Expert Systems with Applications最新文献

筛选
英文 中文
Global optimization of interception guidance law for maneuvering target based on reward reconstruction
IF 7.5 1区 计算机科学
Expert Systems with Applications Pub Date : 2025-04-03 DOI: 10.1016/j.eswa.2025.127372
Wang Zhao , Qinglong Zhang , Ye Zhang , Jingyu Wang
{"title":"Global optimization of interception guidance law for maneuvering target based on reward reconstruction","authors":"Wang Zhao ,&nbsp;Qinglong Zhang ,&nbsp;Ye Zhang ,&nbsp;Jingyu Wang","doi":"10.1016/j.eswa.2025.127372","DOIUrl":"10.1016/j.eswa.2025.127372","url":null,"abstract":"<div><div>To address the maneuvering target interception task under finite field-of-view (FOV), this paper proposes a novel Reward-Guided Efficient Global Policy Learning (RGEGPL) method based on dynamic reward restructuring. The method aims to ensure efficient deep reinforcement learning (DRL) training while achieving a globally optimal policy with high interception accuracy, low average energy consumption, and low total energy consumption. To enhance the efficiency of the DRL process, the paper introduces an action space design based on proportional navigation (PN) to prevent the agent from conducting entirely random exploration during the initial phase in an unknown environment. Additionally, a reward shaping module is employed, along with a rational parameter selection method. To address the spatiotemporal reward coupling problem caused by the introduction of process rewards in the reward shaping module, this paper proposes a Spatiotemporal Coupling Decoupling (SCD) module based on dynamic reward reconstruction. This module effectively resolves the spatiotemporal coupling issue, ensuring the efficiency of the learning process while allowing iterative policy optimization to converge to the globally optimal solution. Through comparative simulations and Monte Carlo (MC) experiments, the results demonstrate that the proposed method achieves a policy with more than three times the interception accuracy of classical methods, and both average and total energy consumption are optimized. Compared to state-of-the-art DRL methods, the interception accuracy improves by 11.91%, and average and total energy consumption increase by over 6.44%. Furthermore, under conditions involving target maneuvering changes and constraints on decision frequency, the proposed method still exhibits strong robustness and strategic advantages. The computational complexity of the trained policy during the stage of use is also more efficient compared to other methods. The proposed method demonstrates superior performance and adaptability across multiple DRL baseline algorithms. Even in resource-constrained environments with limited decision-making frequency, the trained model maintains an advantage over both classical methods and state-of-the-art DRL-based methods.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"279 ","pages":"Article 127372"},"PeriodicalIF":7.5,"publicationDate":"2025-04-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143769135","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
GLMKD: Joint global and local mutual knowledge distillation for weakly supervised lesion segmentation in histopathology images
IF 7.5 1区 计算机科学
Expert Systems with Applications Pub Date : 2025-04-03 DOI: 10.1016/j.eswa.2025.127425
Hangbei Cheng , Xueyu Liu , Jun Zhang , Xiaorong Dong , Xuetao Ma , Yansong Zhang , Hao Meng , Xing Chen , Guanghui Yue , Yidi Li , Yongfei Wu
{"title":"GLMKD: Joint global and local mutual knowledge distillation for weakly supervised lesion segmentation in histopathology images","authors":"Hangbei Cheng ,&nbsp;Xueyu Liu ,&nbsp;Jun Zhang ,&nbsp;Xiaorong Dong ,&nbsp;Xuetao Ma ,&nbsp;Yansong Zhang ,&nbsp;Hao Meng ,&nbsp;Xing Chen ,&nbsp;Guanghui Yue ,&nbsp;Yidi Li ,&nbsp;Yongfei Wu","doi":"10.1016/j.eswa.2025.127425","DOIUrl":"10.1016/j.eswa.2025.127425","url":null,"abstract":"<div><div>Segmenting lesion from histopathology images plays a pivotal role in digital pathology workflow. Generally, fully-supervised segmentation algorithms for such task require pixel-wise manual annotations, which is extremely time-consuming and labor-tedious for the minute and scattered lesion in the high resolution whole-slide histopathology images (WSI). The development of weakly supervised methods, e.g. multiple instance learning (MIL), liberate