Expert Systems最新文献

{"title":"Towards Clustering of Incomplete Mixed-Attribute Data","authors":"Chuyao Zhang,&nbsp;Xinxi Chen,&nbsp;Zexi Tan,&nbsp;Fangqing Gu,&nbsp;Yuzhu Ji,&nbsp;Yiqun Zhang","doi":"10.1111/exsy.70074","DOIUrl":"https://doi.org/10.1111/exsy.70074","url":null,"abstract":"<p>Clustering analysis is one of the most important data mining and knowledge discovery tools in real applications. Since the widespread presence of missing values hampers clustering performance, missing values imputation becomes necessary for data pre-processing. However, for the common datasets composed of both numerical and categorical attributes (also known as mixed-attribute datasets), most existing imputation methods suffer from the following three limitations: (1) Only feasible for a certain type of attribute; (2) Encounter difficulties in considering the interdependence between different types of attributes; (3) Short in exploiting the information provided by the incomplete mix-valued objects. As a result, the original data distribution can be ill-restored, misleading the downstream clustering tasks. This paper therefore proposes a clustering-imputation co-learning method for incomplete mixed-attribute datasets to address these issues. This method integrates imputation and clustering into one learning process, emphasising the interrelationships between mixed attributes during the imputation process and exploiting the information of incomplete objectsduring clustering. It turns out that appropriate recovery of the dataset and accurate clustering can be better achieved through a cross-coupling manner. Experiments on various datasets validate the promising efficacy of the proposed method.</p>","PeriodicalId":51053,"journal":{"name":"Expert Systems","volume":"42 7","pages":""},"PeriodicalIF":3.0,"publicationDate":"2025-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1111/exsy.70074","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143949946","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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}

{"title":"ELWARD: Empowering Language Model With World Insights and Human-Aligned Reward Design","authors":"Yongping Du,&nbsp;Siyuan Li,&nbsp;Rui Yan,&nbsp;Ying Hou,&nbsp;Honggui Han","doi":"10.1111/exsy.70055","DOIUrl":"https://doi.org/10.1111/exsy.70055","url":null,"abstract":"<div>\u0000 \u0000 <p>Large language models (LLMs) have made significant progress in many tasks, but they may also generate biased or misleading outputs. Alignment techniques address this issue by refining models to reflect human values, but high-quality preference datasets are limited. This study introduces a method to train a high-performance reward model (RM) by integrating open knowledge with human feedback. We construct the Open Knowledge and Human Feedback (OK-HF) dataset, comprising 39.8 million open preference data entries and 30,000 human feedback entries. The dual-stage aligning strategy is proposed to combine preference pre-training with domain adaptation, leveraging multi-objective optimization to enhance learning from both preference data and fine-grained human feedback. The Open Knowledge and Human-feedback Reward Model (OKH-RM), designed with the dual-stage aligning strategy on the OK-HF dataset, demonstrates exceptional performance in aligning LLMs with human preferences. The experimental results show that OKH-RM outperforms Llama2-RM, Qwen-RM and Ultra-RM, particularly achieving an accuracy of 85.93% on the Stanford SHP dataset. The model has shown advanced capabilities in detecting low-quality repetitive responses and mitigating biases related to response length.</p>\u0000 </div>","PeriodicalId":51053,"journal":{"name":"Expert Systems","volume":"42 6","pages":""},"PeriodicalIF":3.0,"publicationDate":"2025-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143939460","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}

{"title":"Smartphone Sensor-Based Physiological Parameter Monitoring: Advances, Apps, and Discussions","authors":"Shuni Li,&nbsp;Mingzhi Wang,&nbsp;Xiong Jiang,&nbsp;Xingyao Li,&nbsp;Jiawei Du,&nbsp;Nan Ji,&nbsp;Junxin Chen","doi":"10.1111/exsy.70068","DOIUrl":"https://doi.org/10.1111/exsy.70068","url":null,"abstract":"<div>\u0000 \u0000 <p>With the increasing prevalence of smartphones and advancements in sensors, smartphone-based solutions for physiological parameter monitoring appear to offer notable advantages over traditional methods, potentially enhancing safety, convenience and efficiency. This paper aims to present a systematic survey of smartphone sensor-based physiological parameter monitoring apps, with particular discussions of gaps between their current functional capabilities and recent advances. We conducted a systematic analysis of relevant apps available on the App Store and Google Play, mainly focusing on four vital signs: heart rate (HR), blood pressure(BP), body temperature (BT) and respiratory rate (RR), as well as oxygen saturation (<span></span><math>\u0000 <semantics>\u0000 <mrow>\u0000 <msub>\u0000 <mi>S</mi>\u0000 <mi>p</mi>\u0000 </msub>\u0000 <msub>\u0000 <mi>O</mi>\u0000 <mn>2</mn>\u0000 </msub>\u0000 </mrow>\u0000 <annotation>$$ {S}_p{O}_2 $$</annotation>\u0000 </semantics></math>), blood glucose (BG) and haemoglobin (Hb). The analysis revealed that HR measurement apps were the most prevalent, while BP, <span></span><math>\u0000 <semantics>\u0000 <mrow>\u0000 <msub>\u0000 <mi>S</mi>\u0000 <mi>p</mi>\u0000 </msub>\u0000 <msub>\u0000 <mi>O</mi>\u0000 <mn>2</mn>\u0000 </msub>\u0000 </mrow>\u0000 <annotation>$$ {S}_p{O}_2 $$</annotation>\u0000 </semantics></math>, BT and RR measurement apps were comparatively fewer, and no smartphone sensor-based BG measurement apps were identified. The contact photoplethysmography method is widely adopted by current apps, while non-contact approach holds potential. Novel techniques require further investigation beyond laboratory settings to enhance robustness. Smartphone-based measurement of physiological parameters shows promise, though further research and development are needed to bridge the gap between current capabilities and the demands of accurate, real-world health monitoring.</p>\u0000 </div>","PeriodicalId":51053,"journal":{"name":"Expert Systems","volume":"42 6","pages":""},"PeriodicalIF":3.0,"publicationDate":"2025-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143939462","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}

