计算机科学最新文献

{"title":"Enhancing Brain Tumor Classification Through Optimal Kernel Selection With \u0000 \u0000 \u0000 \u0000 GL\u0000 1\u0000 \u0000 \u0000 $$ {GL}_1 $$\u0000 -Regularization","authors":"Otmane Mallouk,&nbsp;Nour-Eddine Joudar,&nbsp;Mohamed Ettaouil","doi":"10.1002/ima.70150","DOIUrl":"https://doi.org/10.1002/ima.70150","url":null,"abstract":"<div>\u0000 \u0000 <p>Brain tumors, known for their rapid and aggressive growth, are among the most serious and life-threatening diseases worldwide. This makes the development of automated detection methods essential for saving lives. Deep transfer learning has become a highly effective approach for automating brain tumor classification and medical imaging, offering promising solutions on a global scale. However, leveraging a pretrained model typically involves special adaptations. Existing adaptation methods involve freezing or fine-tuning specific layers without considering the contribution level of individual kernels. This work aims to extend the concept of layer-level contributions to the kernel level by employing an adaptive optimization model. Indeed, this paper presents a novel optimization model that incorporates group lasso regularization to control which kernels are frozen and which are fine-tuned. The proposed model selects optimal source features that contribute to the target task. Additionally, the proposed optimization model is solved utilizing proximal gradient descent. The method was evaluated on a three-class brain tumor classification task, distinguishing between glioma, meningioma, and pituitary tumors, using a medical MRI dataset. Several experiments confirm the efficacy of our model in identifying both frozen and fine-tuned kernels, thereby improving data classification. Subsequently, the results obtained are compared with those of state-of-the-art transfer learning methods for comprehensive comparison.</p>\u0000 </div>","PeriodicalId":14027,"journal":{"name":"International Journal of Imaging Systems and Technology","volume":"35 4","pages":""},"PeriodicalIF":3.0,"publicationDate":"2025-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144550970","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":"Distance-Based Unsupervised Local Outlier Detection: Based Values Analysis to Improve Outlier Detection Using Machine Learning","authors":"Atul Kumar Gupta,&nbsp;Rahul Kumar,&nbsp;Jhankar Moolchandani,&nbsp;Vikas Thada,&nbsp;Mohd Asif Shah,&nbsp;Anoop Kumar Tiwari","doi":"10.1049/cmu2.70060","DOIUrl":"https://doi.org/10.1049/cmu2.70060","url":null,"abstract":"<p>Machine learning faces challenges in detecting outliers, especially in high-dimensional datasets. Effective data quality is crucial for better results, and many algorithms identify outliers by analysing outlying aspects of data objects and objects within the dataset. The proposed Advanced Distance-Based Unsupervised Local Outlier Detection (DU-LOD) method improves this process by continuously evaluating and identifying outliers using unsupervised learning and distance-based calculations. DU-LOD identifies outliers by comparing differences between data objects and their neighbours, making it the first method to combine unsupervised local outlier detection with nearest cluster point identification. Experimental analysis through accuracy performance of 96.12%, detection rate performance of 41.89%, precision of 56.12%, and recall of 1.79% proves that our model performs best over the various parameters compared with other existing algorithms. Therefore, measures such as area under the ROC curve (AUC), precision and recall are more appropriate in such a scenario.</p>","PeriodicalId":55001,"journal":{"name":"IET Communications","volume":"19 1","pages":""},"PeriodicalIF":1.5,"publicationDate":"2025-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1049/cmu2.70060","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144551005","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":"Harmonious colourings of temporal matchings","authors":"Duncan Adamson","doi":"10.1016/j.tcs.2025.115437","DOIUrl":"10.1016/j.tcs.2025.115437","url":null,"abstract":"<div><div>Graph colouring is a fundamental problem in computer science, with a large body of research dedicated to both the general colouring problem and restricted cases. <em>Harmonious colourings</em> are one such restriction, where each edge must contain a globally unique pair of colours, i.e. if an edge connects a vertex coloured <em>x</em> with a vertex coloured <em>y</em>, then no other pair of connected vertices can be coloured <em>x</em> and <em>y</em>. Finding such a colouring in the traditional graph setting is known to be NP-hard, even in trees. This paper considers the generalisation of harmonious colourings to <em>Temporal Graphs</em>, specifically <span><math><mo>(</mo><mi>k</mi><mo>,</mo><mi>t</mi><mo>)</mo></math></span><em>-Temporal matchings</em>, a class of temporal graphs where the underlying graph is a matching (a collection of disconnected components containing pairs of vertices), each edge can appear in at most <em>t</em> timesteps, and each timestep can contain at most <em>k</em> other edges. We provide a complete overview of the complexity landscape of finding <em>temporal harmonious colourings</em> for <span><math><mo>(</mo><mi>k</mi><mo>,</mo><mi>t</mi><mo>)</mo></math></span>-matchings. We show that finding a <em>Temporal Harmonious Colouring</em>, a colouring that is harmonious in each timestep, is NP-hard for <em>(k,t)-Temporal Matchings</em> when <span><math><mi>k</mi><mo>≥</mo><mn>2</mn><mo>,</mo><mi>t</mi><mo>≥</mo><mn>4</mn></math></span>, or when <span><math><mi>k</mi><mo>≥</mo><mn>3</mn></math></span> and <span><math><mi>t</mi><mo>≥</mo><mn>2</mn></math></span>. We further show that this problem is inapproximable for <span><math><mi>t</mi><mo>≥</mo><mn>2</mn></math></span> and an unbounded value of <em>k</em>, and that the problem of determining the temporal harmonious chromatic number of a <span><math><mo>(</mo><mn>2</mn><mo>,</mo><mn>3</mn><mo>)</mo></math></span>-temporal matching can be determined in linear time. Finally, we strengthen this result by a set of upper and lower bounds of the temporal harmonious chromatic number both for individual temporal matchings and for the classes of <span><math><mo>(</mo><mi>k</mi><mo>,</mo><mi>t</mi><mo>)</mo></math></span>-temporal matchings, paths, and cycles.</div></div>","PeriodicalId":49438,"journal":{"name":"Theoretical Computer Science","volume":"1053 ","pages":"Article 115437"},"PeriodicalIF":0.9,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144549637","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":"NiNet: A new invertible neural network architecture more suitable for deep image hiding","authors":"Zishun Ni ,&nbsp;Hang Cheng ,&nbsp;Jiaoling Chen ,&nbsp;Yongliang Xu ,&nbsp;Fei Chen ,&nbsp;Meiqing Wang","doi":"10.1016/j.ipm.2025.104275","DOIUrl":"10.1016/j.ipm.2025.104275","url":null,"abstract":"<div><div>Image hiding through the application of invertible neural network (INN) represents a significant branch within the realm of deep image hiding methodologies, characterized by a compact network architecture and a streamlined parameter count. Nonetheless, when juxtaposed with autoencoder-based approaches, existing INN methods often result in inferior image quality. To surmount this challenge, this paper introduces a novel masking-based image hiding paradigm, establishes a new spatial domain transformation for images, and refines the Swin-transformer block. By integrating these innovations, an INN architecture is crafted that is particularly adept for deep image hiding, termed NiNet. The experimental results demonstrate that NiNet can remarkably address the problem of image hiding. In terms of steganographic image quality, NiNet outperforms the current state-of-the-art (SOTA) algorithms by 0.26 dB on the DIV2K dataset, 1.49 dB on the COCO dataset, and 0.39 dB on the ImageNet dataset. Regarding the quality of secret image recovery, NiNet surpasses the SOTA algorithms by 2.06 dB on DIV2K, 1.98 dB on COCO, and 0.50 dB on ImageNet.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"62 6","pages":"Article 104275"},"PeriodicalIF":7.4,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144548908","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Hypergraph negative influence blocking maximization via influence estimation","authors":"Xian-Jie Zhang ,&nbsp;Xiao-Ming Zhang ,&nbsp;Xu-Dong Huang ,&nbsp;Hai-Feng Zhang","doi":"10.1016/j.ipm.2025.104273","DOIUrl":"10.1016/j.ipm.2025.104273","url":null,"abstract":"<div><div>The rapid advancement of internet technology has accelerated the spread of negative information. To block the impact of such diffusion, selecting a set of positive seed nodes for competitive propagation is a viable strategy. However, existing studies have predominantly focused on pairwise user interactions, neglecting group relationships. To address this gap, we employ a hypergraph to model higher-order group interactions, defining the hypergraph influence blocking maximization (HIBM) problem and proposing the hypergraph competitive