Expert Systems with Applications最新文献

{"title":"Autonomous obstacle avoidance decision method for spherical underwater robot based on brain-inspired spiking neural network","authors":"Boyang Zhang,&nbsp;Huiming Xing,&nbsp;Zhicheng Zhang,&nbsp;Weixing Feng","doi":"10.1016/j.eswa.2025.127021","DOIUrl":"10.1016/j.eswa.2025.127021","url":null,"abstract":"<div><div>Autonomous obstacle avoidance is a critical capability for underwater robots to operate safely and sustainably in complex, unfamiliar, and unknown underwater environments. Existing methods often lack information processing and intelligent rapid decision-making ability similar to the human brain, making it difficult to adapt to the complex and challenging underwater environment. To address these limitations, with the spherical underwater robot (SUR) as the research object, a novel brain-inspired spiking neural network, neuromorphic hybrid deep deterministic policy gradient (Neuro-HDDPG), is proposed in this paper. The soft reset membrane potential update mechanism is designed to better represent the variation of spiking neuron membrane potentials. By integrating the spiking neural network and deep reinforcement learning, the proposed Neuro-HDDPG is composed of a soft reset spiking actor normal network (SANN) and deep critic normal network (DCNN). The SANN consists of soft reset improved leaky integrate-and-fire (SR-ILIF) neurons, and the DCNN comprises artificial neurons, realizing autonomous obstacle avoidance exploration of SUR in complex and unknown environments, with more temporal continuity and biological interpretability. To evaluate the obstacle avoidance efficiency of the proposed Neuro-HDDPG, through the ablation studies and comparison experiments with other known methods, the proposed Neuro-HDDPG achieved the highest success rate of 91% and 92%, respectively, in the two underwater evaluation environments with different levels of complexity, demonstrating superior obstacle avoidance performance and forming a reliable and efficient underwater obstacle avoidance decision-making capability. Simultaneously, the concept of combining spiking neural network with deep reinforcement learning provides an intelligent and reliable reference for other unmanned underwater intelligent systems.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"274 ","pages":"Article 127021"},"PeriodicalIF":7.5,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143480044","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":"Q-fractional fuzzy influence pair domination number to locate and control smog area","authors":"Fahad Ur Rehman,&nbsp;Tabasam Rashid,&nbsp;Muhammad Tanveer Hussain","doi":"10.1016/j.eswa.2025.126886","DOIUrl":"10.1016/j.eswa.2025.126886","url":null,"abstract":"&lt;div&gt;&lt;div&gt;An intuitionistic fuzzy graph &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mo&gt;(&lt;/mo&gt;&lt;mi&gt;I&lt;/mi&gt;&lt;mi&gt;F&lt;/mi&gt;&lt;mi&gt;G&lt;/mi&gt;&lt;mo&gt;)&lt;/mo&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; and its extensions could not handle the situation of the form &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;msub&gt;&lt;mrow&gt;&lt;mi&gt;η&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mn&gt;2&lt;/mn&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;mo&gt;=&lt;/mo&gt;&lt;mrow&gt;&lt;mo&gt;{&lt;/mo&gt;&lt;mrow&gt;&lt;mo&gt;(&lt;/mo&gt;&lt;msub&gt;&lt;mrow&gt;&lt;mo&gt;ħ&lt;/mo&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;mo&gt;,&lt;/mo&gt;&lt;mn&gt;0&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;8&lt;/mn&gt;&lt;mo&gt;,&lt;/mo&gt;&lt;mn&gt;0&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;7&lt;/mn&gt;&lt;mo&gt;)&lt;/mo&gt;&lt;/mrow&gt;&lt;mo&gt;,&lt;/mo&gt;&lt;mrow&gt;&lt;mo&gt;(&lt;/mo&gt;&lt;msub&gt;&lt;mrow&gt;&lt;mo&gt;ħ&lt;/mo&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mn&gt;2&lt;/mn&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;mo&gt;,&lt;/mo&gt;&lt;mn&gt;0&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;9&lt;/mn&gt;&lt;mo&gt;,&lt;/mo&gt;&lt;mn&gt;0&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;8&lt;/mn&gt;&lt;mo&gt;)&lt;/mo&gt;&lt;/mrow&gt;&lt;mo&gt;,&lt;/mo&gt;&lt;mrow&gt;&lt;mo&gt;(&lt;/mo&gt;&lt;msub&gt;&lt;mrow&gt;&lt;mo&gt;ħ&lt;/mo&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mn&gt;3&lt;/mn&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;mo&gt;,&lt;/mo&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;mo&gt;,&lt;/mo&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;mo&gt;)&lt;/mo&gt;&lt;/mrow&gt;&lt;mo&gt;}&lt;/mo&gt;&lt;/mrow&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; because &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mn&gt;0&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;8&lt;/mn&gt;&lt;mo&gt;+&lt;/mo&gt;&lt;mn&gt;0&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;7&lt;/mn&gt;&lt;mo&gt;=&lt;/mo&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;5&lt;/mn&gt;&lt;mo&gt;&gt;&lt;/mo&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;, &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mn&gt;0&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;9&lt;/mn&gt;&lt;mo&gt;+&lt;/mo&gt;&lt;mn&gt;0&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;8&lt;/mn&gt;&lt;mo&gt;=&lt;/mo&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;mo&gt;.