计算机科学最新文献

{"title":"A Topology-Aware GNN Learning Approach for Energy Optimization in Multi-Hop LoRa Networks","authors":"Huapeng Yang, Xiping Wu, Han Ji, Zhangqin Huang, Juan Fang","doi":"10.1109/jiot.2025.3606698","DOIUrl":"https://doi.org/10.1109/jiot.2025.3606698","url":null,"abstract":"","PeriodicalId":54347,"journal":{"name":"IEEE Internet of Things Journal","volume":"42 1","pages":""},"PeriodicalIF":10.6,"publicationDate":"2025-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145003143","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Rehabilitation exercise assessment using convolutional neural network with feature prioritization using Whale optimization and Shapely values","authors":"Md. Johir Raihan ,&nbsp;Mainul Islam Labib ,&nbsp;Md. Atiqur Rahman Ahad ,&nbsp;Abdullah-Al Nahid","doi":"10.1016/j.compeleceng.2025.110644","DOIUrl":"10.1016/j.compeleceng.2025.110644","url":null,"abstract":"<div><div>Assessing the rehabilitation score is one of the crucial areas of healthcare research which has the potential to automate the assessment process and to improve the patient’s exercise to their needs. In this paper, we have proposed a framework that uses a metaheuristic algorithm such as the Whale Optimization Algorithm (WOA) to find the most useful set of body joints that can be used to assess the exercise score efficiently. In addition, the metaheuristic algorithm was used to fine-tune the hyperparameters of the convolutional neural network architecture to best fit the data. Further, to understand the model’s inner working, SHapely Additive exPlanation (SHAP) was used to find how each body joint is influencing the model globally to predict the assessment score of an exercise. The proposed framework has been used on two popularly available datasets named KInematic Assessment of MOvement and Clinical Scores for Remote Monitoring of Physical REhabilitation (KIMORE) dataset and University of Idaho – Physical Rehabilitation Movements Data Set (UI-PRMD) to get a better understanding of the proposed methodology. In the KIMORE and UI-PRMD datasets, the WOA algorithm selected 47 out of 75 features and 63 out of 117 features respectively, which is almost half of the initial features. We have obtained a mean absolute deviation of 0.125 and 0.028 in the KIMORE and UI-PRMD datasets respectively. The SHAP analysis performed among the WOA selected features shows that the features of the hand of the KIMORE subject mostly influence the model, whereas the feature of the foot of the UI-PRMD subject is most influential in the model’s prediction.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110644"},"PeriodicalIF":4.9,"publicationDate":"2025-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144997047","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enhanced performance of intelligent hybrid Takagi Sugeno-Integral Backstepping controller for grid-tied PV systems: processor-in-the-loop validation","authors":"Oumaima Echab ,&nbsp;Noureddine Ech-Cherki ,&nbsp;Ilham Nassar-Eddine ,&nbsp;Elmostafa Chetouani ,&nbsp;Abdellatif Obbadi ,&nbsp;Youssef Errami ,&nbsp;Smail Sahnoun ,&nbsp;Mohssin Aoutoul","doi":"10.1016/j.compeleceng.2025.110690","DOIUrl":"10.1016/j.compeleceng.2025.110690","url":null,"abstract":"<div><div>This paper presents an intelligent hybrid Takagi-Sugeno Integral Backstepping Control (TS-IBSC) technique to enhance the performance of grid-tied photovoltaic systems (GTPVS) under rapidly changing environmental conditions. The suggested technique combines TS fuzzy logic with the nonlinear IBSC, which is confirmed through Processor-In-the-Loop (PIL) implementation using the eZdsp F28335 board. This hybrid method is assessed under multiple scenarios, including abrupt climatic changes, real-world irradiance variations, a grid fault, and three-phase load changes. Results corroborate the superior MPPT efficiency of 98.85 %, rapid settling time of 17 ms, and reduced Harmonic Distortion (THD) of 0.59 % under standard conditions. TS-IBSC establishes rapid convergence, enhanced power quality, and overall system stability, outperforming conventional methods. The findings prove its feasibility for real-time practical applications, offering a robust and reliable solution for GTPVS.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110690"},"PeriodicalIF":4.9,"publicationDate":"2025-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144997049","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Electroencephalography-based emotion recognition using a dual-stream multi-scale spatiotemporal convolutional capsule network","authors":"Han Cai,&nbsp;Pengfei Lu,&nbsp;Xiaofang Wang,&nbsp;Yuxing Wang","doi":"10.1016/j.engappai.2025.112144","DOIUrl":"10.1016/j.engappai.2025.112144","url":null,"abstract":"<div><div>In recent years, emotion recognition using electroencephalogram (EEG) signals has gained significant attention. However, shallow convolutional neural networks used in EEG emotion recognition are ineffective at capturing spatial relationships between features, adversely affecting model performance. To solve this problem, this study proposes a dual-stream multi-scale spatiotemporal convolutional capsule network aimed at improving EEG-based emotion recognition. The novelty of our approach is the dual-stream multi-scale spatiotemporal design, which enables parallel extraction and fusion of temporal and spatial EEG features at multiple scales, offering a richer and more discriminative representation for emotion recognition. Specifically, a