Expert Systems with Applications最新文献_第8页

DTS: A Decoupled Task Specificity Approach for Aspect Sentiment Triplet Extraction

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-02-10 DOI: 10.1016/j.eswa.2025.126759

Bao Wang, Guangjin Wang, Peiyu Liu

{"title":"DTS: A Decoupled Task Specificity Approach for Aspect Sentiment Triplet Extraction","authors":"Bao Wang, Guangjin Wang, Peiyu Liu","doi":"10.1016/j.eswa.2025.126759","DOIUrl":"10.1016/j.eswa.2025.126759","url":null,"abstract":"<div><div>Aspect Sentiment Triplet Extraction (ASTE) aims to extract aspect terms, opinion terms, and their corresponding sentiment polarities from customer reviews. Contemporary table-filling approaches prominently construct a task-sharing table for distinct subtasks to exploit the interaction between entities. However, the single table mechanism neglects the need for task-specific knowledge, inevitably causing feature confusion. In this paper, we introduce an innovative method named Decoupled Task Specificity (DTS) to address these issues. Specifically, this model builds a term expert table to learn semantically and syntactically enhanced knowledge for term extraction while constructing another sentiment expert table for sentiment classification by incorporating more comprehensive contextual knowledge. These two task expert tables learn task-specific knowledge to mitigate the gap between knowledge and subtasks. Moreover, two task-specific decoders, based on different region detection strategies, are designed for the decoding of expert tables. Overall, these modules jointly achieve subtask specificity throughout the whole process. Experimental results on public benchmark datasets show the effectiveness and exceptional performance of the DTS.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"273 ","pages":"Article 126759"},"PeriodicalIF":7.5,"publicationDate":"2025-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143420464","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

Effective state monitoring for remote auto-operation systems

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-02-10 DOI: 10.1016/j.eswa.2025.126774

Hongyang Xu , Yiming Ding , Lilan Tu

{"title":"Effective state monitoring for remote auto-operation systems","authors":"Hongyang Xu , Yiming Ding , Lilan Tu","doi":"10.1016/j.eswa.2025.126774","DOIUrl":"10.1016/j.eswa.2025.126774","url":null,"abstract":"<div><div>The study of state monitoring for remote auto-operation systems (RAOSs) is an interesting and realistic issue. From the state monitoring perspective, chaos dynamics is employed to simulate the signal communication of RAOS, and a novel multi-layer network structure is proposed for the system. The system is designed into multiple layers that independently manage different signal types, which can determine the overall network state by monitoring a high-layer node, and fast backtracking can locate the fault location. Meanwhile, a divide-and-conquer management strategy is proposed, constructing a low-layer network composed of small sub-networks that are non-fully connected and decentralized. They can optimize specific tasks independently, enabling secret and fast parallel signal transmission and fault localization. A simple and sensitive non-stationary measure (NS) is used to monitor the RAOS, which accurately monitors the fault location within milliseconds after a transmission. Results demonstrate a 100% fault monitoring rate under noiseless, and the robustness of the proposed method is validated with the signal-to-noise ratio (SNR) of 15. The average fault monitoring accuracy in the low-layer network exceeds 97%, while the middle-layer network can achieve more than 85%. These experiments indicate that our method has significant application potential for state monitoring in RAOSs.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"273 ","pages":"Article 126774"},"PeriodicalIF":7.5,"publicationDate":"2025-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143420566","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

GraphDRL: GNN-based deep reinforcement learning for interactive recommendation with sparse data

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-02-10 DOI: 10.1016/j.eswa.2025.126832

Wenxin Li, Xiao Song, Yuchun Tu

{"title":"GraphDRL: GNN-based deep reinforcement learning for interactive recommendation with sparse data","authors":"Wenxin Li, Xiao Song, Yuchun Tu","doi":"10.1016/j.eswa.2025.126832","DOIUrl":"10.1016/j.eswa.2025.126832","url":null,"abstract":"<div><div>Interactive recommendation (IR) continuously optimizes performance through sustained interactions between users and the system, thereby capturing dynamic changes in user interests more effectively. Due to the advantages of deep reinforcement learning (DRL) in dynamic optimization and decision-making, researchers have integrated DRL models into interactive recommendations. However, the interactive recommendation still faces the problem of data sparsity, and DRL-based recommendation algorithms often suffer from efficiency issues when handling large-scale discrete action spaces. To address these problems, this paper proposes a GNN-based deep reinforcement learning model, GraphDRL. Specifically, we utilize Graph Neural Networks (GNNs) to obtain embedding representations that effectively model the intricate interactions between users and items, alleviating the data sparsity problem. On this basis, we construct a deep reinforcement learning model with a temporal multi-head attention method to capture users’ evolving preferences. Moreover, we propose a dynamic candidate action generation method based on item popularity and embedding representations, which not only more accurately identifies items of interest to users but also reduces the action space, thereby improving recommendation accuracy and efficiency. The superior performance of our algorithm is confirmed through experiments on three public benchmark recommendation datasets and a real-world buyer–supplier interaction dataset.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"273 ","pages":"Article 126832"},"PeriodicalIF":7.5,"publicationDate":"2025-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143420567","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

