Information Sciences最新文献

{"title":"Optimization of inner and general rules","authors":"Beata Zielosko ,&nbsp;Mikhail Moshkov ,&nbsp;Evans Teiko Tetteh","doi":"10.1016/j.ins.2025.122466","DOIUrl":"10.1016/j.ins.2025.122466","url":null,"abstract":"<div><div>The subject of the paper concerns the problem of deriving decision rules from distributed data. The paper examines issues of learning general and inner decision rules from a set of decision trees. Inner rules refer to the routes within decision trees from the root to leaf nodes, while general rules are arbitrary rules derived from attributes found in the set of decision trees. The paper illustrates that the optimization of general decision rules is NP-hard problem, so the authors propose heuristics <span><math><msub><mrow><mi>H</mi></mrow><mrow><mn>1</mn></mrow></msub></math></span> and <span><math><msub><mrow><mi>H</mi></mrow><mrow><mn>2</mn></mrow></msub></math></span> for this issue. Taking into account induction and optimization of inner decision rules an algorithm <span><math><mi>A</mi></math></span> is employed. Additionally, an approach based on global optimization relative to length, support, and sequential optimization is proposed. The presented algorithms were studied considering two perspectives (i) knowledge discovery from data and (ii) knowledge representation. In the first case, it is possible to discover patterns from the data and verify the induced model, in the second case, it is possible to represent knowledge in a comprehensible and explainable way. These elements are important in an era of heterogeneous, distributed data sources. Experiments were carried out on selected datasets from UCI ML and Kaggle repositories. In order to create a distributed data structure, an approach based on reducts induced by a genetic algorithm was employed. Obtained results show that there are cases where the global rule-based classifiers built in the framework of optimization of inner decision rules perform better in terms of accuracy than that of local models induced directly from subtables. In the case of algorithms <span><math><msub><mrow><mi>H</mi></mrow><mrow><mn>1</mn></mrow></msub></math></span> and <span><math><msub><mrow><mi>H</mi></mrow><mrow><mn>2</mn></mrow></msub></math></span>, the low complexity of models based on decision rules induced from a set of decision trees is noted.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"719 ","pages":"Article 122466"},"PeriodicalIF":8.1,"publicationDate":"2025-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144564055","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Global exponential synchronization of competitive neural networks with D operators and mixed delays and an application to secure communication","authors":"Weijing Yan ,&nbsp;Yu Xue ,&nbsp;Xian Zhang","doi":"10.1016/j.ins.2025.122473","DOIUrl":"10.1016/j.ins.2025.122473","url":null,"abstract":"<div><div>The global exponential synchronization (GES) of competitive neural networks (CNNs) with mixed delays and D operators is the main topic of this research. To ensure GES between the drive and response CNNs, effective controllers are designed in this paper. Based on this, we propose a direct analysis approach based on system solutions, which not only eliminates the need to construct the Lyapunov–Krasovskii functional but also simplifies the process of solving the synchronization criteria and reduces the workload and computational complexity to a large extent. Additionally, it is discussed how the proposed synchronization method is applied to secure communication. Eventually, numerical examples and an application example will confirm the theoretical and practical usefulness of the approach. It is important to note that this research is the first to investigate the GES issue for CNNs with D operators and mixed delays.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"719 ","pages":"Article 122473"},"PeriodicalIF":8.1,"publicationDate":"2025-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144524236","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Fractional impulsive fuzzy differential systems with multi-order in (1,2): The solution representation and asymptotical stabilization","authors":"Ngo Van Hoa","doi":"10.1016/j.ins.2025.122461","DOIUrl":"10.1016/j.ins.2025.122461","url":null,"abstract":"<div><div>The work is devoted to the investigation of a semilinear multi-order fractional fuzzy differential system influenced by impulsive effects, utilizing the Caputo derivative within the granular approach of fuzzy functions. By applying a fractional Laplace-like transform, an explicit representation of the system's solution is obtained. Utilizing the Gronwall-Bellman inequality, a sufficient condition for local asymptotic stability is established. Furthermore, a linear state feedback controller is formulated to ensure asymptotic stabilization. To verify the practical applicability of the proposed method, several numerical examples are presented.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"719 ","pages":"Article 122461"},"PeriodicalIF":8.1,"publicationDate":"2025-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144570889","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Hierarchical interpretable vision reasoning driven based depth estimation method in adverse weather conditions through a multi-modal large language model","authors":"Wenfeng Mi ,&nbsp;Chengwei Yang ,&nbsp;Yuzhe Qian ,&nbsp;He Chen","doi":"10.1016/j.ins.2025.122460","DOIUrl":"10.1016/j.ins.2025.122460","url":null,"abstract":"<div><div>Although the most advanced depth estimation methods achieve impressive results under ideal weather conditions, their reliability is significantly reduced in adverse weather conditions such as night, rain, and snow. In this paper, we propose a hierarchical interpretable vision reasoning (HIVR) driven by a multi-modal large language model (MLLM), which offers an effective solution capable of reliable operation in adverse weather environments. First, we construct an interpretable and scalable visual model reasoning framework called HIVR. Subsequently, we fine-tune an MLLM to identify weather conditions, automatically generate invocation programs, and drive HIVR by sequentially invoking relevant models until the depth estimation is completed. We also design a self-supervised monocular depth estimation framework that combines convolutional neural network (CNN) and transformer architectures to extract better features and achieve greater robustness. Additionally, we utilize Generative Adversarial Networks (GANs) to create large-scale synthetic KITTI and Robotcar datasets that include six types of adverse weather conditions. Finally, comprehensive comparative and ablation experiments are conducted to demonstrate the effectiveness of the proposed method.