dual-stream feature construction module is developed to extract multi-scale temporal and spatial features from raw EEG signals. A hybrid spatiotemporal attention mechanism enhances feature fusion, while a capsule-based classifier improves recognition accuracy by modeling relationships between local and global features. Experimental results using subject-dependent 10-fold cross-validation show average accuracies of 97.72%, 97.56%, and 97.82% for valence, arousal, and dominance on the Dataset for Emotion Analysis using Physiological signals(DEAP), with average F1 scores of 97.83%, 97.51%, and 97.68%. For the Dataset for Emotion Analysis using Physiological signals(DREAMER<span><span>¹</span></span>), the average accuracies are 96.48%, 96.32%, and 96.35%, with corresponding F1 scores of 96.23%, 95.74%, and 95.66%. The proposed method outperforms existing state-of-the-art approaches on both datasets, reducing the number of parameters by 58.07% and decreasing inference time by approximately 50.86% and 33.39%, compared to Residual Network 18. Additionally, in the subject-independent leave-one-subject-out cross-validation, the proposed method demonstrated results that significantly outperformed the baseline model across both datasets. Experiments show that the proposed method enhances spatiotemporal feature extraction and improves emotion recognition accuracy. This method reduces computational resource consumption and enhances recognition accuracy, thereby facilitating efficient algorithm development and deployment for real-time emotion monitoring and related applications.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"161 ","pages":"Article 112144"},"PeriodicalIF":8.0,"publicationDate":"2025-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144997368","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"SentireCache: Accelerate Sentiment Classification With Saliency-Based Caching","authors":"Yilong Zhu;Juncheng Jia;Mianxiong Dong;Jun Qi","doi":"10.1109/TAFFC.2025.3578574","DOIUrl":"10.1109/TAFFC.2025.3578574","url":null,"abstract":"Deep Learning methodologies have demonstrated exceptional efficacy in sentiment classification tasks. However, their extended inference times often impede practical deployment, particularly in resource-constrained environments. This paper addresses the challenge of reducing inference time by introducing a novel in-GPU caching approach, termed SentireCache, specifically designed for sentiment classification tasks. While traditional caching methods with the cosine similarity measurement have shown some reduction in inference time, they suffer from low hit rates and accuracy. To overcome this limitation, we incorporate a token filtering mechanism based on saliency into the caching system, along with simplified similarity calculation methods. The effectiveness of our proposed approach is theoretically analyzed. Moreover, extensive experimentation is conducted to compare SentireCache with other state-of-the-art caching methods. The results demonstrate a significant 37.7% reduction in inference time with an average performance degradation of 4.69%.","PeriodicalId":13131,"journal":{"name":"IEEE Transactions on Affective Computing","volume":"16 3","pages":"1349-1361"},"PeriodicalIF":9.8,"publicationDate":"2025-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145003108","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Spike-Driven Lightweight Large Language Model With Evolutionary Computation","authors":"Malu Zhang, Wenjie Wei, Zijian Zhou, Wanlong Liu, Jie Zhang, Ammar Belatreche, Yang Yang","doi":"10.1109/tevc.2025.3606613","DOIUrl":"https://doi.org/10.1109/tevc.2025.3606613","url":null,"abstract":"","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"64 1","pages":""},"PeriodicalIF":14.3,"publicationDate":"2025-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145002917","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"IEEE Access","authors":"","doi":"10.1109/mcom.2025.11152686","DOIUrl":"https://doi.org/10.1109/mcom.2025.11152686","url":null,"abstract":"","PeriodicalId":55030,"journal":{"name":"IEEE Communications Magazine","volume":"35 1","pages":""},"PeriodicalIF":11.2,"publicationDate":"2025-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145003032","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"TechRxiv","authors":"","doi":"10.1109/mcom.2025.11152694","DOIUrl":"https://doi.org/10.1109/mcom.2025.11152694","url":null,"abstract":"","PeriodicalId":55030,"journal":{"name":"IEEE Communications Magazine","volume":"64 1","pages":""},"PeriodicalIF":11.2,"publicationDate":"2025-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145003037","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Women in Engineering","authors":"","doi":"10.1109/mcom.2025.11152668","DOIUrl":"https://doi.org/10.1109/mcom.2025.11152668","url":null,"abstract":"","PeriodicalId":55030,"journal":{"name":"IEEE Communications Magazine","volume":"33 1","pages":""},"PeriodicalIF":11.2,"publicationDate":"2025-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145003039","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Look Back","authors":"","doi":"10.1109/mcom.2025.11152691","DOIUrl":"https://doi.org/10.1109/mcom.2025.11152691","url":null,"abstract":"","PeriodicalId":55030,"journal":{"name":"IEEE Communications Magazine","volume":"64 1","pages":""},"PeriodicalIF":11.2,"publicationDate":"2025-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145003046","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}