposed MSPF models outperforms the best existing baselines by 4.91 %, 4.08 %, and 2.86 % in F1 score on the ConvECPE, ECPE-D-DD, and ECPE-D-IE (for the domain adaptation experiment) datasets, respectively, with further ablation analysis confirming the effectiveness of our framework.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 1","pages":"Article 104356"},"PeriodicalIF":6.9,"publicationDate":"2025-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144864092","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}