pathologists and open up possibilities of further automated quantitative analysis of WSI. The challenge of using image-level annotations for lesion segmentation is the lack of adequate location information, causing severe lesion omission and segmentation inaccuracy. In this study, we propose an novel weakly supervised segmentation framework that combines global and local mutual knowledge distillation for lesion segmentation (named GLMKD). Specifically, GLMKD consists of extraction of both global and local lesion information through multi-scale views to yield multiple high-confidence pseudo masks, and subsequently employs dual-stream mutual knowledge distillation (MKD) technology with information interaction and fusion to guide the student model to achieve accurate segmentation of target lesion. In the MKD, a shape transfer loss is designed to mutually share knowledge between the global and local teachers, which allows the whole system to fully utilize the underlying information from image-level labels and enables the student model to converge to a better optimum. Experimental results demonstrate that our method achieves the good generality of lesion segmentation on glomerular lesion, Camelyon16, and DigestPath2019 datasets with mIoU scores of 84.65%, 82.93%, and 77.39%, respectively, outperforming other image-level weakly supervised methods, and achieving comparable performance to fully-supervised one.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"279 ","pages":"Article 127425"},"PeriodicalIF":7.5,"publicationDate":"2025-04-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143760700","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
Early fault diagnosis of transformers based on relative deterioration analysis
IF 7.5 1区 计算机科学
Expert Systems with Applications Pub Date : 2025-04-03 DOI: 10.1016/j.eswa.2025.127437
Yang Liu, Yingchun Wang, Zhishuo Wang, Suzhan Xue, Yudong Wang
{"title":"Early fault diagnosis of transformers based on relative deterioration analysis","authors":"Yang Liu,&nbsp;Yingchun Wang,&nbsp;Zhishuo Wang,&nbsp;Suzhan Xue,&nbsp;Yudong Wang","doi":"10.1016/j.eswa.2025.127437","DOIUrl":"10.1016/j.eswa.2025.127437","url":null,"abstract":"<div><div>This paper aims to enhance the fault diagnosis accuracy of oil-immersed power transformers by proposing a fault prediction framework based on CECNN-Bi-LSTM and an improved Decision Tree model. The framework introduces a Channel Equalization Module (CE-Block) within the CNN to activate suppressed channels, significantly boosting prediction accuracy. Additionally, the Decision Tree model is improved using the trace distance function to strengthen classification performance. The approach involves decomposing dissolved gas concentration data using Variational Mode Decomposition (VMD) to extract key features, then applying the CECNN-Bi-LSTM model to analyze the time-series data of dissolved gases for high-precision combined predictions. A Generative Adversarial Network (GAN) is used to balance fault classification data, enhancing robustness. Finally, the Entropy Weight-based Relative Degradation Degree (EM-RDA) evaluation standard increases classification feature dimensions and the improved Decision Tree accurately classifies transformer operating states. Simulation and case study results show that the proposed method achieves over 97% diagnostic accuracy, effectively predicting and preventing potential faults in transformers, thereby ensuring the stable operation of the power system.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"279 ","pages":"Article 127437"},"PeriodicalIF":7.5,"publicationDate":"2025-04-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143769129","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
Advancing recommendation systems with DeepMF and hybrid sentiment analysis: Deep learning and Lexicon-based integration
IF 7.5 1区 计算机科学
Expert Systems with Applications Pub Date : 2025-04-03 DOI: 10.1016/j.eswa.2025.127432
Nossayba Darraz , Ikram Karabila , Anas El-Ansari , Nabil Alami , Mostafa El Mallahi
