Information Processing & Management最新文献

{"title":"Coal-gas risk fusion identification and warning based on deep explainable expert learning networks: A case study in mining information systems","authors":"Chengfei Liu ,&nbsp;Enyuan Wang ,&nbsp;Zhonghui Li ,&nbsp;Jiankun Xu ,&nbsp;Shan Yin","doi":"10.1016/j.ipm.2026.104665","DOIUrl":"10.1016/j.ipm.2026.104665","url":null,"abstract":"<div><div>Reliable and interpretable systems for integrated identification and warning of gas hazards are essential for improving workers’ risk recognition and ensuring coal mine safety. Traditional statistical methods lack generalization and fail to capture multi-indicator coupling, whereas deep learning models, despite their accuracy, are limited by poor interpretability. The dataset used in this study was collected from 35 coal mines in Guizhou Province. By integrating field records, risk annotations, and feature analyses, a total of 22,460 samples were generated through a sliding-window approach, with a risk-to-non-risk ratio of approximately 35% to 65%. Based on this dataset and motivated by the aforementioned limitations, we propose a multi-source coal-gas risk fusion early warning method based on a deep-explainable hybrid learning expert network. Data distribution characteristics are adaptively pre-processed using boxplots and MICE, and representative indicators are selected through multi-angle correlation and precursor feature analysis. A one-dimensional multi-scale multi-attention parallel convolution network (MACN) is developed, and explainable risk modeling is achieved by integrating InterpGN with SBM. Experimental results show that the proposed method achieves 97.2% accuracy and 98.6% F1 score, surpassing conventional machine learning and deep learning baselines by 3.90% and 4.35%, respectively. Field validation confirms the capability of the proposed approach to deliver timely and reliable early warnings of coal-gas risks. Unlike post-hoc attribution methods such as SHAP, our model incorporates an endogenous explainable mechanism, marking the first application of such hybrid networks in the field of coal-gas disaster prevention. It provides a new paradigm for intelligent early warning expert information systems and scientific decision-making in coal mines. Significantly enhanced coal mine information integration, risk identification, early warning, and worker safety management capabilities</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104665"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190714","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":"Decoding inner speech via frequency-specific cortical EEG representations","authors":"Yifan Wang ,&nbsp;Hongyu Cai ,&nbsp;Lei Zhou ,&nbsp;Zejia Yang ,&nbsp;Liu Wang ,&nbsp;Jian Zhao ,&nbsp;Zhejun Kuang ,&nbsp;Haiyan Wang ,&nbsp;Dawei Xu","doi":"10.1016/j.ipm.2026.104676","DOIUrl":"10.1016/j.ipm.2026.104676","url":null,"abstract":"<div><div>Neural decoding of inner speech provides a crucial pathway for silent communication and cognitive state recognition. However, existing circular brain electrical activity mapping (BEAM)-based decoding methods suffer from the problem that edge distortion and information loss easily occur, and they are mismatched with the input structure of convolutional networks, ultimately limiting decoding accuracy. To overcome this problem, this paper first proposed a frequency-specific rectangular BEAM sequence construction method that employed an adaptive variation function to generate high-fidelity EEG sequences across different frequency bands, preserving cortical spatial correlations while perfectly matching the rectangular input structure of convolutional networks. Then, based on the continuous temporal and multi-frequency features of rectangular BEAM maps, 3D continuous frequency-spatial and temporal-spatial dynamic representation modules, respectively, were constructed. Finally, a dual-branch 3D convolutional network was designed to achieve joint decoding of inner speech brain activity. The manuscript reorganized the publicly available InnerSpeech dataset, involving 10 subjects, 3 task modalities, and 4 categories, with the first 10 trials selected for each condition. In total, 1200 samples were used to evaluate the proposed model’s ability to accurately classify inner speech. Experimental results showed that the proposed method outperformed the baseline model EEGNet, achieving accuracies of 61.79%, 62.67%, and 53.25% in the Inner, Pron, and Vis classification tasks, respectively, while also providing new insights into the mechanisms of frequency–task coupling and the development of high-precision brain–computer interfaces.