Enhancing bone-conducted speech through a pre-trained transformer with low-rank driven sparsity bootstrapping

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-02-10 DOI: 10.1016/j.eswa.2025.126761

Dongjing Shan , Mengchu Yang , Jiashun Mao , Yamei Luo , Qi Han

{"title":"Enhancing bone-conducted speech through a pre-trained transformer with low-rank driven sparsity bootstrapping","authors":"Dongjing Shan , Mengchu Yang , Jiashun Mao , Yamei Luo , Qi Han","doi":"10.1016/j.eswa.2025.126761","DOIUrl":"10.1016/j.eswa.2025.126761","url":null,"abstract":"<div><div>The traditional Transformer architecture encounters substantial challenges in terms of time complexity when dealing with long sequences. Sequential signals, such as speech and serialized image data, inherently exhibit low-rank properties along the temporal axis. By leveraging this low-rank nature effectively, we can not only prune redundant information to enhance model robustness but also devise a sparsity-bootstrapped attention mechanism that significantly reduces the temporal complexity of Transformer-based models. This study is dedicated to applying a self-supervised, pre-trained model that leverages low-rank driven sparsity bootstrapping to enhance bone-conducted speech and address the challenge of scarce paired speech data. This innovative technique enables communication in noisy environments by directly capturing signals from the human skull or larynx. In our experiments, we benchmark our model against five other state-of-the-art recovery models using a comprehensive set of evaluation criteria. Both objective metrics and subjective assessments consistently demonstrate the superiority of our proposed model, indicating its potential to advance bone-conducted speech enhancement technologies.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"272 ","pages":"Article 126761"},"PeriodicalIF":7.5,"publicationDate":"2025-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143387323","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

Online semantic embedding correlation for discrete cross-media hashing 用于离散跨媒体散列的在线语义嵌入相关性

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-02-10 DOI: 10.1016/j.eswa.2025.126758

Fan Yang, Haoyu Hu , Fumin Ma , Xiaojian Ding, Qiaoxi Zhang, Xinqi Liu

{"title":"Online semantic embedding correlation for discrete cross-media hashing","authors":"Fan Yang, Haoyu Hu , Fumin Ma , Xiaojian Ding, Qiaoxi Zhang, Xinqi Liu","doi":"10.1016/j.eswa.2025.126758","DOIUrl":"10.1016/j.eswa.2025.126758","url":null,"abstract":"<div><div>The rapid expansion of multimedia data has generated an urgent need for efficient retrieval methods. While batch-based cross-modal hashing has advanced precision in retrieval, it becomes inefficient for online streaming data, increasing computation and storage costs. Additionally, existing online methods often overlook the interdependencies among multiple labels in multimodal data, limiting their ability to generate highly discriminative hash codes. To address these issues, we propose a new online hashing method known as Online semantiC Embedding correlAtion for discrete cross-media hashiNg (OCEAN). OCEAN directly extracts key feature information from multimodal data and uses a normalized label inner product to connect the supervised information accumulated over all rounds, embedding rich semantics into hash codes while reducing computational and storage needs. An asymmetric strategy is introduced to enhance class information embedding, circumventing optimization issues from discrete constraints. Furthermore, OCEAN employs an adaptive label association strategy to dynamically learn label correlations, strengthening the semantic depth of supervised information. An online discrete iterative optimization strategy also helps create concise hash codes with improved discriminative power. Experiments on three benchmark databases show that OCEAN outperforms previous methods, offering superior scalability, efficiency, and search performance. Codes are available at <span><span>https://github.com/nufehash/OCEAN</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"272 ","pages":"Article 126758"},"PeriodicalIF":7.5,"publicationDate":"2025-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143387322","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

Two-level attention mechanism with contrastive learning for heterogeneous graph representation learning

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-02-10 DOI: 10.1016/j.eswa.2025.126751