{"title":"Advancing recommendation systems with DeepMF and hybrid sentiment analysis: Deep learning and Lexicon-based integration","authors":"Nossayba Darraz ,&nbsp;Ikram Karabila ,&nbsp;Anas El-Ansari ,&nbsp;Nabil Alami ,&nbsp;Mostafa El Mallahi","doi":"10.1016/j.eswa.2025.127432","DOIUrl":"10.1016/j.eswa.2025.127432","url":null,"abstract":"<div><div>In the hotel industry, ensuring customer satisfaction and providing personalized recommendations are crucial elements for creating a remarkable guest experience. However, traditional recommendation systems encounter several challenges that hinder their effectiveness. These challenges include cold start problems, where it is difficult to make recommendations for new or less-rated items, as well as data sparsity, which limits the availability of relevant information. Additionally, accurately interpreting the diverse sentiments expressed by customers in their reviews poses another significant challenge. This study tackles these challenges by integrating sentiment analysis into hotel recommendation systems, aiming to capture and analyze guest opinions and sentiments from their reviews. This study aims to enhance recommendation systems by integrating a hybrid sentiment analysis approach. The approach combines lexicon-based techniques and deep learning methodologies, using TextBlob with Bag of Words and a Multilayer Perceptron (MLP) algorithm to analyze the sentiment of textual data. The hybrid sentiment analysis approach exhibits an impressive accuracy rate of 88.63%, demonstrating its effectiveness in capturing sentiment from customer reviews. This integration enables recommendation systems to better understand and incorporate customer sentiments, leading to improved personalized recommendations. Moreover, we combine this hybrid sentiment analysis with DeepMF for collaborative hotel recommendations, which yields a remarkable Root Mean Square Error (RMSE) of 0.1. By integrating sentiment analysis into the recommendation system, we gain valuable insights into customer preferences, leading to improved recommendation quality and personalization. This research highlights the potential of sentiment analysis in optimizing customer experience management within the hotel industry, providing a valuable tool for enhancing guest satisfaction and engagement.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"279 ","pages":"Article 127432"},"PeriodicalIF":7.5,"publicationDate":"2025-04-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143760699","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 Hybrid Similarity-Aware Graph Neural Network with Transformer for Node Classification
IF 7.5 1区 计算机科学
Expert Systems with Applications Pub Date : 2025-04-02 DOI: 10.1016/j.eswa.2025.127292
Aman Singh , Shahid Shafi Dar , Ranveer Singh, Nagendra Kumar
{"title":"A Hybrid Similarity-Aware Graph Neural Network with Transformer for Node Classification","authors":"Aman Singh ,&nbsp;Shahid Shafi Dar ,&nbsp;Ranveer Singh,&nbsp;Nagendra Kumar","doi":"10.1016/j.eswa.2025.127292","DOIUrl":"10.1016/j.eswa.2025.127292","url":null,"abstract":"<div><div>Node classification has gained significant importance in graph deep learning with real-world applications such as recommendation systems, drug discovery, and citation networks. Graph Convolutional Networks and Graph Transformers have achieved superior performance in node classification tasks. However, the key concern with Graph Convolutional Networks is over-squashing, which limits their ability to capture long-range dependencies in the network. Additionally, Graph Transformers face scalability challenges, making it difficult to process large graphs efficiently. To address this, we propose a novel framework, A Hybrid <strong>SI</strong>milarity-Aware <strong>G</strong>raph <strong>N</strong>eural <strong>Net</strong>work with Transformer for Node Classification (SIGNNet), which capitalizes local and global structural information, enhances the model’s capability to effectively capture fine-grained relationships and broader contextual patterns within the graph structure. The proposed method leverages Graph Convolutional Networks alongside a score-based mechanism to effectively capture local and global node interactions while addressing the limitations of over-squashing. Our proposed method employs a novel Personalized PageRank-based node sampling method to address scalability issues by generating subgraphs of nodes. Additionally, SIGNNet incorporates a novel attention mechanism, Structure-Aware Multi-Head Attention (SA-MHA) which integrates node structural information for informed attention weighting, enabling the model to prioritize nodes based on topological significance. Extensive experiments demonstrate the significant improvements achieved by the proposed method over existing state-of-the-art methods, with average accuracy gains of 6.03%, 5.47%, 4.78%, 19.10%, 19.61%, 7.22%, 19.54% and 14.94% on Cora, Citeseer, CS, Wisconsin, Texas, Actor, Cornell and Chameleon datasets, respectively.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"279 ","pages":"Article 127292"},"PeriodicalIF":7.5,"publicationDate":"2025-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143747542","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