&lt;/mo&gt;&lt;mn&gt;7&lt;/mn&gt;&lt;mo&gt;&gt;&lt;/mo&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;, and &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;mo&gt;+&lt;/mo&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;mo&gt;=&lt;/mo&gt;&lt;mn&gt;2&lt;/mn&gt;&lt;mo&gt;&gt;&lt;/mo&gt;&lt;mn&gt;1&lt;/mn&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;. In this article, we proposed the concept of a q-fractional fuzzy influence graph &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mo&gt;(&lt;/mo&gt;&lt;mi&gt;q&lt;/mi&gt;&lt;msub&gt;&lt;mrow&gt;&lt;mi&gt;f&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mi&gt;r&lt;/mi&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;mi&gt;F&lt;/mi&gt;&lt;mi&gt;I&lt;/mi&gt;&lt;mi&gt;G&lt;/mi&gt;&lt;mo&gt;)&lt;/mo&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;. A &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;q&lt;/mi&gt;&lt;msub&gt;&lt;mrow&gt;&lt;mi&gt;f&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mi&gt;r&lt;/mi&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;mi&gt;F&lt;/mi&gt;&lt;mi&gt;I&lt;/mi&gt;&lt;mi&gt;G&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; can indicate degrees of membership and non-membership 100% independently using the q-intercept of a straight line. We explore some ideas like a strongest q-fractional fuzzy influence pair &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mo&gt;(&lt;/mo&gt;&lt;mi&gt;S&lt;/mi&gt;&lt;mi&gt;G&lt;/mi&gt;&lt;mi&gt;q&lt;/mi&gt;&lt;msub&gt;&lt;mrow&gt;&lt;mi&gt;f&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mi&gt;r&lt;/mi&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;mi&gt;F&lt;/mi&gt;&lt;mi&gt;I&lt;/mi&gt;&lt;mi&gt;P&lt;/mi&gt;&lt;mo&gt;)&lt;/mo&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;, strong q-fractional fuzzy influence pair &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mo&gt;(&lt;/mo&gt;&lt;mi&gt;S&lt;/mi&gt;&lt;mi&gt;q&lt;/mi&gt;&lt;msub&gt;&lt;mrow&gt;&lt;mi&gt;f&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mi&gt;r&lt;/mi&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;mi&gt;F&lt;/mi&gt;&lt;mi&gt;I&lt;/mi&gt;&lt;mi&gt;P&lt;/mi&gt;&lt;mo&gt;)&lt;/mo&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;, weak q-fractional fuzzy influence pair &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mo&gt;(&lt;/mo&gt;&lt;mi&gt;W&lt;/mi&gt;&lt;mi&gt;q&lt;/mi&gt;&lt;msub&gt;&lt;mrow&gt;&lt;mi&gt;f&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mi&gt;r&lt;/mi&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;mi&gt;F&lt;/mi&gt;&lt;mi&gt;I&lt;/mi&gt;&lt;mi&gt;P&lt;/mi&gt;&lt;mo&gt;)&lt;/mo&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;, q-fractional fuzzy influence cut-node &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mo&gt;(&lt;/mo&gt;&lt;mi&gt;q&lt;/mi&gt;&lt;msub&gt;&lt;mrow&gt;&lt;mi&gt;f&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mi&gt;r&lt;/mi&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;mi&gt;F&lt;/mi&gt;&lt;mi&gt;I&lt;/mi&gt;&lt;mi&gt;C&lt;/mi&gt;&lt;mi&gt;N&lt;/mi&gt;&lt;mo&gt;)&lt;/mo&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;, q-fractional fuzzy influence bridge &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mo&gt;(&lt;/mo&gt;&lt;mi&gt;q&lt;/mi&gt;&lt;msub&gt;&lt;mrow&gt;&lt;mi&gt;f&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mi&gt;r&lt;/mi&gt;&lt;/mrow&gt;&lt;/msub&gt;&lt;mi&gt;F&lt;/mi&gt;&lt;mi&gt;I&lt;/mi&gt;&lt;mi&gt;B&lt;/mi&gt;&lt;mo&gt;)&lt;/mo&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;, q-fractional fuzzy influence cut-pair &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mo&gt;(&lt;/mo&gt;&lt;mi&gt;q&lt;/mi&gt;&lt;msub&gt;&lt;mrow&gt;&lt;mi&gt;f&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mi&gt;r&lt;/mi&gt;&lt;","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"274 ","pages":"Article 126886"},"PeriodicalIF":7.5,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143474266","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":"Online