Information Processing & Management最新文献

{"title":"MIMOSA: A unified multi-view multi-order contrastive learning framework for bundle recommendation","authors":"Xiangyu Li ,&nbsp;Yao Mu ,&nbsp;Yuying Lin","doi":"10.1016/j.ipm.2025.104446","DOIUrl":"10.1016/j.ipm.2025.104446","url":null,"abstract":"<div><div>Bundle recommendation has emerged as a vital online service that provides users with personalized collections of items likely to attract them. While existing methods attempt to integrate multi-view information, they often struggle to effectively leverage associations between entities (i.e., users and bundles), particularly in distinguishing between associated and non-associated entities. To address this, our study proposes an innovative bundle recommendation approach termed multi-view multi-order contrastive learning (MIMOSA), which simultaneously models the underlying relationships across multiple views and between different entities. The approach introduces a unified graph-based contrastive learning framework that organizes both intra- and inter-entity associations using different orders of proximity within the user-bundle graph. Specifically, MIMOSA customizes tactics for positive sample discovery and contrastive loss calculation to capture the heterogeneous semantics of various entity associations. This enables the alignment of associated entity embeddings while effectively dispersing non-associated ones. Additionally, a center-matching strategy is designed to efficiently coordinate multi-view entity representations, thereby accelerating the contrastive learning process. Extensive experiments on two large-scale datasets, Youshu and NetEase, demonstrate MIMOSA’s superior performance over baseline methods. The results show that compared to the best baseline, our proposed approach achieves average improvements of 2.87% (R@20) and 3.02% (N@20) on Youshu, and 5.19% (R@20) and 4.07% (N@20) on NetEase.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 2","pages":"Article 104446"},"PeriodicalIF":6.9,"publicationDate":"2025-10-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145361775","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":"DLGTrust: Graph neural network-based trust evaluation using dynamic line graph","authors":"Minglong Cheng ,&nbsp;Tingting Xu ,&nbsp;Wei Chen ,&nbsp;Weidong Fang ,&nbsp;Minda Yao ,&nbsp;Jueting Liu ,&nbsp;Zehua Wang","doi":"10.1016/j.ipm.2025.104455","DOIUrl":"10.1016/j.ipm.2025.104455","url":null,"abstract":"<div><div>Trust serves as the foundation for ensuring secure interactions among network entities. However, existing trust evaluation models suffer from vulnerability to global attacks, dependency on multi-layer stacking, and difficulty adapting to dynamically sparse networks. To address these limitations, a graph neural network-based dynamic and robust trust evaluation model is proposed, named DLGTrust. Dynamic line graph snapshots are used to explicitly map indirect trust to direct connections, enhancing the model’s capability to capture complex interactions and its adaptability to sparse data. By integrating a multimodal spatial feature extraction and gated recurrent unit-driven spatiotemporal fusion mechanism, fine-grained modeling of complex interactions is achieved. Simultaneously, adversarial perturbation injection and global robustness constraints are introduced to enhance the model’s defense against global attacks. Experimental results on three real-world datasets show that the comprehensive performance of DLGTrust is improved by at least 26.3% compared to the baseline model. The F1-macro in both observed and unobserved node scenarios is over 98%. Under bad-mouthing, good-mouthing, and global attack rates each set to 10%, the F1-macro is improved by 24.5%, 23.7%, and 76.7%, respectively. The robustness and defense capability of DLGTrust are enhanced. Consequently, DLGTrust offers effective support for secure interactions among entities.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 2","pages":"Article 104455"},"PeriodicalIF":6.9,"publicationDate":"2025-10-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145361776","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":"Ophiuchus: Privacy-preserving training service with user-controlled pseudo-noise information generation","authors":"Longlong Sun ,&nbsp;Hui Li ,&nbsp;Qingcai Luo ,&nbsp;Yanguo Peng ,&nbsp;Jiangtao Cui","doi":"10.1016/j.ipm.2025.104443","DOIUrl":"10.1016/j.ipm.2025.104443","url":null,"abstract":"<div><div>Cloud-based learning services face the risk of privacy information leakage. Thus, many cryptography-based inference schemes have been proposed. However, the requirement for protecting backpropagation and gradient updates makes training much more complex and costly than inference. To fill the gap between inference and training, we optimize the amount of cryptographic computation during the backpropagation. Specifically, we demonstrate that parameter gradients are separable, thus extending an existing private inference scheme to private training. Furthermore, we pioneer the integration of normalizer-free learning into private training, circumventing the normalization layers, which are cryptography-unfriendly. Moreover, to defend against reconstruction attacks, we construct pseudo-noise information by introducing contrastive loss, which is based on the confidentiality of labels and the randomness of positive–negative pairs. Putting it all together, we propose <em>Ophiuchus</em>, a private CNN training framework for image recognition. Empirical results on several benchmark datasets demonstrate that <em>Ophiuchus</em> achieves accuracy comparable to plain training and the backpropagation only incurs an additional overhead of <span><math><mrow><mn>2</mn><mo>.