Hybrid machine learning approach for parallel machine scheduling under uncertainty
IF 7.5 1区 计算机科学
Expert Systems with Applications Pub Date : 2025-04-02 DOI: 10.1016/j.eswa.2025.127427
Aleksandar Stanković , Goran Petrović , Rajko Turudija , Danijel Marković , Žarko Ćojbašić
{"title":"Hybrid machine learning approach for parallel machine scheduling under uncertainty","authors":"Aleksandar Stanković ,&nbsp;Goran Petrović ,&nbsp;Rajko Turudija ,&nbsp;Danijel Marković ,&nbsp;Žarko Ćojbašić","doi":"10.1016/j.eswa.2025.127427","DOIUrl":"10.1016/j.eswa.2025.127427","url":null,"abstract":"<div><div>Today’s manufacturing companies face numerous challenges in a dynamic and highly complex business environment. Effective planning, within the management of the production system, has a key role in achieving business success and achieving a competitive advantage for any company. The main setting of the research is the integration of three phases into one intelligent system. The first phase of the research consists of big data optimization of the planning model in the parallel connection of machines, the second phase of the experiment includes the application of different machine learning models, while the third represents the optimization of the planning model in the parallel connection of machines with stochastic processing times, which represents one of the more difficult NP problems of combinatorial optimization. The integration of machine learning models and job planning models in parallel machine connection under conditions of uncertainty is a big challenge. One of the reasons for the application of machine learning models is the influence of the input optimization parameters on the observed objective function. By choosing optimal optimization parameters, it is possible to solve the problem of parallel machine planning with stochastic processing times. The research in the paper aims to significantly improve the performance and reliability of machine planning in various industrial environments by proposing a robust and adaptive solution that can adapt to dynamic conditions and provide optimal results. The main purpose of the paper is the application and integration of an artificial intelligence model in a planning system in order to increase productivity, thereby increasing the competitiveness of small and medium-sized enterprises on the market. These tools can be relatively easily adapted to the needs of the company and would thus enable a better organization of business activities, as well as lower costs and greater business flexibility. The results of the experiment show the success of the proposed methodology.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"279 ","pages":"Article 127427"},"PeriodicalIF":7.5,"publicationDate":"2025-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143747545","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
BiTA: Bi-directional tuning for lossless acceleration in large language models
IF 7.5 1区 计算机科学
Expert Systems with Applications Pub Date : 2025-04-02 DOI: 10.1016/j.eswa.2025.127305
Feng Lin , Hanling Yi , Yifan Yang , Hongbin Li , Xiaotian Yu , Guangming Lu , Rong Xiao