sequential Extreme learning Machine (OSELM) based denoising of encrypted image","authors":"Biniyam Ayele Belete,&nbsp;Demissie Jobir Gelmecha,&nbsp;Ram Sewak Singh","doi":"10.1016/j.eswa.2025.126999","DOIUrl":"10.1016/j.eswa.2025.126999","url":null,"abstract":"<div><div>With the increasing demand for secure digital communication, image encryption is essential for safeguarding sensitive information from unauthorized access. Conventional encryption techniques frequently encounter difficulties when dealing with noise and maintaining image quality during transmission. This study presents an innovative approach to image encryption that integrates an Online Sequential Extreme Learning Machine (OSELM) autoencoder and chaotic systems with DNA code for enhanced image encryption and denoising. The proposed method amalgamates an OSELM autoencoder, a hyperchaotic system incorporating two mersisters, a two-dimensional sine map, and DNA coding to effectively protect images from various types of noise such as Gaussian, Salt and Pepper, Quantization, Speckle, Sensor, and Environmental noises. OSELM-based denoising boosts encryption images’ resilience to noise attacks, ensuring strong security and image quality. Simulation results have shown that the proposed method achieved a key space of around 10²⁴⁰ or 2⁷⁹⁷, with information entropy values near 8 for encrypted images. This method also attains Number of Pixels Change Rate (NPCR) values between 99.59 % and 99.64 % and Unified Average Changing Intensity (UACI) values ranging from 32.97 % to 33.92 %, Peak Signal to Noise Ratio (PSNR) values of denoised images between 23.63 and 37.45, indicating outstanding performance in terms do both security and noise resilience. Additionally, histogram analysis, correlation analysis, and Mean Squared Error (MSE) results highlight the algorithm’s strong resistance to statistical attacks. At the same time, NPCR and UACI values affirm its robustness against differential attacks. The algorithm exhibits high sensitivity to key variations of up to 10^-16, ensuring robust protection even with slight changes to the encryption of the decryption key. Finally, the NIST SP800-22 statistical tests confirm the randomness of the encrypted image’s bitstream, reinforcing its cryptographic strength. This dual approach effectively addresses the challenges of image security and noise resilience, safeguarding the integrity and clarity of digital images. This makes it highly suitable for the secure transmission and storage of digital images in diverse fields such as medicine, photography, biology, astronomy, and defense.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"274 ","pages":"Article 126999"},"PeriodicalIF":7.5,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143480045","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 sparse orthogonal basis clustering for unsupervised feature selection","authors":"Jianyu Miao ,&nbsp;Jingjing Zhao ,&nbsp;Tiejun Yang ,&nbsp;Yingjie Tian ,&nbsp;Yong Shi ,&nbsp;Mingliang Xu","doi":"10.1016/j.eswa.2025.126890","DOIUrl":"10.1016/j.eswa.2025.126890","url":null,"abstract":"<div><div>Unsupervised Feature Selection (UFS), which identifies the optimal-related feature subset from the original feature set to lower the dimensionality of data without label information, has had a high profile in recent years. Given the absence of label information, the existing UFS approaches usually utilize graph and manifold learning techniques to retain the intrinsic structure of the data. The inclusion of irrelevant and redundant features and noise, would inevitably lower the quality of the structure. For this purpose, in this paper, we come up with Robust Sparse Orthogonal Basis Clustering (RSOBC), a novel method for UFS that integrates feature selection process with clustering task into a unified framework. Instead of explicitly utilizing the pre-computed local information, such a strategy focuses on exploring the inherent clustering structures of data. RSOBC leverages the log-based function as the loss to lessen the effect of noise and outliers, thereby enhancing its robustness. To select the more useful and discriminative features, the <span><math><msub><mrow><mi>ℓ</mi></mrow><mrow><mn>2</mn><mo>,</mo><mn>1</mn></mrow></msub></math></span> norm is employed as the sparse regularization to encourage sparsity of the projection matrix. Meanwhile, we adopt the low redundancy regularization to make the weights of the correlated features small. In this way, the correlated features cannot be selected simultaneously. Consequently, the projection matrix, centroid matrix and cluster label matrix are learned simultaneously, such that the intrinsic structure is constructed in a more accurate way. The resulting optimization can be readily tackled by multi-block Alternating Direction Method of Multipliers (ADMM) based algorithm. Comprehensive experiments have been carried out on nine diverse real-world datasets. The results demonstrate that RSOBC surpasses many state-of-the-art UFS approaches, which indicates its effectiveness and superiority.