susceptible–infected (HCSI) diffusion model. We then prove the monotonicity and submodularity of the objective function, enabling the development of a greedy algorithm. To address the high computational complexity of the greedy approach, we further propose an efficient heuristic method, hypergraph competitive influence estimation (HCIE), which leverages a probabilistic formula to estimate the influence of negative seed sets under the blocking effect of positive nodes, thereby selecting positive nodes to minimize the propagation of negative information. Experimental results demonstrate that the HCIE method reduces the diffusion of negative information by over 10% on certain datasets, showcasing its effectiveness and robustness. Additionally, the HCIE method achieves performance close to that of the greedy algorithm while significantly reducing computational time.</div></div>","PeriodicalId":50365,"journal":{"name":"Information Processing & Management","volume":"62 6","pages":"Article 104273"},"PeriodicalIF":7.4,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144534285","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Marimba: A Python framework for structuring and processing FAIR scientific image datasets","authors":"Christopher J. Jackett ,&nbsp;Kevin Barnard ,&nbsp;Franziska Althaus ,&nbsp;Nicolas Mortimer ,&nbsp;David Webb ,&nbsp;Candice Untiedt ,&nbsp;Aaron Tyndall ,&nbsp;Ian Jameson ,&nbsp;Bec Gorton ,&nbsp;Carlie Devine ,&nbsp;Joanna Strzelecki ,&nbsp;Peter H. Thrall ,&nbsp;Ben Scoulding","doi":"10.1016/j.softx.2025.102251","DOIUrl":"10.1016/j.softx.2025.102251","url":null,"abstract":"<div><div>The rapid advancement of scientific imaging technologies has created significant challenges in managing large-scale image datasets while maintaining compliance with FAIR (Findable, Accessible, Interoperable, and Reusable) data principles. We present Marimba, an open-source Python framework for structuring, processing, and packaging scientific image datasets. Marimba enhances data management through unified workflow processing, automated metadata embedding, efficient data handling, and standardized dataset packaging while integrating with the image FAIR Digital Object (iFDO) metadata standard. The framework's capabilities were evaluated through four diverse marine case studies involving multi-instrument microscopy, automated plankton imagery, deep-sea coral surveys, and historical image digitization. Marimba successfully processed datasets ranging from thousands to hundreds of thousands of images and videos, demonstrating robust performance and scalability. Marimba's modular architecture enables customization for specific research requirements while ensuring consistent data management practices. Results demonstrate Marimba's potential to advance scientific image data management by improving workflow efficiency, data quality, and adherence to FAIR principles throughout the research data lifecycle.</div></div>","PeriodicalId":21905,"journal":{"name":"SoftwareX","volume":"31 ","pages":"Article 102251"},"PeriodicalIF":2.4,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144534720","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":"Beyond the Repertory Grid: New Approaches to Constructivist Knowledge Acquisition Tool Development","authors":"Jeffrey M. Bradshaw, Kenneth M. Ford, Jack R. Adams‐Webber, John H. Boose","doi":"10.1002/j.1098-111x.1993.tb00007.x","DOIUrl":"https://doi.org/10.1002/j.1098-111x.1993.tb00007.x","url":null,"abstract":"<jats:italic>Personal construct theory</jats:italic> provides both a plausible theoretical foundation for knowledge acquisition and a practical approach to modeling. Yet, only a fraction of the ideas latent in this theory have been tapped. Recently, several researchers have been taking another look at the theory, to discover new ways that it can shed light on the foundations and practice of knowledge acquisition. These efforts have led to the development of a new generation of constructivist knowledge acquisition systems: DDUCKS, ICONKAT, and KSSn/KRS. These tools extend repertory grid techniques in various ways and integrate them with ideas springing from complementary perspectives. New understandings of relationships between personal construct theory, assimilation theory, logic, semantic networks, and decision analysis have formed the underpinnings of these systems. Theoretical progress has fostered practical development in system architecture, graphical forms of knowledge representation, analysis and induction techniques, and group use of knowledge acquisition