计算机科学最新文献_第6页

Managing real-time constraints through monitoring and analysis-driven edge orchestration

IF 3.7 2区计算机科学

Journal of Systems Architecture Pub Date : 2025-04-04 DOI: 10.1016/j.sysarc.2025.103403

Daniel Casini , Paolo Pazzaglia , Matthias Becker

{"title":"Managing real-time constraints through monitoring and analysis-driven edge orchestration","authors":"Daniel Casini , Paolo Pazzaglia , Matthias Becker","doi":"10.1016/j.sysarc.2025.103403","DOIUrl":"10.1016/j.sysarc.2025.103403","url":null,"abstract":"<div><div>Emerging real-time applications are increasingly moving to distributed heterogeneous platforms, under the promise of more powerful and flexible resource capabilities. This shift inevitably brings new challenges. The design space to deploy chains of threads is more complex, and sound estimates of worst-case execution times are harder to obtain. Additionally, the environment is more dynamic, requiring additional runtime flexibility on the part of the application itself. In this paper, we present an optimization-based approach to this problem. First, we present a model and real-time analysis for modern distributed edge applications. Second, we propose a design-time optimization problem to show how to set the main parameters characterizing such applications from a time-predictability perspective. Then, we present an orchestration and runtime decision-making mechanism that monitors execution times and allows for runtime reconfigurations, spanning from graceful degradation policies to re-distributions of workload. A prototypical implementation of the proposed approach based on the QNX RTOS and its evaluation on a realistic case study based on an edge-based valet parking application conclude the paper.</div></div>","PeriodicalId":50027,"journal":{"name":"Journal of Systems Architecture","volume":"163 ","pages":"Article 103403"},"PeriodicalIF":3.7,"publicationDate":"2025-04-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143776864","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An incorporation of metaheuristic algorithm and two-stage deep learnings for fault classified framework for diesel generator maintenance

IF 7.5 2区计算机科学

Engineering Applications of Artificial Intelligence Pub Date : 2025-04-04 DOI: 10.1016/j.engappai.2025.110688

Thanh-Phuong Nguyen

{"title":"An incorporation of metaheuristic algorithm and two-stage deep learnings for fault classified framework for diesel generator maintenance","authors":"Thanh-Phuong Nguyen","doi":"10.1016/j.engappai.2025.110688","DOIUrl":"10.1016/j.engappai.2025.110688","url":null,"abstract":"<div><div>Diesel generators play a vital role in providing reliable power to ensure uninterrupted power supply. However, effective fault classification in these systems is challenging due to their complexity. This paper proposes a two-stage framework that combines deep learning with metaheuristic optimization for fault classification of diesel generators in Artificial Internet of Things (AIoT) systems. The first stage involves employing a Long short-term memory convolutional neural network (LSTM-CNN) model for accurate feature extraction and fault detection. The improved particle swarm optimization (IPSO) algorithm is employed to optimize the hyperparameters of the LSTM-CNN model, resulting in an enhanced IPSO-LSTM-CNN framework. A comprehensive performance evaluation is conducted by comparing the developed algorithm with other models, including recurrent neural networks (RNN), CNN, gated recurrent units (GRU), LSTM, and CNN-LSTM. The IPSO-LSTM-CNN obtains the most significant gains of many evaluation benchmarks when compared to other state-of-the-art algorithms. In terms of fault classification accuracy and robustness, the created model performs better than the alternative methods, confirming its usefulness in enhancing the operational efficiency and dependability of diesel generators in AIoT frameworks. This research provides a completed IPSO-LSTM-CNN framework in AI application for failure diagnosis of industrial machinery in maintenance service.</div></div>","PeriodicalId":50523,"journal":{"name":"Engineering Applications of Artificial Intelligence","volume":"151 ","pages":"Article 110688"},"PeriodicalIF":7.5,"publicationDate":"2025-04-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143769271","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}

引用次数: 0

A Dynamic Burst Assembly Approach for High-Priority and Self-Similar Traffic in High-Speed Optical Burst Switched Network

IF 1.7 4区计算机科学

International Journal of Communication Systems Pub Date : 2025-04-04 DOI: 10.1002/dac.70079