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"274 ","pages":"Article 126890"},"PeriodicalIF":7.5,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143474352","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":"Efficient data stream clustering via elastic sparse representation and Bayesian dictionary learning","authors":"Yao Li ,&nbsp;Ming Chi ,&nbsp;Xiaodong Liu","doi":"10.1016/j.eswa.2025.126889","DOIUrl":"10.1016/j.eswa.2025.126889","url":null,"abstract":"<div><div>Existing data stream clustering algorithms face two key challenges: (1) reducing resource consumption by designing algorithms that can handle continuous data streams; (2) efficiently processing large-scale data and identifying the intrinsic structures of data objects. To address these challenges, this paper introduces an efficient data stream clustering method via elastic sparse representation and Bayesian dictionary learning (ESRBDL). Firstly, we control the size of the landmark windows to ensure data object richness while using fuzzy rules to limit the number of data objects, thereby managing continuous data streams. Secondly, the elastic penalty is introduced to enhance model flexibility, balancing sparsity while improving the identification of different data characteristics. Thirdly, we apply Bayesian theory to infer the true posterior distribution from the initial dictionary distributions, effectively identifying intrinsic relationships among data objects. Finally, we use the spectral clustering algorithm to cluster data streams. Additionally, comparative experiments were conducted on five synthetic and six real datasets to benchmark the proposed method against advanced data stream clustering methods. The experimental results demonstrate the effectiveness and robustness of ESRBDL in data stream clustering.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"274 ","pages":"Article 126889"},"PeriodicalIF":7.5,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143474271","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":"LLM-AE-MP: Web Attack Detection Using a Large Language Model with Autoencoder and Multilayer Perceptron","authors":"Jing Yang ,&nbsp;Yuangui Wu ,&nbsp;Yuping Yuan ,&nbsp;Haozhong Xue ,&nbsp;Sami Bourouis ,&nbsp;Mahmoud Abdel-Salam ,&nbsp;Sunil Prajapat ,&nbsp;Lip Yee Por","doi":"10.1016/j.eswa.2025.126982","DOIUrl":"10.1016/j.eswa.2025.126982","url":null,"abstract":"<div><div>Web applications store sensitive data, making them prime targets for cybercriminals and posing national security risks. This study introduces a new approach to distinguishing legitimate and malicious hypertext transfer protocol (HTTP) requests using an autoencoder (AE). The integration of AE allows for efficient feature distillation, enhancing the sensitivity of the model to anomalies in HTTP traffic. The AE framework is combined with a transductive long short-term memory (TLSTM) network, which is trained with an advanced generative adversarial network (GAN). Using GAN promotes an adaptive learning environment, significantly boosting the robustness and generalizability of our method against evolving web attack vectors. TLSTM uses transductive learning to focus on data points near the test set, improving the adaptability of the model to outperform traditional LSTM models. In our GAN, the generator purposely excludes gradients from the most influential batch elements, improving the ability of the model to generate diverse and generalized outputs. After training the AE, its latent representations are passed to a multilayer perceptron (MLP) for detection tasks. To address the imbalanced classification in MLP, we use a reinforcement learning (RL) strategy. The RL approach strategically adjusts incentives, enhancing the performance of the model in identifying less frequent but critical malicious instances, thereby supporting a balanced security assessment. Our evaluations using the CSIC 2010 (Spanish National Research Council 2010), FWAF (web application firewall), and HttpParams datasets show that our method outperforms existing techniques, achieving (Accuracy, F-measure, geometric mean (G-means), and area under the curve (AUC)) reaching (90.937%, 89.755%, 88.446%, 0.838), (89.055, 90.663%, 88.334%, 0.847) and (92.242%, 93.774%, 91.356%, 0.897), respectively. Moreover, our model achieves efficient runtime and memory usage across the datasets, providing a practical solution for real-time web attack detection. These results confirm the effectiveness of the model in security contexts, representing a substantial advancement in web attack detection and the improvement of investigative strategies.