Knowledge-Based Systems最新文献

英文中文

CoreNet: Leveraging context-aware representations via MLP networks for CTR prediction

IF 7.2 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Knowledge-Based Systems

Pub Date : 2025-02-15 DOI: 10.1016/j.knosys.2025.113154

Khoi N.P. Dang , Thu Thuy Tran , Ta Cong Son , Tran Tien Anh , Duc Anh Nguyen , Nguyen Van Son

Click-through rate (CTR) prediction is pivotal for industrial recommendation systems and has been driving intensive research. Recent studies emphasized the effectiveness of adaptive methods that use context-aware representations to enhance predictions by dynamically adjusting feature representations across instances and overcoming fixed embedding limitations. The typical architecture for learning context-aware representations involves a network block built on Multi-Head Self-Attention (MHA) or Multi-Layer Perceptron (MLP). Despite promising results, three main challenges arise from these methods. First, relying on a single network block limits the learning potential of the model by providing only one perspective on the interactions. Second, implementing the MHA mechanism requires multiple attention layers for its effectiveness, thereby increasing the complexity of the model. Third, using only a vanilla MLP makes it difficult to combine implicit and explicit feature interactions, which is crucial for successful CTR solutions. To address these issues, we propose a novel model called Context-Aware Net (CoreNet). CoreNet incorporates an advanced module, Context-Aware Module (CAM), which employs a combination of MLP and Hadamard products to generate comprehensive context-aware representations. The CAM component integrates a two-stream network with first-order and second-order aware streams, extracting insights from different perspectives to complement each other and enhance overall performance. Extensive experiments on four public datasets consistently demonstrate that CoreNet outperforms other state-of-the-art models. Notably, our CAM component is lightweight and model-agnostic, facilitating seamless integration into streaming CTR models to enhance performance in a plug-and-play manner¹.

{"title":"CoreNet: Leveraging context-aware representations via MLP networks for CTR prediction","authors":"Khoi N.P. Dang , Thu Thuy Tran , Ta Cong Son , Tran Tien Anh , Duc Anh Nguyen , Nguyen Van Son","doi":"10.1016/j.knosys.2025.113154","DOIUrl":"10.1016/j.knosys.2025.113154","url":null,"abstract":"<div><div>Click-through rate (CTR) prediction is pivotal for industrial recommendation systems and has been driving intensive research. Recent studies emphasized the effectiveness of adaptive methods that use context-aware representations to enhance predictions by dynamically adjusting feature representations across instances and overcoming fixed embedding limitations. The typical architecture for learning context-aware representations involves a network block built on Multi-Head Self-Attention (MHA) or Multi-Layer Perceptron (MLP). Despite promising results, three main challenges arise from these methods. First, relying on a single network block limits the learning potential of the model by providing only one perspective on the interactions. Second, implementing the MHA mechanism requires multiple attention layers for its effectiveness, thereby increasing the complexity of the model. Third, using only a vanilla MLP makes it difficult to combine implicit and explicit feature interactions, which is crucial for successful CTR solutions. To address these issues, we propose a novel model called <strong>Co</strong>ntext-Awa<strong>re Net</strong> (<em>CoreNet</em>). CoreNet incorporates an advanced module, Context-Aware Module (<em>CAM</em>), which employs a combination of MLP and Hadamard products to generate comprehensive context-aware representations. The CAM component integrates a two-stream network with first-order and second-order aware streams, extracting insights from different perspectives to complement each other and enhance overall performance. Extensive experiments on four public datasets consistently demonstrate that <em>CoreNet</em> outperforms other state-of-the-art models. Notably, our CAM component is lightweight and model-agnostic, facilitating seamless integration into streaming CTR models to enhance performance in a plug-and-play manner<span><span><sup>1</sup></span></span>.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"312 ","pages":"Article 113154"},"PeriodicalIF":7.2,"publicationDate":"2025-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143429900","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An efficient algorithm for fast discovery of high-efficiency patterns

IF 7.2 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Knowledge-Based Systems