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"719 ","pages":"Article 122460"},"PeriodicalIF":8.1,"publicationDate":"2025-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144570891","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Event-triggered sliding mode control for interval type-2 fuzzy interconnected systems under Markov-model-based hybrid cyberattacks","authors":"Xiulin Wang ,&nbsp;Feng Li ,&nbsp;Sangmoon Lee ,&nbsp;Hao Shen","doi":"10.1016/j.ins.2025.122467","DOIUrl":"10.1016/j.ins.2025.122467","url":null,"abstract":"<div><div>This paper studies the event-triggered sliding mode control problem for interval type-2 fuzzy interconnected systems under hybrid cyberattacks modeled by Markov processes. Under the framework of an interval type-2 fuzzy model, the nonlinear relationship in the interconnected system is modeled and analyzed, and a fuzzy sliding mode controller is designed which can effectively resist hybrid cyberattacks. To rationally utilize limited channel resources, a decentralized event-triggered mechanism is used to reduce unnecessary information transmission. By constructing a Lyapunov function, some criteria are deduced to ensure that the closed-loop system is stochastically passive and the reachability of the designed sliding area is guaranteed. Finally, a four-area networked interconnected system is used to verify the effectiveness of the decentralized fuzzy sliding mode control strategy.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"719 ","pages":"Article 122467"},"PeriodicalIF":8.1,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144535102","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Robust stochastic configuration network based perception model for furnace temperature in municipal solid waste incineration process","authors":"Jingcheng Guo ,&nbsp;Zhe Dong ,&nbsp;Wei Guo ,&nbsp;Aijun Yan","doi":"10.1016/j.ins.2025.122477","DOIUrl":"10.1016/j.ins.2025.122477","url":null,"abstract":"<div><div>Addressing the challenging issue of reduced accuracy in furnace temperature perception models due to asymmetric outliers in operational data of the municipal solid waste incineration process, we propose a novel robust stochastic configuration network perception model for furnace temperature in municipal solid waste incineration processes. A skewed t distribution with the heavy-tailed characteristic is adopted to model the prior distribution of outliers in the operational data of the solid waste incineration process. Subsequently, the maximum likelihood estimation method is employed to solve the output weights of the furnace temperature perception model. Additionally, the output weights and prior distribution position hyperparameters are updated by the expectation-conditional maximization algorithm. Comparative experimental results demonstrate the superiority of our proposed method in terms of both accuracy in furnace temperature perception and robustness against outliers. Our approach holds significant potential for application in critical fields such as furnace temperature prediction and optimal control during the incineration process.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"719 ","pages":"Article 122477"},"PeriodicalIF":8.1,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144570888","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-objective CNN optimization: A robust framework for automated model design","authors":"Sefa Aras ,&nbsp;Elif Aras ,&nbsp;Eyüp Gedikli ,&nbsp;Hamdi Tolga Kahraman","doi":"10.1016/j.ins.2025.122468","DOIUrl":"10.1016/j.ins.2025.122468","url":null,"abstract":"<div><div>Convolutional neural networks (CNNs) have demonstrated high performance in classifying image data. However, CNNs require expert-level hyperparameter tuning and involve substantial computational complexity, which hinders their effective deployment in real-time and IoT systems. Research on CNNs indicates that optimal hyperparameter configurations can enhance inference speed while improving classification accuracy. We developed a novel approach that uses multi-objective optimization to design CNN models automatically. Our method tunes hyperparameters to balance classification accuracy and inference speed. We define classification performance and inference speed as objectives and balance them using a Pareto-optimal strategy. Unlike traditional approaches, MoCNN systematically explores Pareto-optimal trade-offs between classification performance and computational efficiency, enabling a fully automated architecture search without manual intervention. In this study, we select NSGA-II as our preferred MOEA while ensuring the framework remains flexible enough to accommodate other evolutionary strategies. Experimental evaluations on benchmark datasets (CIFAR-10, CIFAR-100, and FRUITS-360) demonstrate that MoCNN reduces inference time by up to 72.02% on average and improves classification accuracy by 6.72% compared to manually tuned CNN architectures. By eliminating the need for heuristic hyperparameter selection, MoCNN enhances scalability and is particularly well suited for real-time, mobile AI, and edge-computing applications. Our results show that MoCNN outperforms state-of-the-art optimization frameworks in both computational efficiency and predictive performance, highlighting its potential for deployment in scenarios where accuracy and speed are critical.