</mo><mn>3</mn><mtext>%–</mtext><mn>5</mn><mo>.</mo><mn>6</mn><mtext>%</mtext></mrow></math></span>. Our scheme can improve the performance between <span><math><mrow><mn>1</mn><mo>.</mo><mn>6</mn><mo>×</mo><mtext>–</mtext><mn>4</mn><mo>.</mo><mn>1</mn><mo>×</mo></mrow></math></span> compared to the current private training schemes. Notably, the constructed pseudo-noise outperforms random noise in both aspects of privacy and utility.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 2","pages":"Article 104443"},"PeriodicalIF":6.9,"publicationDate":"2025-10-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145361774","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":"HAC-FRL: A learning-driven distributed task allocation framework for large-scale warehouse automation","authors":"Hanjie Gu ,&nbsp;Yanyong Feng ,&nbsp;Deke Yu ,&nbsp;Junwei Fang ,&nbsp;Yuliang Sun ,&nbsp;Fengjun Hu ,&nbsp;Ezzeddine Touti","doi":"10.1016/j.ipm.2025.104430","DOIUrl":"10.1016/j.ipm.2025.104430","url":null,"abstract":"<div><div>Hybrid Auction-Consensus with Fine-tuned Recurrent Learning (HAC-FRL) is a framework that is presented in this research for the purpose of distributed task allocation in large-scale warehouse automation. For the purpose of enhancing conflict resolution, accelerating recovery, and optimizing energy use, HAC-FRL incorporates proximal optimization with a mine blast algorithm for training data execution. Unlike previous approaches, which are plagued by agent conflicts, inefficient learning, and poor deadlock recovery, HAC-FRL gives robots the ability to dynamically alter their strategy prior to the assignment of tasks. When compared to baseline approaches, simulation trials using 1000 robots and 5000 tasks indicate considerable gains. These advantages include a 26.2 % increase in task success rate, a 1.24 % reduction in deadlocks, an 84 % faster recovery, a 38 % higher energy efficiency, and a 62 % lower message loss. In light of these findings, it is clear that HAC-FRL offers a solution that is both fault-tolerant and scalable, enabling multi-agent task allocation that is both reliable and efficient in terms of energy consumption for mission-critical systems. The system that has been suggested improves the dependability and scalability of warehouse automation by guaranteeing that learning is efficient and distributed coordination is resilient.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 2","pages":"Article 104430"},"PeriodicalIF":6.9,"publicationDate":"2025-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145320901","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":"Surpassing probabilistic based community detection in flow-based mobility networks","authors":"Yanzhong Yin ,&nbsp;Qunyong Wu","doi":"10.1016/j.ipm.2025.104441","DOIUrl":"10.1016/j.ipm.2025.104441","url":null,"abstract":"<div><div>Understanding community structures in flow-based mobility networks is critical for analysing regional integration patterns, yet existing methods face two key limitations: (1) current modularity optimization algorithms struggle with resolution limits, and (2) failure to combine both local and global community detection method in flow-based mobility networks. To address these gaps, this study develops a novel framework integrating surpassing probability theory with community detection. The surpassing probability-based Leiden method (SPBL) first reshuffles flow weights to overcome resolution limits in the Leiden algorithm, enabling identification of macro-communities. Next, the two-phase surpassing probability community detection (TPSPCD) algorithm systematically decomposes these communities into granular sub-communities while preserving critical anchor relationships. The framework further introduces an Aggregate Surpassing Degree (ASD) metric to quantify the relative strength of internal versus external community connections. Our results revealed distinct core-periphery patterns within flow-based mobility networks, with strong community cohesion around key node centres. This study concludes that the proposed community detection method effectively captures localized interactions in flow-based mobility networks. This work advances both the theory and application of community detection in flow-based mobility networks, offering planners actionable tools for regional development.