Pub Date : 2025-02-15 DOI: 10.1016/j.knosys.2025.113157

Irfan Yildirim

The high-efficiency pattern mining (HEPM) problem has recently emerged as a variant of the high-utility pattern mining problem, aiming to identify patterns with the highest profit-to-investment ratio by considering both their utilities and investments. However, due to its vast search space, the HEPM problem is inherently difficult and complex to solve. Existing HEPM algorithms suffer from inefficiencies in runtime and memory usage due to inadequate search space pruning. This study introduces a new algorithm named EHEPM to address this issue more effectively. EHEPM introduces four new upper-bound models to enhance search space pruning and presents two data structures for the accurate and efficient calculation of pattern efficiency and upper-bound values. Experimental results conducted on various datasets demonstrate that EHEPM outperforms existing algorithms in terms of runtime, memory consumption, number of join operations, and scalability.

引用次数: 0

Large language models can better understand knowledge graphs than we thought

IF 7.2 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Knowledge-Based Systems

Pub Date : 2025-02-15 DOI: 10.1016/j.knosys.2025.113060

Xinbang Dai , Yuncheng Hua , Tongtong Wu , Yang Sheng , Qiu Ji , Guilin Qi

When we integrate factual knowledge from knowledge graphs (KGs) into large language models (LLMs) to enhance their performance, the cost of injection through training increases with the scale of the models. Consequently, there is significant interest in developing prompt strategies that effectively incorporate KG information into LLMs. However, the community has not yet comprehensively understood how LLMs process and interpret KG information in different input formats and organizations within prompts, and researchers often rely on trial and error. To address this gap, we design extensive experiments to empirically study LLMs’ comprehension of different KG prompts. At the literal level, we reveal LLMs’ preferences for various input formats (from linearized triples to fluent natural language text). At the attention distribution level, we discuss the underlying mechanisms driving these preferences. We then investigate how the organization of structured knowledge impacts LLMs and evaluate LLMs’ robustness in processing and utilizing KG information in practical scenarios. Our experiments show that (1) linearized triples are more effective than fluent NL text in helping LLMs understand KG information and answer fact-intensive questions; (2) Different LLMs exhibit varying preferences for different organizational formats of triples; (3) LLMs with larger scales are more susceptible to noisy, incomplete subgraphs.

{"title":"Large language models can better understand knowledge graphs than we thought","authors":"Xinbang Dai , Yuncheng Hua , Tongtong Wu , Yang Sheng , Qiu Ji , Guilin Qi","doi":"10.1016/j.knosys.2025.113060","DOIUrl":"10.1016/j.knosys.2025.113060","url":null,"abstract":"<div><div>When we integrate factual knowledge from knowledge graphs (KGs) into large language models (LLMs) to enhance their performance, the cost of injection through training increases with the scale of the models. Consequently, there is significant interest in developing prompt strategies that effectively incorporate KG information into LLMs. However, the community has not yet comprehensively understood how LLMs process and interpret KG information in different input formats and organizations within prompts, and researchers often rely on trial and error. To address this gap, we design extensive experiments to empirically study LLMs’ comprehension of different KG prompts. At the literal level, we reveal LLMs’ preferences for various input formats (from linearized triples to fluent natural language text). At the attention distribution level, we discuss the underlying mechanisms driving these preferences. We then investigate how the organization of structured knowledge impacts LLMs and evaluate LLMs’ robustness in processing and utilizing KG information in practical scenarios. Our experiments show that (1) linearized triples are more effective than fluent NL text in helping LLMs understand KG information and answer fact-intensive questions; (2) Different LLMs exhibit varying preferences for different organizational formats of triples; (3) LLMs with larger scales are more susceptible to noisy, incomplete subgraphs.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"312 ","pages":"Article 113060"},"PeriodicalIF":7.2,"publicationDate":"2025-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143429897","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}

引用次数: 0

LLM Knowledge-Driven Target Prototype Learning for Few-Shot Segmentation

IF 7.2 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Knowledge-Based Systems

Pub Date : 2025-02-14 DOI: 10.1016/j.knosys.2025.113149

Pengfang Li , Fang Liu , Licheng Jiao , Shuo Li , Xu Liu , Puhua Chen , Lingling Li , Zehua Hao