{"title":"BiTA: Bi-directional tuning for lossless acceleration in large language models","authors":"Feng Lin ,&nbsp;Hanling Yi ,&nbsp;Yifan Yang ,&nbsp;Hongbin Li ,&nbsp;Xiaotian Yu ,&nbsp;Guangming Lu ,&nbsp;Rong Xiao","doi":"10.1016/j.eswa.2025.127305","DOIUrl":"10.1016/j.eswa.2025.127305","url":null,"abstract":"<div><div>Large language models (LLMs) typically employ autoregressive generation during inference, leading to high memory bandwidth demand and consequently extended latency. An effective strategy to mitigate this inefficiency is speculative decoding, which reduces the number of model inference calls, thereby lowering memory bandwidth requirements. In this paper, we propose BiTA (<strong>Bi</strong>-directional <strong>T</strong>uning for lossless <strong>A</strong>cceleration), an innovative speculative decoding method that expedites LLMs through streamlined semi-autoregressive generation and draft verification. BiTA enhances LLMs with a parameter-efficient design called bi-directional tuning, enabling semi-autoregressive generation, while leveraging an efficient tree-based decoding mechanism to perform draft candidate generation and verification in parallel, ensuring that the outputs of accelerated LLMs remain identical to those of their original autoregressive counterparts. As a lightweight plug-in module, BiTA seamlessly boosts the inference efficiency of existing LLMs without requiring additional assistance models or incurring significant extra memory costs. Applying BiTA, LLaMA-2-70B-Chat achieves a <span><math><mrow><mn>2</mn><mo>.</mo><mn>7</mn><mo>×</mo></mrow></math></span> speedup on the MT-Bench benchmark. Extensive experiments confirm that BiTA surpasses state-of-the-art speculative decoding methods. The code is available at <span><span>https://github.com/linfeng93/BiTA</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"279 ","pages":"Article 127305"},"PeriodicalIF":7.5,"publicationDate":"2025-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143760587","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
How do LLMs perform on Turkish? A multi-faceted multi-prompt evaluation
IF 7.5 1区 计算机科学
Expert Systems with Applications Pub Date : 2025-04-02 DOI: 10.1016/j.eswa.2025.127421
Mustafa Burak Topal , Aysun Bozanta , Ayşe Başar
{"title":"How do LLMs perform on Turkish? A multi-faceted multi-prompt evaluation","authors":"Mustafa Burak Topal ,&nbsp;Aysun Bozanta ,&nbsp;Ayşe Başar","doi":"10.1016/j.eswa.2025.127421","DOIUrl":"10.1016/j.eswa.2025.127421","url":null,"abstract":"<div><div>Turkish is a resourced language, but it remains underresearched, causing it to lag behind recent advances focused on LLM in NLP. Comprehensive evaluations and standardized benchmarks are crucial for advancing Turkish LLMs, as they help identify strengths and weaknesses. This study aims to evaluate large language models (LLMs) in Turkish, focusing on their performance in understanding and trustworthiness tasks. Our analysis examines the models’ prompt robustness and compares fine-tuned LLMs with their chat-based counterparts. We evaluated 10 open-source models for 11 different tasks using 17 datasets. These data sets comprised original Turkish sources and translations from English. We also included Turkish and multilingual pre-trained language models (PLMs) as baselines for certain tasks. The gemma2-9b-it model outperformed other chat LLMs in both understanding and trustworthiness tasks. However, in fine-tuning experiments, no single model emerged as the best, with the top PLM achieving results comparable to the best LLM. Significant performance variations on paraphrased prompts highlight the need for improved robustness, which can be achieved by fine-tuning as our results suggest. Models like Trendyol-8B-chat-v2.0 and wiroai-turkish-llm-8b, adapted to Turkish via instruction tuning, often surpassed the LLMs they are based on. This suggests that adapting gemma2-9b-it to Turkish might lead to a model that is stronger than gemma-2-9b-it, the best model in our evaluation. This study evaluates Turkish LLMs and shares key insights from the experiments.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"279 ","pages":"Article 127421"},"PeriodicalIF":7.5,"publicationDate":"2025-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143769130","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
LLM-guided fuzzy kinematic modeling for resolving kinematic uncertainties and linguistic ambiguities in text-to-motion generation
IF 7.5 1区 计算机科学
Expert Systems with Applications Pub Date : 2025-04-02 DOI: 10.1016/j.eswa.2025.127283
Ali Asghar Manjotho , Tekie Tsegay Tewolde , Ramadhani Ally Duma , Zhendong Niu