{"title":"Approaches to Automatic Classification, Detection and Segmentation of Breast Arterial Calcification Using Deep Learning","authors":"Dominic Maguire,&nbsp;John D. Thompson,&nbsp;Sunil Vadera,&nbsp;Katy Szczepura","doi":"10.1111/exsy.70069","DOIUrl":"https://doi.org/10.1111/exsy.70069","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Objective</h3>\u0000 \u0000 <p>Cardiovascular disease (CVD) is the leading cause of premature death in the United Kingdom with one type, coronary artery disease, killing more than two times as many women as breast cancer. Recently, researchers have noted that breast arterial calcification (BAC), which is regularly observed as an incidental finding on mammograms, could be used to risk-stratify women for CVD. However, identifying BAC is known to be a tedious, expensive and time-consuming process. Thus, this paper investigates deep learning models for BAC classification, object detection and segmentation.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methodology</h3>\u0000 \u0000 <p>A data set, annotated under the guidance of two consultant radiologists, was created using data augmentation. This was used to evaluate several alternative deep learning models.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>A modified ResNet22 classification network achieved a test accuracy of 80%, indicating that this method could be used as a flag for the presence or absence of BAC. We also used this network for feature extraction in a YOLOv4 BAC object detection network. Despite improving on a recent similar study, this latter network performed poorly with very low average precision scores at several thresholds. More promising was our DeepLabv3+-based BAC segmentation network, which reached similar high global accuracy scores to three recent studies and a BFScore of over 70% specifically for BAC. It also performed satisfactorily on an unseen data set.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>These results show the potential for using classification and segmentation models as part of a pipeline for detecting BAC.</p>\u0000 </section>\u0000 </div>","PeriodicalId":51053,"journal":{"name":"Expert Systems","volume":"42 6","pages":""},"PeriodicalIF":3.0,"publicationDate":"2025-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1111/exsy.70069","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143939459","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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}

{"title":"Enhancing Analytic Hierarchy Process Modelling Under Uncertainty With Fine-Tuning LLM","authors":"Haeun Park,&nbsp;Hyunjoo Oh,&nbsp;Feng Gao,&nbsp;Ohbyung Kwon","doi":"10.1111/exsy.70051","DOIUrl":"https://doi.org/10.1111/exsy.70051","url":null,"abstract":"<p>Given that decision-making typically encompasses stages such as problem recognition, the generation of alternatives, and the selection of the optimal choice, Large Language Models (LLMs) are progressively being integrated into tasks requiring the enumeration and comparative evaluation of alternatives, thereby promoting more rational decision-making frameworks. Analysing the extent to which LLMs exhibit meaningful performance at each stage of the decision-making process has thus become a critical area of inquiry. In particular, LLMs hold the potential to identify latent relationships within contextual information and data related to the problem domain. This capability enables them to propose novel evaluation criteria or alternatives that may otherwise be overlooked by human designers. This study seeks to advance the modelling and evaluation of the analytical hierarchy process (AHP), a widely utilised multiple criteria decision making (MCDM) method, by leveraging LLMs. To achieve this, a methodology was developed for constructing AHP models using LLMs fine-tuned with domain-specific documents. The performance of the proposed methodology was assessed by evaluating the extent to which its outputs aligned with reference hierarchies and criteria created by human experts under predefined AHP frameworks. Additionally, the study examined the model's efficacy in generating complete AHP hierarchies and criteria in scenarios where these were not predefined. For empirical validation, the proposed methodology was applied to assess and improve the management performance of six-sector agricultural enterprises. Comparative analysis of the LLM-based AHP results with human expert evaluations was conducted to determine the validity and robustness of the approach. The findings provide insights into the potential of LLMs to contribute to structured decision-making and enhance the application of MCDM methods.</p>","PeriodicalId":51053,"journal":{"name":"Expert Systems","volume":"42 6","pages":""},"PeriodicalIF":3.0,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1111/exsy.70051","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143925945","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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}