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104676"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190570","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":"SR-DANet: Sparse-aware super-resolution network based on dynamic adaptive branch for medical image segmentation","authors":"Haotian Lu ,&nbsp;Desheng Kong ,&nbsp;Xinjian Wei ,&nbsp;Xiaoxuan Xu ,&nbsp;Jing Xu","doi":"10.1016/j.ipm.2026.104660","DOIUrl":"10.1016/j.ipm.2026.104660","url":null,"abstract":"<div><div>Accurate medical image segmentation is critical for reliable diagnosis and treatment planning. However, low-resolution images limit accurate segmentation of small or thin structures, and deeper encoder-decoder networks alone cannot overcome this sampling constraint. Super-resolution helps only when it is jointly modeled with segmentation in a task-aware way, so that the enhanced details truly support precise delineation of anatomically and clinically relevant regions. In addition, most existing deep learning methods fail to exploit the sparse characteristics of medical images. In such images, informative anatomical or pathological structures occupy only a small portion of the entire field, while large background areas contain limited diagnostic value. Conventional networks process all pixels or patches uniformly, resulting in redundant computation and suboptimal feature representation. To address these limitations, we propose SR-DANet, a sparse-aware super-resolution segmentation framework based on dynamic adaptive branch. SR-DANet consists of three key modules: mask generator, hash-based sparse attention mechanism and dynamic adaptive branch module. The mask generator identifies informative regions and guides selective feature refinement. The hash-based sparse attention mechanism restricts attention computation to relevant tokens, improving both accuracy and efficiency. And the dynamic adaptive branch module adjusts receptive fields across layers, enabling flexible feature extraction from fine details to global semantics. We evaluate SR-DANet on BraTS 2021 (1470 cases), BraTS-TCGA-LGG (65 cases), and MSD Colon Cancer (190 cases) datasets. Our method achieves Dice scores of 92.45%, 96.91%, and 93.63%, respectively, outperforming state-of-the-art baselines by 0.45% on average. These results demonstrate that SR-DANet effectively enhances segmentation accuracy and efficiency by leveraging the sparse nature of medical images.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104660"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190568","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":"Bridging the gap in cross-domain graph anomaly detection: Enhanced source utilization and label accuracy","authors":"Cairui Yan ,&nbsp;Xu Cheng ,&nbsp;Likang Wu ,&nbsp;Yingyuan Xiao ,&nbsp;Hongke Zhao ,&nbsp;Wenguang Zheng","doi":"10.1016/j.ipm.2026.104678","DOIUrl":"10.1016/j.ipm.2026.104678","url":null,"abstract":"<div><div>Cross-domain graph anomaly detection is crucial for maintaining data security and system stability, as it can identify and respond to anomalous behaviors spanning various data domains. However, current methods are limited by their inability to fully leverage the complex source domain information and accurately assign labels due to the ambiguity and variability of anomalies across domains. To address these challenges, we propose an innovative approach, namely Bridging the Gap in Cross-Domain Anomaly Detection (BGAD). Our method employs high-frequency information-enriched node initialization and anomaly-aware denoised auto-encoders to deeply mine insights from the source domain and align them with the target domain. Additionally, a feedback-driven label shifting correction mechanism dynamically enhances label prediction precision. Our method has been rigorously tested on four different datasets. Compared with the most advanced CD-GAD method CDFS, BGAD has increased the AUC-ROC by 15.7%, significantly reducing false positives and false negatives, demonstrating outstanding effectiveness and practicality.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104678"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190667","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":"Functional data clustering and outlier identification via subgroup analysis with minimum spanning tree","authors":"Chao Liu ,&nbsp;Shimin Lin ,&nbsp;Jiaqi Li","doi":"10.1016/j.ipm.2026.104646","DOIUrl":"10.1016/j.ipm.2026.104646","url":null,"abstract":"<div><div>Functional data clustering with outliers poses critical challenges in cluster number selection consistency, computational efficiency, and robustness to contamination. To address these issues, this paper proposes a novel functional data subgroup clustering algorithm based on minimum spanning tree (FDSC-MST) method. Specifically, FDSC-MST constructs a minimum spanning tree (MST) using both geographical distance and functional curve similarity, then applies subgroup pairwise penalties only to MST edges. An alternating direction method of multipliers (ADMM) is utilized to solve the optimization problem. For contaminated data, we further propose an effective functional data subgroup clustering and outlier detection algorithm via minimum spanning tree (FDSCOD-MST), which simultaneously performs clustering and outlier detection while enhancing robustness and retaining computational efficiency. Comprehensive simulations demonstrate that FDSC-MST outperforms existing methods in cluster number selection consistency and clustering accuracy, even under outlier contamination. When compared to state-of-the-art methods with consistent cluster number selection, the proposed FDSC-MST requires only 17% of the computational time when the sample size is 200, with time savings becoming more significant as sample size grows. Validation on two real-world datasets (bike sharing system data and travel trajectory data) further confirms the practical effectiveness of the proposed framework.