Shamandeep Singh, Simranjit Singh, Bikrampal Kaur, Prabhjot Kaur Chahal

{"title":"A Dynamic Burst Assembly Approach for High-Priority and Self-Similar Traffic in High-Speed Optical Burst Switched Network","authors":"Shamandeep Singh, Simranjit Singh, Bikrampal Kaur, Prabhjot Kaur Chahal","doi":"10.1002/dac.70079","DOIUrl":"https://doi.org/10.1002/dac.70079","url":null,"abstract":"<div>\u0000 \u0000 <p>Dynamic burst size selection is a challenging process in the optical burst switching (OBS) networks for efficient burst assembly. In this manuscript, a dynamic burst-size assembly approach is proposed to standardize the data burst size in OBS networks. The proposed approach utilizes hysteresis properties in the burst size decider module (BSDM) to decide the data burst size. The inculcation of the dynamic burst assembly algorithm (DBAA) focuses on the nonlinear features to handle the blocking problem during the burst assembly process. DBAA involves a priority evaluator mechanism to determine the importance of each incoming packet at the ingress node. This provides a dynamic decision-making strategy to standardize the data burst size with change in transition count number (TCN). The performance of the proposed approach is evaluated on the self-similar traffic model with burstiness, ranging from <i>H</i> = 0.5–0.7. The experimental results show a decrease in the average queuing delay by 14.59% and an improved average burst utilization by 23.36% compared with the hybrid (time/length) approach. However, the proposed DBAA attains better burst utilization with a significant reduction in queuing delay. Furthermore, the consistency value of burst sizes indicates that DBAA performs better in terms of burst utilization than existing approaches.</p>\u0000 </div>","PeriodicalId":13946,"journal":{"name":"International Journal of Communication Systems","volume":"38 7","pages":""},"PeriodicalIF":1.7,"publicationDate":"2025-04-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143769989","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Causal discovery based on hierarchical reinforcement learning

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-04-04 DOI: 10.1016/j.eswa.2025.127466

Jingchi Jiang , Rujia Shen , Chao Zhao , Yi Guan , Xuehui Yu , Xuelian Fu

{"title":"Causal discovery based on hierarchical reinforcement learning","authors":"Jingchi Jiang , Rujia Shen , Chao Zhao , Yi Guan , Xuehui Yu , Xuelian Fu","doi":"10.1016/j.eswa.2025.127466","DOIUrl":"10.1016/j.eswa.2025.127466","url":null,"abstract":"<div><div>Conditional independence (CI) tests in causal discovery can determine a set of Markov equivalence classes w.r.t. the observed data by checking whether each pair of variables is d-separated under faithfulness and Markov assumptions. However, CI tests are intractable for high-dimensional conditional variables. Motivated by the advantages of reinforcement learning in exploring the solution space, firstly, we propose a causal discovery framework based on hierarchical reinforcement learning (CD-HRL). This framework trains both the discovery of the causal skeleton and the identification of direction using two interdependent high-level and low-level policies seperately. Dividing causal discovery into two distinct subtasks to high-level and low-level policies enhances exploration efficiency and minimises error accumulation. The high-level policy iteratively generates causal skeletons as subgoals for instructing the low-level policy, which then identifies causal directions of individual pairs of variables. Secondly, to avoid redundant exploration of familiar causal structures, we incorporate a memory module into the high-level agent and predefine an augmented reward that combines a causal score function and a curiosity item for exploring unknown causal structures. Lastly, experiments on both synthetic and real datasets show that the proposed approach outperforms the state-of-the-art methods under various data-generating procedures, which follow linear, nonlinear, and ordinary differential equations with additive Gaussian noise. The code for our CD-HRL method is available online in <span><span>https://github.com/HITshenrj/CD-HRL</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"279 ","pages":"Article 127466"},"PeriodicalIF":7.5,"publicationDate":"2025-04-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143777701","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Combining N-grams and graph convolution for text classification

IF 7.2 1区计算机科学

Applied Soft Computing Pub Date : 2025-04-04 DOI: 10.1016/j.asoc.2025.113092

Tarık Üveys Şen , Mehmet Can Yakit , Mehmet Semih Gümüş , Orhan Abar , Gokhan Bakal