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 2","pages":"Article 104441"},"PeriodicalIF":6.9,"publicationDate":"2025-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145320902","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":"XR-based cultural heritage information spaces: Towards a new paradigm of information encoding and decoding guided by embodied cognition","authors":"Xiaoguang Wang ,&nbsp;Jingyi Fu ,&nbsp;Sipeng Luo ,&nbsp;Ke Zhao ,&nbsp;Qingyu Duan","doi":"10.1016/j.ipm.2025.104445","DOIUrl":"10.1016/j.ipm.2025.104445","url":null,"abstract":"<div><div>Embodied engagement in immersive information contexts is an emerging trend in informatics research. This study conceptualises XR-based cultural heritage information spaces as multimodal, open-ended environments shaped by embodied interaction. Based on a meta-synthesis of 67 empirical studies, we identify 22 core components and eight main categories across five dimensions of embodied interaction, forming a theoretical model of information flow and cognitive development in XR-based cultural heritage spaces. Embodied interaction is framed as a dual-phase information process: structured encoding of physical, sociocultural, and embodied elements, and user-driven decoding through emotional-aesthetic resonance and cognitive integration. The human body is positioned as an active human-computer interaction interface and processor, enabling the transformation of encoded content into self-constructed meaning. We propose a dual-path evaluation framework that combines system-level performance metrics with embodied-level cognitive and affective outcomes to assess interaction effectiveness. Furthermore, we highlight the role of physiological measures in capturing real-time user responses and enhancing the observability and interpretability of encoding-decoding dynamics. This study advances a human-centred perspective on information processing and offers actionable insights for analysing information behaviour in immersive environments.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 2","pages":"Article 104445"},"PeriodicalIF":6.9,"publicationDate":"2025-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145320903","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":"Unlocking knowledge-sharing live streaming e-commerce: An LLM-empowered analytics framework for book sales prediction","authors":"Runyu Chen,&nbsp;Junru Xiao,&nbsp;Luqi Chen,&nbsp;Xiaohe Sun","doi":"10.1016/j.ipm.2025.104444","DOIUrl":"10.1016/j.ipm.2025.104444","url":null,"abstract":"<div><div>Streamers’ discourse plays a key role in shaping purchasing decisions in live streaming e-commerce, especially in knowledge-sharing formats where product promotion is combined with information delivery. Previous studies have shown that streamers’ discourse can influence product sales, with few studies systematically extracting semantic features across different dimensions and quantifying their impact on sales prediction performance. The main contribution of our research is the design of a predictive framework for sales in knowledge-sharing live streaming. The framework integrates social support theory with fine-tuned large language models (LLMs) to systematically extract multi-dimensional semantic cues from streamers’ discourse for sales prediction. We collected data from 80 live streams across 35 Douyin rooms over two months for our experiments. In the social support classification experiment, the fine-tuned Ernie-SFT model outperformed the best baseline LLM, with improvements of 11.12% in accuracy, 11.87% in weighted F1-score, and 7.83% in macro F1-score. In the sales prediction experiments, we validated the proposed framework using four mainstream classifiers and observed consistent performance gains. The best-performing classifier achieved improvements of 12.53% in accuracy, 10.83% in weighted F1-score, and 4.24% in macro F1-score. These findings highlight the strong predictive value of social support features embedded in streamers’ discourse, offering actionable insights for streamers and enabling data-driven optimization strategies for platforms.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 2","pages":"Article 104444"},"PeriodicalIF":6.9,"publicationDate":"2025-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145320910","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":"Advancing cross-domain emergency classification with multi-view adversarial learning","authors":"Yuhan Xie ,&nbsp;Chen Lyu ,&nbsp;Zheng Qu ,&nbsp;Chunmei Liu","doi":"10.1016/j.ipm.2025.104442","DOIUrl":"10.1016/j.ipm.2025.104442","url":null,"abstract":"<div><div>The growing volume of natural and man-made emergency data requires effective real-time classification across various emergency domains on social media. However, current Unsupervised Domain Adaptation (UDA) methods for emergency data classification face two key challenges: predominant grounding in natural disaster contexts that limits generalizability, and difficulty handling domain shifts caused by heterogeneous distributions, linguistic variations, and emotional expressions. To overcome these challenges, we propose Multi-View Adversarial Neural Networks for Robust Unsupervised Domain Adaptation (MARDA), a novel framework that integrates adversarial domain adaptation with multi-view feature learning. First, a cross-view processor consisting of semantic and emotional processors, along with interactive integrators, is designed to generate rich and comprehensive multi-view feature representations. Second, an adaptive weighted domain enhancer is developed to dynamically balance contributions from multiple views, effectively aggregating discriminative information in various domains. Third, an adversarial cross-view optimizer is proposed that employs a minimax game and feature consistency regularization, thereby enhancing cross-domain generalization. Experimental results on four real-world emergency datasets with 24,008 samples show that MARDA outperforms advanced baselines by 7.39% and exceeds large language models by 1.59% in average F1-Score, demonstrating its effectiveness as a generalized solution for cross-domain emergency event classification.