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104646"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190670","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":"Diffusion -driven group anomaly detection in spatiotemporal trajectories: Robust masked sequence imputation for enhanced pattern discovery","authors":"Qi Ouyang ,&nbsp;Hongchang Chen ,&nbsp;Shuxin Liu ,&nbsp;Ran Li ,&nbsp;Kai Wang ,&nbsp;Yingle Li","doi":"10.1016/j.ipm.2026.104650","DOIUrl":"10.1016/j.ipm.2026.104650","url":null,"abstract":"<div><div>Detecting group anomalies in spatiotemporal trajectories is vital for smart city security but remains challenging due to noise and complex dynamics. Existing diffusion models often suffer from “conditional contamination,” where observed anomalies inadvertently guide reconstruction, masking the very deviations they aim to detect. To address this, we propose Diffusion-Driven Group Anomaly Detection (DGAD). Specifically, we introduce an Interleaved Window Masking Strategy that segments data to enforce mutual supervision and reveal latent patterns. We pair this with an Unconditional Imputation Mechanism that conditions generation on forward noise instead of partial observations. This prevents anomaly leakage and significantly widens the divergence between normal and abnormal behaviors. Furthermore, a Denoising Weighted Voting module aggregates outputs across diffusion steps to mitigate uncertainty and enhance stability. Extensive experiments on synthetic and real-world datasets show that DGAD consistently outperforms state-of-the-art methods, improving F1-score by 1.1% on average and 1.8% in high-noise conditions. Code and case studies are available at [<span><span>https://github.com/oyq-star/DAGD-main</span><svg><path></path></svg></span>].</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104650"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146081586","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":"GAI-HIQ: Developing a health information quality assessment indicator system for generative artificial intelligence","authors":"Jia Yi ,&nbsp;Fei Du ,&nbsp;Yuchen Nie ,&nbsp;Wencui Liang ,&nbsp;Xiaoyu Zhou ,&nbsp;Jinjuan Chen ,&nbsp;Guangying Li ,&nbsp;Mimi Liu ,&nbsp;Yalan Lv ,&nbsp;Wenlong Zhao ,&nbsp;Xiaorong Hou","doi":"10.1016/j.ipm.2026.104651","DOIUrl":"10.1016/j.ipm.2026.104651","url":null,"abstract":"<div><div>The rapid adoption of generative artificial intelligence (GAI) in healthcare has raised concerns about the quality of its outputs, yet existing health information assessment tools are not designed for AI-generated content. This study developed the GAI-HIQ assessment scale. Using Latent Dirichlet Allocation on 341 relevant publications, we initially identified three core dimensions and 15 secondary indicators for evaluating health information quality. A two-round Delphi consultation with 20 experts (100 % response rate) refined the framework to three core dimensions and 13 secondary indicators, achieving a statistically significant level of expert consensus (Kendall’s <em>W</em> = 0.288, <em>P</em> &lt; 0.05). The analytic hierarchy process was then applied to calculate indicator weights, ensuring structured prioritization of quality dimensions (all judgment matrices yielded CR values below 0.1). The GAI-HIQ provides a consensus-based framework for evaluating health information generated by AI, offering practical tools for developers to optimize algorithms, for healthcare institutions to regulate applications, and for patients to assess information reliability.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104651"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146081584","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":"Fusion decomposition and backbone gathering based multimodal sentiment analysis under uncertain missing modalities","authors":"Hongxiang Sun ,&nbsp;Zhizhong Liu ,&nbsp;Dianhui Chu ,&nbsp;Quan Z. Sheng ,&nbsp;Zhaowei Liu ,&nbsp;Yingjie Wang ,&nbsp;Adnan Mahmood","doi":"10.1016/j.ipm.2026.104674","DOIUrl":"10.1016/j.ipm.2026.104674","url":null,"abstract":"<div><div>Recently, Multimodal Sentiment Analysis (MSA) under uncertain missing modalities has become a critical challenge in the field of emotional computing. Although some effective models have been developed, these models largely fail to consider the influence of modalities’ quality on the multimodal fusion and missing modalities’ completion. To tackle