Mahnaz Moradi , Parham Moradi , Azadeh Faroughi , Mahdi Jalili

{"title":"Two-level attention mechanism with contrastive learning for heterogeneous graph representation learning","authors":"Mahnaz Moradi , Parham Moradi , Azadeh Faroughi , Mahdi Jalili","doi":"10.1016/j.eswa.2025.126751","DOIUrl":"10.1016/j.eswa.2025.126751","url":null,"abstract":"<div><div>This paper introduces a novel method, M2CHGNN, for generating node representations in heterogeneous graphs by combining an attention mechanism with contrastive learning. The approach leverages meta-structures and meta-paths to capture complex hidden structures within heterogeneous graphs. While meta-paths identify diverse interaction patterns between nodes, meta-structures uncover intricate structural arrangements, revealing detailed relationships and capturing both local and higher-order structures simultaneously. Each meta-path or meta-structure results in a homogeneous graph to obtain node representations. Within each homogeneous subgraph, two views of node representation are then employed. In the first view, a node attention mechanism assesses the influence of neighboring nodes during embedding extraction, emphasizing the features of influential neighbors. Concurrently, a second set of embeddings is derived using the graph-topology view, highlighting structural relationships and further enriching the representations. Then, an additional attention layer is applied to determine the significance of each homogeneous subgraph within each view, resulting in a weighted, aggregated node representation. To learn robust and informative node representations across these two views, we use contrastive learning to align and distinguish representations between the node-attention and graph-topology views, alongside intra-view contrastive learning to refine each view individually. To train the node representations, we combine contrastive loss with a cross-entropy loss function, which enhances the model’s ability to generate high-quality node representations for heterogeneous graphs. The effectiveness of M2CHGNN was evaluated in three different applications, including link prediction, data classification, and clustering. The experimental results demonstrated the superiority of the proposed method in comparison with baseline and state-of-the-art graph embedding methods.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"273 ","pages":"Article 126751"},"PeriodicalIF":7.5,"publicationDate":"2025-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143428995","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

Unsupervised extractive opinion summarization based on text simplification and sentiment guidance

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-02-10 DOI: 10.1016/j.eswa.2025.126760

Rui Wang , Tian Lan , Zufeng Wu , Leyuan Liu

引用次数: 0

Urban traffic monitoring based on deep learning on an embedded GPU

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-02-10 DOI: 10.1016/j.eswa.2025.126847

Fredy Nocua M , Wilson-Javier Pérez-Holguín , Camilo Pardo-Beainy

{"title":"Urban traffic monitoring based on deep learning on an embedded GPU","authors":"Fredy Nocua M , Wilson-Javier Pérez-Holguín , Camilo Pardo-Beainy","doi":"10.1016/j.eswa.2025.126847","DOIUrl":"10.1016/j.eswa.2025.126847","url":null,"abstract":"<div><div>This paper presents a deep learning-based system for urban traffic monitoring, focusing on the detection and tracking of motorcycles using embedded hardware, due to the high accident rates of this type of vehicle. Different convolutional neural network (CNN) models were evaluated, including MobileNet-v1-SSD, YOLOv5, and Faster R-CNN, implemented on an NVIDIA Graphics Processing Units (GPUs) board as the Jetson Xavier NX®. The MobileNet-v1-SSD model stands out for its balance between precision (90 %), recall (66 %), and latency (∼10 ms), making it ideal for real-time applications. Additionally, a tracking algorithm based on optical flow using the Lucas-Kanade method was developed, complemented with logic for creating and deleting identities (IDs), enabling object tracking in dynamic scenarios with partial occlusions. The system includes a methodology for calculating key traffic variables such as speed and direction by correlating pixels with real-world distances through camera calibration. This approach demonstrates the feasibility of developing complex image-processing applications based on resource-constrained platforms by leveraging the features of efficient embedded systems such as General Purpose GPUs.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"273 ","pages":"Article 126847"},"PeriodicalIF":7.5,"publicationDate":"2025-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143420468","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

Identifying optimal technique of reducing dimensionality of scour influencing hydraulic parameters applying SWOT analysis

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-02-09 DOI: 10.1016/j.eswa.2025.126829