Few-Shot Segmentation (FSS) aims to segment new class objects in a query image with few support images. The prototype-based FSS methods first model a target prototype and then match it with the query feature for segmentation. Recent research has focused on mining visual features to model the prototype. However, modeling the target prototype using visual features alone is not sufficient to represent target objects due to appearance differences between targets in support and query images. To address this limitation, based on the generalizable knowledge implied in the Large Language Model (LLM), we propose an LLM Knowledge-Driven Target Prototype Learning method (KD-TPL) to learn a robust prototype for the target object in the query image. Specifically, a knowledge-driven semantic prior generator is constructed to mine semantic priors in the query image applied to LLM knowledge. Based on the modeled semantic priors, a knowledge-driven hybrid prototype learner is designed to learn a representative target prototype. A knowledge-driven query feature enhancer is developed to enhance the semantics of the query feature. Finally, competitive comparison and ablation experimental results on COCO-

2 0^{i}

and PASCAL-

5^{i}

demonstrate the effectiveness of our method.

{"title":"LLM Knowledge-Driven Target Prototype Learning for Few-Shot Segmentation","authors":"Pengfang Li , Fang Liu , Licheng Jiao , Shuo Li , Xu Liu , Puhua Chen , Lingling Li , Zehua Hao","doi":"10.1016/j.knosys.2025.113149","DOIUrl":"10.1016/j.knosys.2025.113149","url":null,"abstract":"<div><div>Few-Shot Segmentation (FSS) aims to segment new class objects in a query image with few support images. The prototype-based FSS methods first model a target prototype and then match it with the query feature for segmentation. Recent research has focused on mining visual features to model the prototype. However, modeling the target prototype using visual features alone is not sufficient to represent target objects due to appearance differences between targets in support and query images. To address this limitation, based on the generalizable knowledge implied in the Large Language Model (LLM), we propose an LLM Knowledge-Driven Target Prototype Learning method (KD-TPL) to learn a robust prototype for the target object in the query image. Specifically, a knowledge-driven semantic prior generator is constructed to mine semantic priors in the query image applied to LLM knowledge. Based on the modeled semantic priors, a knowledge-driven hybrid prototype learner is designed to learn a representative target prototype. A knowledge-driven query feature enhancer is developed to enhance the semantics of the query feature. Finally, competitive comparison and ablation experimental results on COCO-<span><math><mrow><mn>2</mn><msup><mrow><mn>0</mn></mrow><mrow><mi>i</mi></mrow></msup></mrow></math></span> and PASCAL-<span><math><msup><mrow><mn>5</mn></mrow><mrow><mi>i</mi></mrow></msup></math></span> demonstrate the effectiveness of our method.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"312 ","pages":"Article 113149"},"PeriodicalIF":7.2,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143436940","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

TVC Former: A transformer-based long-term multivariate time series forecasting method using time-variable coupling correlation graph

IF 7.2 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Knowledge-Based Systems

Pub Date : 2025-02-14 DOI: 10.1016/j.knosys.2025.113147

Zhenyu Liu , Yuan Feng , Hui Liu , Ruining Tang , Bo Yang , Donghao Zhang , Weiqiang Jia , Jianrong Tan

Long-term multivariate time series forecasting is crucial in various domains that require the effective modeling of intervariable dependencies in series data. However, existing methods tend to capture these dependencies directly across the entire series, thus neglecting the local dynamic characteristics of the intervariable correlation patterns caused by locality differences and the dynamic variability of the series. To address this, we propose TVC Former, a forecasting model that uses a time-variable coupling correlation graph (TVC graph). The TVC graph treats local window-level subsequences as nodes and explicitly models local intervariable dependence. Its structure is dynamically and adaptively generated to effectively represent task-specific valuable intervariable local correlation patterns while eliminating irrelevant ones. Specifically, a sparsified graph structure is initialized based on the correlation between the input historical series and statistical similarity between the subsequences. It is then optimized using a pattern capture-fusion sparsification unit with learnable parameters. In addition, we propose a time-variable joint-encoding framework with a transformer encoder as the backbone. By introducing local head markers and a graph neural network, the framework effectively captures the intervariable dependencies using the TVC graph. Experiments on seven real-world datasets demonstrate the superiority of TVC Former in long-term forecasting tasks.