these issues, we propose a Fusion Decomposition and Backbone Gathering based MSA model under uncertain missing modalities, FDBG, which includes four innovative works. (1) Inspired by the pyramid convolution structure, we devise a pyramid multi-head attention mechanism, which serves as an important calculation foundation for the feature learning in FDBG; (2) To improve the quality of each modality, we suggest a pyramid prior enhancement method that combines the current sample with several similar historical samples; (3) We put forward a fusion decomposition method to effectively complete the missing modalities, which first fuses the three modalities with the pyramid self-attention, and then decomposes the fusion feature into three modalities and completes the missing modalities with the self-attention interaction mechanism; and (4) To eliminate the interference, we introduce a backbone gathering method to translate and fuse the three modalities into a single modality. Based on three public benchmark datasets (IEMOCAP, MELD and CMU-MOSI), we conduct extensive experiments and prove that FDBG outperforms the 12 baseline models. The source codes of this work are available at <span><span>https://github.com/SHX-AI/FDBGMSA</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104674"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190669","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":"Feature structure co-optimized augmented network for graph anomaly detection","authors":"Yingyue Zhang,&nbsp;Huifang Ma,&nbsp;Rui Bing,&nbsp;Meihuizi Jia,&nbsp;Xiaofei Wang","doi":"10.1016/j.ipm.2026.104661","DOIUrl":"10.1016/j.ipm.2026.104661","url":null,"abstract":"<div><div>Graph Anomaly Detection (GAD) aims to identify nodes that deviate from attribute or structural norms. Despite recent advances, redundant features and structural noise continue to obscure anomaly signals and bias node representations. We propose FSCAN (Feature-Structure Co-optimized Augmented Network), a plug-in module that jointly optimizes feature selection and graph refinement. FSCAN leverages spectral contribution-based selection, grounded in the Spiked Covariance Model and Singular Value Decomposition, to identify label-relevant features, while its adaptive refinement prunes noisy edges from head nodes and augments connections for tail nodes to enhance information propagation. Experiments on benchmark datasets show that FSCAN outperforms state-of-the-art baselines. Specifically, on the Questions dataset, when integrated with a supervised backbone, FSCAN achieves a GMean of 67.35%, representing a 6.63% improvement over the best baseline; when integrated with an unsupervised backbone, the GMean reaches 22.49%, corresponding to a 10.98% improvement over the baseline. Our code is publicly available at <span><span>https://github.com/mala-tang/FSCAN</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104661"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190594","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 named entity recognition method for Chinese aviation accident investigation reports based on prescriptive data augmentation and fine-weighted voting","authors":"Hui Jin ,&nbsp;Yinxiao Hu ,&nbsp;Jiayu Chen ,&nbsp;Hongjuan Ge ,&nbsp;Weiwei Liu ,&nbsp;Jiayi Zhao","doi":"10.1016/j.ipm.2026.104663","DOIUrl":"10.1016/j.ipm.2026.104663","url":null,"abstract":"<div><div>Extracting comprehensive information from aviation accident investigation (AAI) reports based on named entity recognition (NER) methods is beneficial for accident analysis and prevention. In this study, a novel NER method for identifying complex entities in Chinese AAI reports is proposed, which composed of prescriptive data augmentation (PDA), improved SoftLexicon based on part of speech (SLPS), and fine-weighted voting (FWV), named as PDA-SLPS-FWV briefly. Firstly, a PDA method is proposed to address the syntactic errors commonly occur in sentences generated by existing data augmentation methods. Subsequently, the SoftLexicon is improved by incorporating part of speech feature, and sub-models are trained using SLPS and the augmented datasets. Furthermore, the FWV method is proposed for ensemble learning, which assigns category-specific weights to each sub-model. Finally, a domain-specific AAI dataset is constructed based on 42 Chinese AAI reports, and a series of experiments are conducted on this dataset along with two widely-used benchmark datasets, Weibo and OntoNotes. The results indicate that our method observe significant improvements on the AAI dataset, achieving a 2.8 % increase in F1-score compared to SoftLexicon. Furthermore, on the two benchmark datasets, our method also exhibited competitive performance, outperforming SoftLexicon and performing comparably to recent state-of-the-art baselines.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"63 5","pages":"Article 104663"},"PeriodicalIF":6.9,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190572","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}