{"title":"Combining N-grams and graph convolution for text classification","authors":"Tarık Üveys Şen , Mehmet Can Yakit , Mehmet Semih Gümüş , Orhan Abar , Gokhan Bakal","doi":"10.1016/j.asoc.2025.113092","DOIUrl":"10.1016/j.asoc.2025.113092","url":null,"abstract":"<div><div>Text classification, a cornerstone of natural language processing (NLP), finds applications in diverse areas, from sentiment analysis to topic categorization. While deep learning models have recently dominated the field, traditional n-gram-driven approaches often struggle to achieve comparable performance, particularly on large datasets. This gap largely stems from deep learning’ s superior ability to capture contextual information through word embeddings. This paper explores a novel approach to leverage the often-overlooked power of n-gram features for enriching word representations and boosting text classification accuracy. We propose a method that transforms textual data into graph structures, utilizing discriminative n-gram series to establish long-range relationships between words. By training a graph convolution network on these graphs, we derive contextually enhanced word embeddings that encapsulate dependencies extending beyond local contexts. Our experiments demonstrate that integrating these enriched embeddings into an long-short term memory (LSTM) model for text classification leads to around 2% improvements in classification performance across diverse datasets. This achievement highlights the synergy of combining traditional n-gram features with graph-based deep learning techniques for building more powerful text classifiers.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"175 ","pages":"Article 113092"},"PeriodicalIF":7.2,"publicationDate":"2025-04-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143776892","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}

引用次数: 0

IEEE Transactions on Neural Networks and Learning Systems Publication Information

IF 10.2 1区计算机科学

IEEE transactions on neural networks and learning systems Pub Date : 2025-04-04 DOI: 10.1109/TNNLS.2025.3550761

引用次数: 0

Directional sound transmission and reception of the beluga whale (Delphinapterus leucas). 白鲸（Delphinapterus leucas）的定向声音传播和接收。

IF 3.1 3区计算机科学

Bioinspiration & Biomimetics Pub Date : 2025-04-04 DOI: 10.1088/1748-3190/adc5bd

Wenzhan Ou, Zhongchang Song, Caroline E C Goertz, T Aran Mooney, Sophie Dennison, Chuang Zhang, Yu Zhang, Manuel Castellote

{"title":"Directional sound transmission and reception of the beluga whale (<i>Delphinapterus leucas</i>).","authors":"Wenzhan Ou, Zhongchang Song, Caroline E C Goertz, T Aran Mooney, Sophie Dennison, Chuang Zhang, Yu Zhang, Manuel Castellote","doi":"10.1088/1748-3190/adc5bd","DOIUrl":"10.1088/1748-3190/adc5bd","url":null,"abstract":"<p><p>The biosonar system of odontocetes enables directional sound transmission and reception. Beluga whales (<i>Delphinapterus leucas</i>) are notable among odontocetes as they can alter the shape of their fatty melon during sound transmission, potentially suggesting distinct acoustic capabilities. In this study, we developed a biosonar model of a beluga whale using computed tomography scanning and structural reconstruction to examine directional transmission and reception in this species. This model could modulate sounds into a directional beam using either single or dual sources. Across frequencies from 5 to 60 kHz, the directivity indices for the left and right sound sources ranged from 4.83 to 15.2 dB and 4.81-14.7 dB, respectively. When both sound sources were used simultaneously, there was an average increase of at least 2.26 dB in energy and 0.68 dB in the directivity index compared to using a single source. Additionally, beam steering was achieved in the dual-source transmission by introducing a timing difference between the two sources. The simulations indicated that sound reception was frequency-dependent, with the greatest sensitivity to lateral sounds at low frequencies and to forward sounds at high frequencies. These results suggested that both transmission and reception in beluga whales were directional and frequency-dependent.</p>","PeriodicalId":55377,"journal":{"name":"Bioinspiration & Biomimetics","volume":" ","pages":""},"PeriodicalIF":3.1,"publicationDate":"2025-04-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143733382","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}

引用次数: 0

Towards Real-World Aerial Vision Guidance with Categorical 6D Pose Tracker

IF 23.6 1区计算机科学

IEEE Transactions on Pattern Analysis and Machine Intelligence Pub Date : 2025-04-04 DOI: 10.1109/tpami.2025.3558237

Jingtao Sun, Yaonan Wang, Danwei Wang

引用次数: 0

A Survey of Deep Transfer Learning in Automatic Modulation Classification

IF 8.6 1区计算机科学

IEEE Transactions on Cognitive Communications and Networking Pub Date : 2025-04-04 DOI: 10.1109/tccn.2025.3558027

Xiang Wang, Yurui Zhao, Zhitao Huang

引用次数: 0

Rapid Optimization of Superposition Codes for Multi-Hop NOMA MANETs via Deep Unfolding