{"title":"TVC Former: A transformer-based long-term multivariate time series forecasting method using time-variable coupling correlation graph","authors":"Zhenyu Liu , Yuan Feng , Hui Liu , Ruining Tang , Bo Yang , Donghao Zhang , Weiqiang Jia , Jianrong Tan","doi":"10.1016/j.knosys.2025.113147","DOIUrl":"10.1016/j.knosys.2025.113147","url":null,"abstract":"<div><div>Long-term multivariate time series forecasting is crucial in various domains that require the effective modeling of intervariable dependencies in series data. However, existing methods tend to capture these dependencies directly across the entire series, thus neglecting the local dynamic characteristics of the intervariable correlation patterns caused by locality differences and the dynamic variability of the series. To address this, we propose TVC Former, a forecasting model that uses a time-variable coupling correlation graph (TVC graph). The TVC graph treats local window-level subsequences as nodes and explicitly models local intervariable dependence. Its structure is dynamically and adaptively generated to effectively represent task-specific valuable intervariable local correlation patterns while eliminating irrelevant ones. Specifically, a sparsified graph structure is initialized based on the correlation between the input historical series and statistical similarity between the subsequences. It is then optimized using a pattern capture-fusion sparsification unit with learnable parameters. In addition, we propose a time-variable joint-encoding framework with a transformer encoder as the backbone. By introducing local head markers and a graph neural network, the framework effectively captures the intervariable dependencies using the TVC graph. Experiments on seven real-world datasets demonstrate the superiority of TVC Former in long-term forecasting tasks.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"314 ","pages":"Article 113147"},"PeriodicalIF":7.2,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143474490","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Multi-scale representation learning for heterogeneous networks via Hawkes point processes

IF 7.2 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Knowledge-Based Systems

Pub Date : 2025-02-14 DOI: 10.1016/j.knosys.2025.113150

Qi Li, Fan Wang

In the field of dynamic heterogeneous network representation learning, current research methods have certain limitations. These limitations are mainly observed in the manual design of meta-paths, the handling of node attribute sparsity, and the fusion of dynamic heterogeneous information. To overcome these challenges, this paper presents a multi-scale representation learning method for heterogeneous networks via Hawkes point processes called MSRL. MSRL models the self-excitation effect among historical events by integrating the Hawkes process and captures the facilitating effect of external structures on event occurrence through a ternary closure process. This study employs the integration of time series analysis with neighbourhood interaction information to automate the extraction of the node pair representation. The MSRL model treats edges as time-stamped events, which not only captures the temporal dependencies between events, but also addresses the imbalance between different node types and the challenge of information fusion from a multi-granularity perspective. In particular, the model enhances the accurate estimation of the probability of node pairs forming edges by analysing the interactions between node pairs and their neighbours, which significantly improves the accuracy of tasks such as prediction. To validate the effectiveness of the MSRL model, an extensive experimental evaluation is conducted in this paper. The experimental results show that the MSRL model outperforms existing baseline models on several benchmark datasets, demonstrating its significant advantages and potential applications in the field of dynamic heterogeneous network representation learning.

{"title":"Multi-scale representation learning for heterogeneous networks via Hawkes point processes","authors":"Qi Li, Fan Wang","doi":"10.1016/j.knosys.2025.113150","DOIUrl":"10.1016/j.knosys.2025.113150","url":null,"abstract":"<div><div>In the field of dynamic heterogeneous network representation learning, current research methods have certain limitations. These limitations are mainly observed in the manual design of meta-paths, the handling of node attribute sparsity, and the fusion of dynamic heterogeneous information. To overcome these challenges, this paper presents a multi-scale representation learning method for heterogeneous networks via Hawkes point processes called MSRL. MSRL models the self-excitation effect among historical events by integrating the Hawkes process and captures the facilitating effect of external structures on event occurrence through a ternary closure process. This study employs the integration of time series analysis with neighbourhood interaction information to automate the extraction of the node pair representation. The MSRL model treats edges as time-stamped events, which not only captures the temporal dependencies between events, but also addresses the imbalance between different node types and the challenge of information fusion from a multi-granularity perspective. In particular, the model enhances the accurate estimation of the probability of node pairs forming edges by analysing the interactions between node pairs and their neighbours, which significantly improves the accuracy of tasks such as prediction. To validate the effectiveness of the MSRL model, an extensive experimental evaluation is conducted in this paper. The experimental results show that the MSRL model outperforms existing baseline models on several benchmark datasets, demonstrating its significant advantages and potential applications in the field of dynamic heterogeneous network representation learning.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"312 ","pages":"Article 113150"},"PeriodicalIF":7.2,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143420707","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Automatically resolving conflicts between expert systems: An experimental approach using large language models and fuzzy cognitive maps from participatory modeling studies