{"title":"LLM-guided fuzzy kinematic modeling for resolving kinematic uncertainties and linguistic ambiguities in text-to-motion generation","authors":"Ali Asghar Manjotho ,&nbsp;Tekie Tsegay Tewolde ,&nbsp;Ramadhani Ally Duma ,&nbsp;Zhendong Niu","doi":"10.1016/j.eswa.2025.127283","DOIUrl":"10.1016/j.eswa.2025.127283","url":null,"abstract":"<div><div>Generating realistic and coherent human motions from text descriptions is essential for applications in computer vision, computer animations, and digital environments. However, existing text-to-motion generation models often overlook kinematic uncertainties and linguistic ambiguities, leading to unnatural and misaligned motion sequences. To address these issues, we propose a novel framework that integrates fuzzy kinematic modeling with large language model (LLM) guidance to jointly model kinematic uncertainties and resolve linguistic ambiguities. Our approach first extracts rich kinematic attributes from raw motion data and converts them into fuzzy kinematic facts (FKFs), which serve as an uncertainty-aware motion representation across different kinematic hierarchies. Simultaneously, we refine ambiguous text descriptions by extracting contextual terms using LLM-guided few-shot in-context learning, enhancing text with additional semantic clarity. These FKFs and contextual terms are then used to train a diffusion-based motion generation model, ensuring semantically accurate and physically plausible motion synthesis. To further enhance kinematic structural consistency in FKF representations, we introduce a Graph-Augmented Self-Attention (GASA) module, which injects spatio-temporal relational constraints into the diffusion process, improving motion coherence and structural integrity. Evaluations on HumanML3D and KIT-ML datasets demonstrate that our method outperforms state-of-the-art models, achieving the lowest FID scores (0.052 and 0.091) and reducing kinematic uncertainty footprint by 21.1% and 17.7%, respectively. The source code and additional resources are publicly available at <span><span>https://alimanjotho.github.io/llm-fqk-t2m</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"279 ","pages":"Article 127283"},"PeriodicalIF":7.5,"publicationDate":"2025-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143760586","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
DAG-HFC: Dual-domain attention and graph optimization network for heterogeneous graph feature completion
IF 7.5 1区 计算机科学
Expert Systems with Applications Pub Date : 2025-04-01 DOI: 10.1016/j.eswa.2025.127293
Yihao Jiang , Zhaowei Liu , Yongchao Song , Yao Shan , Tengjiang Wang , Dong Yang
{"title":"DAG-HFC: Dual-domain attention and graph optimization network for heterogeneous graph feature completion","authors":"Yihao Jiang ,&nbsp;Zhaowei Liu ,&nbsp;Yongchao Song ,&nbsp;Yao Shan ,&nbsp;Tengjiang Wang ,&nbsp;Dong Yang","doi":"10.1016/j.eswa.2025.127293","DOIUrl":"10.1016/j.eswa.2025.127293","url":null,"abstract":"<div><div>Heterogeneous graphs consist of multiple types of edges and nodes. Existing heterogeneous graph neural networks can be understood as a node feature smoothing process guided by the graph structure, which can accurately simulate complex relationships in the real world. However, due to privacy and data scarcity in the real world, some node features are inevitably missing. In addition, with the increase of model depth and the aggregation of multiple types of meta-paths, node embeddings tend to be consistent, resulting in semantic confusion and overfitting problems. To improve the quality of node embeddings, we propose a heterogeneous graph feature completion network (DAG-HFC). DAG-HFC reconstructs the adjacency matrix through topology-based initial feature filling and Jaccard similarity, and fuses the meta-path adjacency matrix to integrate high-order semantic relations and structural information, enhancing the ability to model deep interactions between nodes. The designed Dual-domain differential multi-head attention mechanism combines global feature capture with local feature refinement to effectively alleviate semantic confusion and over-smoothing problems. We conduct extensive experiments on four heterogeneous graph datasets and show that DAG-HFC can significantly improve performance compared with other methods.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"278 ","pages":"Article 127293"},"PeriodicalIF":7.5,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143748629","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
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
相关产品
×
本文献相关产品
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信