{"title":"Heterogeneous Graph Distillation for Stance Prediction","authors":"Yibing Lu,&nbsp;Jingyun Sun,&nbsp;Yang Li","doi":"10.1111/exsy.70058","DOIUrl":"https://doi.org/10.1111/exsy.70058","url":null,"abstract":"<div>\u0000 \u0000 <p>Stance prediction is a critical task in public opinion analysis, aiming to identify users' viewpoints on specific events. Existing research often relies on user interactions for stance inference but generally underutilizes multi-source heterogeneous information such as user entities, opinion text, issues and topics. To address this limitation, this study proposes a stance prediction approach based on heterogeneous entity modeling. By integrating four types of heterogeneous entities to capture similarity in users' participation in issues, the proposed method improves stance inference accuracy. Specifically, we design a heterogeneous graph knowledge extraction framework that fully incorporates both content features and structural semantic information of various entities. First, we construct a heterogeneous information network to capture different types of social media entities and their interactions, learning rich feature representations in the process. Next, we employ matrix factorization to assess users' preferences toward specific issues. Finally, by introducing a knowledge distillation mechanism, the approach significantly enhances prediction accuracy with only a modest increase in computational cost. Experimental results on public datasets demonstrate that our method outperforms existing baselines, verifying its effectiveness.</p>\u0000 </div>","PeriodicalId":51053,"journal":{"name":"Expert Systems","volume":"42 6","pages":""},"PeriodicalIF":3.0,"publicationDate":"2025-05-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143919342","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}

{"title":"Towards Optimal Guidance of Autonomous Swarm Drones in Dynamic Constrained Environments","authors":"Yunes Alqudsi,&nbsp;Murat Makaraci","doi":"10.1111/exsy.70067","DOIUrl":"https://doi.org/10.1111/exsy.70067","url":null,"abstract":"<p>As autonomous drone swarms become increasingly important for complex missions, there remains a critical need for integrated approaches that can simultaneously handle task allocation and safe navigation in dynamic environments. This paper addresses the challenge of optimally allocating tasks and generating collision-free trajectories for drone swarms operating in obstacle-rich settings. Our proposed Swarm Allocation and Route Generation (SARG) framework integrates optimal task assignment with dynamically feasible trajectory planning, enabling efficient mission completion while ensuring safe navigation through complex 3D workspaces. Using quadrotors as our experimental platform, the framework incorporates both Drone-to-Obstacle and Drone-to-Drone collision avoidance algorithms, alongside a modified path planning algorithm that enhances simultaneous graph search efficiency. Our extensive experiments demonstrate that the SARG framework significantly improves performance over existing approaches. The SARG framework, while maintaining a 100% collision avoidance rate in dense environments, achieves a 21.6% reduction in the computation time of the simultaneous graph searching phase compared to conventional methods, contributing to overall system efficiency. These results establish SARG as a viable solution for real-world autonomous drone swarm applications in complex, dynamic settings. Supporting Information, including animated simulations, are available at https://youtu.be/56oabPTUz4g.</p>","PeriodicalId":51053,"journal":{"name":"Expert Systems","volume":"42 6","pages":""},"PeriodicalIF":3.0,"publicationDate":"2025-05-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1111/exsy.70067","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143919343","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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}

{"title":"RETRACTION: Linear Pricing Game Based Power Control With Resource Allocation and Interference Management in Device-to-Device Communication for IoT Applications","authors":"","doi":"10.1111/exsy.70071","DOIUrl":"https://doi.org/10.1111/exsy.70071","url":null,"abstract":"<p>\u0000 <b>RETRACTION:</b> <span>K. Pandey</span> and <span>R. Arya</span>, “ <span>Linear Pricing Game Based Power Control With Resource Allocation and Interference Management in Device-to-Device Communication for IoT Applications</span>,” <i>Expert Systems</i> <span>40</span>, no. <span>5</span> (<span>2023</span>): e13094, https://doi.org/10.1111/exsy.13094.\u0000 </p><p>The above article, published online on 01 July 2022 in Wiley Online Library (wileyonlinelibrary.com), has been retracted by agreement between the journal Editor-in-Chief, David Camacho; and John Wiley &amp; Sons Ltd. The retraction has been agreed upon due to several critical issues. The rationale for the research and the specific challenges it intends to address are not adequately explained. Furthermore, the analysis is brief, lacking sufficient detail, and fails to evaluate the advantages and issues of existing works in the field. Finally, in the validation phase, simulations are used rather than real data and there is no comparison to other state-of-the-art techniques. The editors consider the results and conclusions reported in this article unreliable. The authors do not agree with the retraction.</p>","PeriodicalId":51053,"journal":{"name":"Expert Systems","volume":"42 6","pages":""},"PeriodicalIF":3.0,"publicationDate":"2025-05-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1111/exsy.70071","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143919448","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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}