IF 7.2 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Knowledge-Based Systems

Pub Date : 2025-02-14 DOI: 10.1016/j.knosys.2025.113151

Ryan Schuerkamp , Hannah Ahlstrom , Philippe J. Giabbanelli

A mental model is an individual’s internal representation of knowledge that enables reasoning in a given domain. Cognitive dissonance arises in a mental model when there is internal conflict, causing discomfort, which individuals seek to minimize by resolving the dissonance. Modelers frequently use fuzzy cognitive maps (FCMs) to represent mental models and perspectives on a system and facilitate reasoning. Dissonance may arise in FCMs when two individuals with conflicting mental models interact (e.g., in a hybrid agent-based model with FCMs representing individuals’ mental models). We define cognitive dissonance for FCMs and develop an algorithm to automatically resolve it by leveraging large language models (LLMs). We apply our algorithm to our real-world case studies and find our approach can successfully resolve the dissonance, suggesting LLMs can broadly resolve conflict within expert systems. Additionally, our method may identify opportunities for knowledge editing of LLMs when the dissonance cannot be satisfactorily resolved through our algorithm.

心智模式是一个人对知识的内在表征，它能够在特定领域进行推理。当心智模型中存在内部冲突时，就会产生认知失调，从而引起不适，而个体则会通过解决失调来尽量减少不适。建模者经常使用模糊认知图（FCM）来表示心智模型和对系统的看法，并促进推理。当两个具有相互冲突的心智模型的个体相互作用时（例如，在代表个体心智模型的模糊认知图的基于代理的混合模型中），模糊认知图中可能会出现不和谐。我们定义了 FCM 的认知失调，并开发了一种利用大型语言模型（LLM）自动解决认知失调的算法。我们将算法应用于实际案例研究，发现我们的方法可以成功解决认知失调问题，这表明 LLM 可以广泛解决专家系统中的冲突。此外，当我们的算法无法令人满意地解决不和谐问题时，我们的方法还可以确定对 LLM 进行知识编辑的机会。

{"title":"Automatically resolving conflicts between expert systems: An experimental approach using large language models and fuzzy cognitive maps from participatory modeling studies","authors":"Ryan Schuerkamp , Hannah Ahlstrom , Philippe J. Giabbanelli","doi":"10.1016/j.knosys.2025.113151","DOIUrl":"10.1016/j.knosys.2025.113151","url":null,"abstract":"<div><div>A mental model is an individual’s internal representation of knowledge that enables reasoning in a given domain. Cognitive dissonance arises in a mental model when there is internal conflict, causing discomfort, which individuals seek to minimize by resolving the dissonance. Modelers frequently use fuzzy cognitive maps (FCMs) to represent mental models and perspectives on a system and facilitate reasoning. Dissonance may arise in FCMs when two individuals with conflicting mental models interact (e.g., in a hybrid agent-based model with FCMs representing individuals’ mental models). We define cognitive dissonance for FCMs and develop an algorithm to automatically resolve it by leveraging large language models (LLMs). We apply our algorithm to our real-world case studies and find our approach can successfully resolve the dissonance, suggesting LLMs can broadly resolve conflict within expert systems. Additionally, our method may identify opportunities for knowledge editing of LLMs when the dissonance cannot be satisfactorily resolved through our algorithm.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"313 ","pages":"Article 113151"},"PeriodicalIF":7.2,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143444236","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Overlapping community-based malicious user detection scheme in social networks 社交网络中基于重叠社区的恶意用户检测方案

IF 7.2 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Knowledge-Based Systems