Sudarshan Mondal, Buddhadev Nandi, Subhasish Das

{"title":"Identifying optimal technique of reducing dimensionality of scour influencing hydraulic parameters applying SWOT analysis","authors":"Sudarshan Mondal, Buddhadev Nandi, Subhasish Das","doi":"10.1016/j.eswa.2025.126829","DOIUrl":"10.1016/j.eswa.2025.126829","url":null,"abstract":"<div><div>Spur dike-like structures (SDLS) are built in rivers to control erosion and reroute water flow. Predicting scour depth around SDLS is essential for structural stability, environmental protection, and public safety, depending on riverbed conditions, geometry, sediment properties, flow velocity, and many other parameters. Feature reduction simplifies modeling, improves computational efficiency and focuses on key variables influencing scour depth. To enhance feature reduction, this paper presents a systematically validated framework using dimensional analysis and 13 state-of-the-art dimensionality reduction (DR) techniques. First, dimensional analysis identifies key parameters, followed by using various DR techniques to reduce features. By methodical testing and assessment, the best DR strategies are determined. A SWOT analysis is performed to evaluate the strengths, weaknesses, opportunities, and threats associated with each DR technique. According to SWOT analysis, subset regression analysis (SSR) emerges as the optimal DR technique compared to others, both quantitatively and qualitatively. According to quantitative evaluation, SSR outperforms Principal component analysis (PCA) (61.2 %) and Discriminant analysis (DA) (18.33 %), achieving the highest weighted mean percentage (70 %) among all. Furthermore, SSR excels by its ability to reduce computational complexity while maintaining high predictive accuracy. The efficiency and robustness of SSR are further supported by its low standard error of 0.21 and optimal selection of predictors, ensuring reduced computational demands without compromising accuracy. This strategic assessment provides in-depth insights into the technical effectiveness and practical viability of DR techniques. Through dimensional analysis and evaluating DR techniques, this study advances feature reduction methods for improved decision-making in fields like scouring around SDLS.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"272 ","pages":"Article 126829"},"PeriodicalIF":7.5,"publicationDate":"2025-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143394524","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

A comprehensive ATM security framework for detecting abnormal human activity via granger causality-inspired graph neural network optimized with eagle-strategy supply-demand optimization

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-02-09 DOI: 10.1016/j.eswa.2025.126731

Aniruddha Prakash Kshirsagar , H. Azath

{"title":"A comprehensive ATM security framework for detecting abnormal human activity via granger causality-inspired graph neural network optimized with eagle-strategy supply-demand optimization","authors":"Aniruddha Prakash Kshirsagar , H. Azath","doi":"10.1016/j.eswa.2025.126731","DOIUrl":"10.1016/j.eswa.2025.126731","url":null,"abstract":"<div><div>The daily increase of criminal activity has made real-time human activity detection crucial for the protection & surveillance of public spaces, including bank-automated teller machines (ATM). To overcome the difficulty of online identification of anomalous activity in bank automated teller machines. A Comprehensive ATM Security Framework for Detecting Abnormal Human Activity via Granger Causality-Inspired Graph Neural Network optimized with Eagle-Strategy Supply-Dem& Optimization (ATM-DAHA-GCIGNN-ESSDO) is proposed in this manuscript. Initially, the input videos are gathered from DCSASS Dataset & UCF Crime Dataset. Then, the video is pre-processed by using Reverse Lognormal Kalman Filtering (RLKF) for cleaning noisy data. Granger Causality-Inspired Graph Neural Network (GCIGNN) is employed for detect abnormal human activities in ATM machine. Abuse, Arrest, Arson, Assault, Burglary, Explosion, Fighting, Road Accidents, Robbery, Shooting, Shoplifting, Stealing, V&alism for DCSASS Dataset & Abuse, Arrest, Assault, Arson, Burglary, Explosion, Fighting, Normal Videos, Road Accidents, Shoplifting, Shooting, Robbery, Stealing, V&alism for UCF Crime Dataset. The Eagle-Strategy Supply-Dem& Optimization (ESSDO) is implemented to enhance the parameters of GCIGNN. The proposed method is implemented & the efficiency is estimated using some performance metrics, like Accuracy, Recall, F1-score, precision, False Discovery Rate & Computational time. The performance of the ATM-DAHA-GCIGNN-ESSDO approach attains 24.39%, 35.71%, & 25.55% higher Accuracy; 22.15%, 24.21%, & 43.52% higher Recall. The proposed ATM-DAHA-GCIGNN-ESSDO framework outperforms the existing approaches for identifying aberrant human activity in ATM & criminal situations. Finally, the proposed approach demonstrates its potential as a reliable solution for real-time security & surveillance applications.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"272 ","pages":"Article 126731"},"PeriodicalIF":7.5,"publicationDate":"2025-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143372144","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