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"719 ","pages":"Article 122468"},"PeriodicalIF":8.1,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144524234","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"DEEP-CWS: Distilling Efficient pre-trained models with Early exit and Pruning for scalable Chinese Word Segmentation","authors":"Shiting Xu","doi":"10.1016/j.ins.2025.122470","DOIUrl":"10.1016/j.ins.2025.122470","url":null,"abstract":"<div><div>Chinese Word Segmentation (CWS) is essential for a broad spectrum of tasks in natural language processing (NLP). However, the high inference cost of large pre-trained models like BERT and RoBERTa restricts their scalability in practical deployments. To overcome this limitation, we introduce <strong>DEEP-CWS</strong>, a novel approach for efficient CWS that distills pre-trained transformer models into lightweight CNNs, incorporating pruning, early exit mechanisms, and ONNX optimization to improve inference speed significantly. Our method achieves over 100 times speedup in inference latency relative to the teacher model without compromising segmentation quality, with an F1 score of 97.81 on the PKU benchmark. These characteristics make DEEP-CWS particularly well-suited for real-time scenarios and large-scale processing. Extensive experiments on public benchmarks and a legal-domain dataset validate the robustness and transferability of our framework. We also release our code base to support reproducibility and future research.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"719 ","pages":"Article 122470"},"PeriodicalIF":8.1,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144524235","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Semi-supervised anomaly detection via reinforcement learning-enabled method with causal inference","authors":"Xiangwei Chen ,&nbsp;Ruliang Xiao ,&nbsp;Zhixia Zeng ,&nbsp;Shi Zhang ,&nbsp;Xin Du","doi":"10.1016/j.ins.2025.122463","DOIUrl":"10.1016/j.ins.2025.122463","url":null,"abstract":"<div><div>Semi-supervised anomaly detection is critical in ensuring system reliability. However, existing methods rely on correlation rather than causality, potentially causing misinterpretations due to confounding factors. Although reinforcement learning-based approaches can detect known and unknown anomalies with limited labeled samples, they still face issues such as insufficient use of prior knowledge, limited model flexibility, and incomplete reward feedback. To address the above problems, this paper innovatively constructs a counterfactual causal reinforcement learning model, termed <strong>Tri</strong>ple-Assisted <strong>C</strong>ausal <strong>R</strong>einforcement <strong>L</strong>earning <strong>A</strong>nomaly <strong>D</strong>etector (Tri-CRLAD). The model leverages causal inference to extract the intrinsic causal feature, enhancing the agent's utilization of prior knowledge and improving its generalization capability. In addition, Tri-CRLAD integrates a triple decision support mechanism that includes a historical similarity-based sampling strategy, an adaptive threshold smoothing approach, and an adaptive decision reward mechanism. These mechanisms further enhance the flexibility and generalization ability of the model, enabling it to effectively respond to various complex and dynamically changing environments. Experimental results across seven diverse sensor signal datasets demonstrate that Tri-CRLAD outperforms nine state-of-the-art baseline methods, exhibiting enhanced robustness against data scarcity, with a performance drop up to 23 percentage points smaller than that of baseline methods under minimal known anomalies. Our code is available at <span><span>https://github.com/Aoudsung/Tri-CRLAD/</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"719 ","pages":"Article 122463"},"PeriodicalIF":8.1,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144535100","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An innovative approach to ensemble learning in bankruptcy prediction using support vector machines and meta fuzzy functions","authors":"Pınar Karadayı Ataş ,&nbsp;Nihat Tak ,&nbsp;Süreyya Özöğür-Akyüz ,&nbsp;Birsen Eygi Erdogan","doi":"10.1016/j.ins.2025.122450","DOIUrl":"10.1016/j.ins.2025.122450","url":null,"abstract":"<div><div>The categorization of banks into successful and unsuccessful is essential for ensuring financial stability, effective risk management, and appropriate regulatory oversight. This study introduces a new ensemble modeling method for bank classification that combines meta fuzzy functions (MFFs) with support vector machines (SVMs). We predict bank status (failed or successful) by analyzing financial ratios, such as liquidity, profitability, and solvency metrics, using a dataset of Turkish commercial banks. Gaussian kernel-based SVMs, known for their strong classification performance, serve as the ensemble's base classifiers. Linear kernel SVMs are employed for comparison with previous studies. Because the data structure is a panel data, the proposed approach is compared with a single panel logistic regression model and a previously proposed ensemble approach. The results show that the MFF-based ensemble outperforms both baseline models, achieving an accuracy of [85.4%] and an AUC-ROC score of [87%]. This work demonstrates how ensemble learning using MFFs can enhance bank classification, providing a strong tool for financial analysts and policymakers in times of economic instability.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"719 ","pages":"Article 122450"},"PeriodicalIF":8.1,"publicationDate":"2025-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144535097","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}