Pub Date : 2025-02-14 DOI: 10.1016/j.knosys.2025.113139

Ke Gu , Deng Yang , Wenwu Zhao , Xiong Li

Currently social networks have become an important platform for social interaction and information dissemination. However, the existence of malicious social users poses a huge security threat to social networks and their information, especially it is very difficult to detect these malicious social users in overlapping communities. In this paper, we propose a malicious user detection scheme for overlapping communities-based social networks. In our scheme, we first construct a new overlapping community detection method, which is used to determine the community core and node label update order based on node influence and node relationship strength. Then we propose a malicious user detection method for overlapping communities, which is based on the changes of node attribute and node trust. In our detection method, the change trends of community attribute and message propagation influence of a node in different overlapping communities are used to determine whether the node is malicious with its specific community. Further, related experimental results show our malicious user detection scheme is effective to detect malicious users in overlapping communities.

{"title":"Overlapping community-based malicious user detection scheme in social networks","authors":"Ke Gu , Deng Yang , Wenwu Zhao , Xiong Li","doi":"10.1016/j.knosys.2025.113139","DOIUrl":"10.1016/j.knosys.2025.113139","url":null,"abstract":"<div><div>Currently social networks have become an important platform for social interaction and information dissemination. However, the existence of malicious social users poses a huge security threat to social networks and their information, especially it is very difficult to detect these malicious social users in overlapping communities. In this paper, we propose a malicious user detection scheme for overlapping communities-based social networks. In our scheme, we first construct a new overlapping community detection method, which is used to determine the community core and node label update order based on node influence and node relationship strength. Then we propose a malicious user detection method for overlapping communities, which is based on the changes of node attribute and node trust. In our detection method, the change trends of community attribute and message propagation influence of a node in different overlapping communities are used to determine whether the node is malicious with its specific community. Further, related experimental results show our malicious user detection scheme is effective to detect malicious users in overlapping communities.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"312 ","pages":"Article 113139"},"PeriodicalIF":7.2,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143420710","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A modified single-objective genetic algorithm for solving the rural postman problem with load-dependent costs

IF 7.2 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Knowledge-Based Systems

Pub Date : 2025-02-14 DOI: 10.1016/j.knosys.2025.113146

David De Santis , Mercedes Landete , Xavier Cabezas , José María Sanchis , Juanjo Peiró

This study addresses the rural postman problem with load-dependent costs, a variant of the arc routing problem where the traversal cost of an edge depends on its length and the vehicle’s load. The objective is to find a minimum-cost tour that services all required edges, a problem of particular importance when the demand weight is significant compared to the vehicle’s curb weight. We present an integer linear programming model for the problem and propose a heuristic algorithm based on bio-inspired methodologies to efficiently obtain near-optimal solutions within short computing times. The effectiveness of the approach is demonstrated through computational experiments on benchmark instances, and the results highlight the practicality of the proposed methods.

引用次数: 0

When fractional calculus meets robust learning: Adaptive robust loss functions

IF 7.2 1区计算机科学 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Knowledge-Based Systems

Pub Date : 2025-02-13 DOI: 10.1016/j.knosys.2025.113136

Mert Can Kurucu , Müjde Güzelkaya , Ibrahim Eksin , Tufan Kumbasar

In deep learning, robust loss functions are crucial for addressing challenges like outliers and noise. This paper introduces a novel family of adaptive robust loss functions, Fractional Loss Functions (FLFs), generated by deploying the fractional derivative operator into conventional ones. We demonstrate that adjusting the fractional derivative order

α

allows generating a diverse spectrum of FLFs while preserving the essential properties necessary for gradient-based learning. We show that tuning

α

gives the unique property to morph the loss landscape to reduce the influence of large residuals. Thus,

α

serves as an interpretable hyperparameter defining the robustness level of FLFs. However, determining

α

prior to training requires a manual exploration to pinpoint an FLF that aligns with the learning tasks. To overcome this issue, we reveal that FLFs can balance robustness against outliers while increasing penalization of inliers by tuning

α

. This inherent feature allows transforming

α

to an adaptive parameter as a trade-off that ensures balanced learning of

α

is feasible. Thus, FLFs can dynamically adapt their loss landscape, facilitating error minimization while providing robustness during training. We performed experiments across diverse tasks and showed that FLFs significantly enhanced performance. Our source code is available at https://github.com/mertcankurucu/Fractional-Loss-Functions.