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 2","pages":"Article 104442"},"PeriodicalIF":6.9,"publicationDate":"2025-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145320909","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":"Unraveling urban problem patterns in Zhejiang: Evolutionary trajectories and driving factors based on public complaint data","authors":"Xiangfu Kong,&nbsp;Bo Dong","doi":"10.1016/j.ipm.2025.104438","DOIUrl":"10.1016/j.ipm.2025.104438","url":null,"abstract":"<div><div>This study systematically investigates the formation, evolution, and drivers of urban problem patterns across 87 county-level administrative units in Zhejiang Province, China. By classifying approximately 3.5 million public complaints into 37 distinct categories, urban problem distributions in these counties are quantified from 2018 to 2022. The cluster analysis reveals seven distinct problem patterns: pollution and municipal infrastructure, consumer protection, land use, housing purchase, garbage management, financial service, and pandemic control, which exhibit marked temporal persistence and spatial clustering. Using fuzzy logic to calculate membership degrees, the evolution of problem patterns exhibits predictable trajectories: emerging patterns showed consistent growth, while dominant patterns gradually declined, with occasional deviations. Analysis of eight socioeconomic factors indicates that development-driven patterns—land use, housing purchase, and consumer protection—are strongly correlated with socioeconomic indicators such as land acquisition area, property sales volume, and retail activity, necessitating enhanced regulatory oversight of economic activities. Trend-driven patterns, exemplified by pollution and municipal infrastructure problems, correlate less with short-term fluctuations and more with stable socioeconomic conditions, underscoring the importance of systematic planning and coordinated interventions to mitigate structural regional imbalances. Event-driven patterns, including garbage management, financial service, and pandemic control, demonstrate minimal association with socioeconomic factors, indicating their susceptibility to notable events and highlighting the need for rapid crisis management.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 2","pages":"Article 104438"},"PeriodicalIF":6.9,"publicationDate":"2025-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145320904","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-domain sequential recommendation via multi-sequence and multi-task learning","authors":"Liwei Pan ,&nbsp;Weike Pan ,&nbsp;Zhong Ming","doi":"10.1016/j.ipm.2025.104426","DOIUrl":"10.1016/j.ipm.2025.104426","url":null,"abstract":"<div><div>In recent years, more and more researchers and practitioners have focused on multi-domain CTR prediction and achieved great success. Though users’ behaviors often exhibit sequentiality, little effort has been made on multi-domain sequential recommendation (MDSR). Most existing works on MDSR sort the interactions from all domains in chronological order and then predict the next interacted items in each domain. However, they neglect separate interaction sequences in each domain. Therefore, they cannot exploit the commonalities and differences among different domains well. Cross-domain sequential recommendation (CDSR) models are usually designed for performance improvement in one target domain rather than in each domain. Although extending a CDSR model to an MDSR one directly or indirectly is feasible, it will result in high time complexity. Meanwhile, they often ignore data imbalance across different domains, which might cause negative transfer.</div><div>As a response, we propose a novel MDSR solution called multi-sequence multi-task learning (MML). Our MML consists of three modules, including hybrid-domain sequential preference learning (HSPL), intra-domain sequential preference learning (ISPL) and multi-task learning &amp; prediction (MLP). Specifically, HSPL aims to learn hybrid-domain sequential preferences. Meanwhile, we construct augmented sequences and leverage contrastive learning to learn more unbiased hybrid-domain sequential preferences for alleviating negative transfer. ISPL is designed to capture intra-domain sequential preferences. In the MLP module, three specific tasks and a behavior regularizer are leveraged to ensure that each module can learn the corresponding preferences sufficiently and enhance knowledge transfer among different domains. We conduct extensive experiments on some public datasets using different backbone models and show that our MML is able to achieve significantly better performance than the state-of-the-art methods in two or more domains in most cases. Meanwhile, our MML can achieve the same time complexity as the MDSR models only using hybrid interaction sequences. The source code can be found at <span><span>https://github.com/plw2019/MML</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 2","pages":"Article 104426"},"PeriodicalIF":6.9,"publicationDate":"2025-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145320911","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}