tools.","PeriodicalId":14089,"journal":{"name":"International Journal of Intelligent Systems","volume":"39 1","pages":""},"PeriodicalIF":7.0,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144547050","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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":"Comparison of industrial controllers: Disturbance rejection in an up-and-running industrial flotation process","authors":"Frida Norlund ,&nbsp;Kristian Soltesz ,&nbsp;Margret Bauer","doi":"10.1016/j.jprocont.2025.103478","DOIUrl":"10.1016/j.jprocont.2025.103478","url":null,"abstract":"<div><div>Linear quadratic (LQ) control optimizes a quadratic cost function while following a linear model. It is commercially available in the process industry but often not labeled as such and infrequently used. Froth flotation is a process in the minerals industry that extracts precious metals from a slurry of finely ground rock in consecutive tanks called cells. Flotation cells are often arranged in two parallel streams and the main control task is to regulate the cell levels. In this work, a commercial solution is used to assess the performance of LQ control and compare it to existing PI/PID controllers. Comparing control solutions in the process industry in general and the minerals industry in particular is fraught with difficulty because operating conditions change frequently, especially when studying the capability to reject disturbances. In the flotation process studied here, the disturbances act on two parallel lines, one controlled by an LQ algorithm and the other with the existing PI/PID controllers. This work develops data assessment strategies to isolate events of interest and analyses them both with classical metrics such as integrated absolute error, and with metrics relevant for the operators, such as maximum level deviations. Both these metrics clearly show that the LQ controller performs significantly and consistently better than the PI/PID controllers.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"152 ","pages":"Article 103478"},"PeriodicalIF":3.3,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144536031","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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":"Log-aesthetic curves and generalized Archimedean spirals","authors":"Péter Salvi","doi":"10.1016/j.cagd.2025.102468","DOIUrl":"10.1016/j.cagd.2025.102468","url":null,"abstract":"<div><div>We show that the radials of log-aesthetic curves are generalized Archimedean spirals. Examining the logarithmic curvature histogram reveals that these radials have an inherent similarity to the associated log-aesthetic curves, and can also be used as computationally inexpensive approximants. Different kinds of fits are proposed and discussed through examples. A possible generalization of log-aesthetic curves based on generalized Archimedean spirals is also explored.</div></div>","PeriodicalId":55226,"journal":{"name":"Computer Aided Geometric Design","volume":"121 ","pages":"Article 102468"},"PeriodicalIF":1.3,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144534813","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":"Closing the Gap Between Modeling to Make Sense and Modeling to Implement Systems","authors":"Marc Linster","doi":"10.1002/j.1098-111x.1993.tb00004.x","DOIUrl":"https://doi.org/10.1002/j.1098-111x.1993.tb00004.x","url":null,"abstract":"We view knowledge acquisition for knowledge‐based systems as a constructive model‐building process. From this view we derive several requirements for knowledge modeling environments. We concentrate on those requirements that arise if one wants to support both <jats:italic>modeling to make sense</jats:italic> and <jats:italic>modeling to implement systems</jats:italic> with a single language. For example, among other things, such languages should support multifaceted, bottom‐up construing of observed behavior and they should have operational semantics. We introduce the operational modeling language OMOS, an experimental study that—in a KADS‐like fashion—allows multifaceted model building from a method and a domain point of view, but, unlike KADS conceptual models, results in directly operational systems. Finally, we compare OMOS to other recent developments to highlight differences in the approaches.","PeriodicalId":14089,"journal":{"name":"International Journal of Intelligent Systems","volume":"10 1","pages":""},"PeriodicalIF":7.0,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144547054","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}