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"274 ","pages":"Article 126982"},"PeriodicalIF":7.5,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143488654","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":"Image authentication and encryption algorithm based on RSA cryptosystem and chaotic maps","authors":"Deep Singh,&nbsp;Sandeep Kumar","doi":"10.1016/j.eswa.2025.126883","DOIUrl":"10.1016/j.eswa.2025.126883","url":null,"abstract":"<div><div>In the era of today’s open network communication, security issues of private data over these networks are at their peak, and it is challenging to secure data transmission. Therefore, to overcome these security crises, cryptographers need to develop efficient and robust encryption algorithms. This article proposes a new multilayer image ciphering scheme along with a modified authentication technique for grayscale and color images by using the RSA public key cryptosystem and two chaotic maps. The chaotic key sequences generated through a 2D piecewise smooth non-linear chaotic map (2D-PSNCM) are utilized to perform permutation of rows and columns to enhance the level of confusion. Also, bitwise XOR is applied to fulfill an intricate level of diffusion. To decrease the correlation and to strengthen the security level, the scrambling of pixels is created by using a suitable number of iterations of the Baker map. For authenticity purposes, the modified digital signature is proposed by using the RSA public key cryptosystem for partially encrypted images. The proposed digital signature is employed for further encryption purposes as well. The scheme provides multilayer security and is well-protected against different brute-force attacks. The secret keys and their arrangement in each layer of encryption are imperative for the correct decryption. The proposed algorithm’s robustness and efficiency are verified using multiple RGB and grayscale images. Security analysis demonstrates the validation by comparing it with some recent cryptographic image encryption algorithms. These results show the proposed algorithm’s robustness, such as vast sizeable keyspace, very low correlation coefficient, high sensitivity for keys, and good information entropy. Altogether, this proves the high efficiency of the proposed algorithm in resisting statistical attacks.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"274 ","pages":"Article 126883"},"PeriodicalIF":7.5,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143474268","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":"Efficiency is the rule: Domain adaptive semantic segmentation with minimal annotations","authors":"Tianyu Huai ,&nbsp;Junhang Zhang ,&nbsp;Xingjiao Wu ,&nbsp;Jian Jin ,&nbsp;Liang He","doi":"10.1016/j.eswa.2025.126892","DOIUrl":"10.1016/j.eswa.2025.126892","url":null,"abstract":"<div><div>Active domain adaptation aims to select a few yet informative samples of the target domain for manual annotation to improve model performance. However, a critical observation in our research is the less-than-ideal domain alignment of existing active domain adaptive semantic segmentation (ADASS) methods. Specifically, they only measure the complementarity between target samples and the source domain but neglect the degree of domain shift in the active sample selection process. Furthermore, the impact of hard samples on domain alignment and model discriminative ability is underestimated. To tackle these issues, we propose a framework that contains a novel main-sub anchor modeling method and Confusing Sample Selection (CSS) and Offset Sample Selection (OSS) strategies. While improving the model performance of the ADASS task, we also consider that there remains a substantial resource demand. To solve this issue, we introduce the Instance Assignment Module (IAM). Extensive experiments on GTAV <span><math><mo>→</mo></math></span> Cityscapes and SYNTHIA <span><math><mo>→</mo></math></span> Cityscapes benchmarks, demonstrate that our method sets a new standard in both weakly supervised domain adaptive semantic segmentation (WDASS) and ADASS tasks, achieving the optimal trade-off between annotation cost and model performance.