{"title":"Enhancing Deepfake Audio Detection: A ResNet Framework Based on Hybrid Features and Self-Attention Mechanism","authors":"Lian Huang,&nbsp;Jixiang Yang,&nbsp;Jinhong Zhao,&nbsp;Chunxiang Wu","doi":"10.1111/exsy.70054","DOIUrl":"https://doi.org/10.1111/exsy.70054","url":null,"abstract":"<div>\u0000 \u0000 <p>Due to the successful application of deep learning, audio spoofing detection has made significant progress. Spoofed audio with speech synthesis or voice conversion can be detected by many countermeasures well. However, an automatic speaker verification system is still vulnerable to spoofing attacks such as replay or deepfake audio. Deepfake audio, generated using text-to-speech (TTS) and voice conversion (VC) algorithms, poses a particularly significant challenge. To address this vulnerability, we propose a novel framework incorporating hybrid features and a self-attention mechanism for enhanced spoofing detection. Our approach is distinguished by the following key contributions: (1) A novel dual-path feature extraction architecture, leveraging parallel convolutional neural networks (CNNs) and Short-Time Fourier Transform (STFT) with Mel-frequency filtering to capture complementary deep learning and Mel-spectrogram features, respectively; (2) A max-pooling-based feature fusion strategy, concatenating the extracted features to preserve crucial discriminative information; (3) The integration of a self-attention mechanism to dynamically weight and focus on salient temporal-spectral patterns within the fused feature representation; (4) A ResNet-based classifier, augmented with linear layers, for robust spoofing classification. Rigorous evaluation on the ASVspoof 2021 dataset demonstrates the efficacy of our proposed framework. We achieve state-of-the-art performance, attaining Equal Error Rate (EER) of 9.67% in the physical access (PA) scenario and 8.94% in the deepfake task. These results correspond to substantial relative improvements of 74.60% and 60.05%, respectively, compared to the best-performing baseline systems. These findings underscore the superior discriminative power of our hybrid feature approach, highlighting its ability to capture richer utterance details compared to conventional single-modality feature representations. This work offers a promising new direction for developing robust ASV systems resilient to increasingly sophisticated spoofing attacks.</p>\u0000 </div>","PeriodicalId":51053,"journal":{"name":"Expert Systems","volume":"42 6","pages":""},"PeriodicalIF":3.0,"publicationDate":"2025-05-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143919447","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}

{"title":"SoK: Federated Learning and Unlearning for Medical Image Analysis","authors":"Khaoula ElBedoui,&nbsp;Walid Barhoumi,&nbsp;Jungwon Cho","doi":"10.1111/exsy.70063","DOIUrl":"https://doi.org/10.1111/exsy.70063","url":null,"abstract":"<div>\u0000 \u0000 <p>Medical image analysis is a critical component of modern healthcare, enabling accurate disease diagnosis and effective patient treatment. However, the process is fraught with challenges, including inter- and intra-observer variability, time constraints, and data-related issues such as privacy, heterogeneity and accessibility. Within this framework, Federated Learning (FL) has emerged as a promising solution, allowing collaborative model training across distributed healthcare entities without sharing sensitive patient data. This study provides a comprehensive Systematization of Knowledge (SoK) review of FL and its extension, Federated Unlearning (FU), within the context of medical image analysis. FL enables privacy-preserving, decentralised model training, while FU addresses the ‘Right To Be Forgotten’, ensuring compliance with data protection regulations like GDPR and HIPAA. We explore the opportunities and challenges of FL and FU, detailing their methodologies, frameworks, datasets, and evaluation metrics. The review highlights the potential of FL and FU to enhance diagnostic accuracy, improve patient care, and foster trust in AI-driven healthcare systems. We also identify research gaps and propose future directions for advancing FL and FU in medical imaging, emphasising the need for interdisciplinary collaboration and the development of dedicated frameworks. Thus, this study aims to bridge the gap between theoretical advancements and practical applications, paving the way for more robust and privacy-compliant AI models in healthcare.</p>\u0000 </div>","PeriodicalId":51053,"journal":{"name":"Expert Systems","volume":"42 6","pages":""},"PeriodicalIF":3.0,"publicationDate":"2025-05-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143914507","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}