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"274 ","pages":"Article 126892"},"PeriodicalIF":7.5,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143474273","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":"Matching of manual operation trajectories and warehouse operation information: A data chain Construction method based on indoor positioning technology","authors":"Yunhai Xiang ,&nbsp;Kun Wang ,&nbsp;Xinru Wu","doi":"10.1016/j.eswa.2025.127016","DOIUrl":"10.1016/j.eswa.2025.127016","url":null,"abstract":"<div><div>Accurate data collection in manual warehouses faces significant challenges due to the reliance on singular information collection method and the operators’ flexibility, which impedes data-driven, intelligent decision-making in warehouse operations. This paper addresses this problem to construct the data chain using indoor positioning technology (DCC-IPS). A unique feature of the proposed approach is the integration of the operators’ positioning data with the layout, operations, and tasks in the warehouse, facilitating a deep fusion of new external data and internal business data. Experiments conducted at Southwest Jiaotong University’s laboratory have demonstrated that the DCC-IPS achieves a matching accuracy exceeding 80%. Compared to traditional scanning with PDA, DCC-IPS reduces the delay in operation recognition by 20 s in the experimental scenario. Furthermore, by utilizing the data chain for evaluating operators’ capability and optimizing task assignments, our numerical experiments showed a 22.13% increase in efficiency over random assignments. These results highlight the accuracy and effectiveness of DCC-IPS in enhancing operational efficiency in warehouses.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"274 ","pages":"Article 127016"},"PeriodicalIF":7.5,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143488652","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":"A deep attention-based encoder for the prediction of type 2 diabetes longitudinal outcomes from routinely collected health care data","authors":"Enrico Manzini ,&nbsp;Bogdan Vlacho ,&nbsp;Josep Franch-Nadal ,&nbsp;Joan Escudero ,&nbsp;Ana Génova ,&nbsp;Elisenda Reixach ,&nbsp;Erich Andrés ,&nbsp;Israel Pizarro ,&nbsp;Dídac Mauricio ,&nbsp;Alexandre Perera-Lluna","doi":"10.1016/j.eswa.2025.126876","DOIUrl":"10.1016/j.eswa.2025.126876","url":null,"abstract":"<div><div>Recent evidence indicates that Type 2 Diabetes Mellitus (T2DM) is a complex and highly heterogeneous disease involving various pathophysiological and genetic pathways, which presents clinicians with challenges in disease management. While deep learning models have made significant progress in helping practitioners manage T2DM treatments, several important limitations persist. In this paper we propose DARE, a model based on the transformer encoder, designed for analyzing longitudinal heterogeneous diabetes data. The model can be easily fine-tuned for various clinical prediction tasks, enabling a computational approach to assist clinicians in the management of the disease. We trained DARE using data from over 200,000 diabetic subjects from the primary healthcare SIDIAP database, which includes diagnosis and drug codes, along with various clinical and analytical measurements. After an unsupervised pre-training phase, we fine-tuned the model for predicting three specific clinical outcomes: i) occurrence of comorbidity, ii) achievement of target glycemic control (defined as glycated hemoglobin <span><math><mrow><mo>&lt;</mo><mn>7</mn><mtext>%</mtext></mrow></math></span>) and iii) changes in glucose-lowering treatment. In cross-validation, the embedding vectors generated by DARE outperformed those from baseline models (comorbidities prediction task <span><math><mrow><mi>A</mi><mi>U</mi><mi>C</mi><mo>=</mo><mn>0</mn><mo>.</mo><mn>88</mn></mrow></math></span>, treatment prediction task <span><math><mrow><mi>A</mi><mi>U</mi><mi>C</mi><mo>=</mo><mn>0</mn><mo>.</mo><mn>91</mn></mrow></math></span>, HbA1c target prediction task <span><math><mrow><mi>A</mi><mi>U</mi><mi>C</mi><mo>=</mo><mn>0</mn><mo>.</mo><mn>82</mn></mrow></math></span>). Our findings suggest that attention-based encoders improve results with respect to different deep learning and classical baseline models when used to predict different clinical relevant outcomes from T2DM longitudinal data.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"274 ","pages":"Article 126876"},"PeriodicalIF":7.5,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143479888","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}