{"title":"When fractional calculus meets robust learning: Adaptive robust loss functions","authors":"Mert Can Kurucu , Müjde Güzelkaya , Ibrahim Eksin , Tufan Kumbasar","doi":"10.1016/j.knosys.2025.113136","DOIUrl":"10.1016/j.knosys.2025.113136","url":null,"abstract":"<div><div>In deep learning, robust loss functions are crucial for addressing challenges like outliers and noise. This paper introduces a novel family of adaptive robust loss functions, Fractional Loss Functions (FLFs), generated by deploying the fractional derivative operator into conventional ones. We demonstrate that adjusting the fractional derivative order <span><math><mi>α</mi></math></span> allows generating a diverse spectrum of FLFs while preserving the essential properties necessary for gradient-based learning. We show that tuning <span><math><mi>α</mi></math></span> gives the unique property to morph the loss landscape to reduce the influence of large residuals. Thus, <span><math><mi>α</mi></math></span> serves as an interpretable hyperparameter defining the robustness level of FLFs. However, determining <span><math><mi>α</mi></math></span> prior to training requires a manual exploration to pinpoint an FLF that aligns with the learning tasks. To overcome this issue, we reveal that FLFs can balance robustness against outliers while increasing penalization of inliers by tuning <span><math><mi>α</mi></math></span>. This inherent feature allows transforming <span><math><mi>α</mi></math></span> to an adaptive parameter as a trade-off that ensures balanced learning of <span><math><mi>α</mi></math></span> is feasible. Thus, FLFs can dynamically adapt their loss landscape, facilitating error minimization while providing robustness during training. We performed experiments across diverse tasks and showed that FLFs significantly enhanced performance. Our source code is available at <span><span>https://github.com/mertcankurucu/Fractional-Loss-Functions</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":49939,"journal":{"name":"Knowledge-Based Systems","volume":"312 ","pages":"Article 113136"},"PeriodicalIF":7.2,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143420703","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

首页上一页

下一页尾页

类型

全部化学•材料生命科学医学物理工程技术环境•农林材料科学地球科学法学管理学化学环境科学与生态学计算机科学教育学经济学农林科学人文科学生物学数学物理与天体物理心理学综合性期刊其他工业工程理学历史学农学文学信息工程

数据库

全部 ACS Publications Elsevier ieeexplore Springer The Royal Society of Chemistry Wiley

期刊

Knowledge-Based Systems

全部 Acc. Chem. Res. ACS Applied Bio Materials ACS Appl. Electron. Mater. ACS Appl. Energy Mater. ACS Appl. Mater. Interfaces ACS Appl. Nano Mater. ACS Appl. Polym. Mater. ACS BIOMATER-SCI ENG ACS Catal. ACS Cent. Sci. ACS Chem. Biol. ACS Chemical Health & Safety ACS Chem. Neurosci. ACS Comb. Sci. ACS Earth Space Chem. ACS Energy Lett. ACS Infect. Dis. ACS Macro Lett. ACS Mater. Lett. ACS Med. Chem. Lett. ACS Nano ACS Omega ACS Photonics ACS Sens. ACS Sustainable Chem. Eng. ACS Synth. Biol. Anal. Chem. BIOCHEMISTRY-US Bioconjugate Chem. BIOMACROMOLECULES Chem. Res. Toxicol. Chem. Rev. Chem. Mater. CRYST GROWTH DES ENERG FUEL Environ. Sci. Technol. Environ. Sci. Technol. Lett. Eur. J. Inorg. Chem. IND ENG CHEM RES Inorg. Chem. J. Agric. Food. Chem. J. Chem. Eng. Data J. Chem. Educ. J. Chem. Inf. Model. J. Chem. Theory Comput. J. Med. Chem. J. Nat. Prod. J PROTEOME RES J. Am. Chem. Soc. LANGMUIR MACROMOLECULES Mol. Pharmaceutics Nano Lett. Org. Lett. ORG PROCESS RES DEV ORGANOMETALLICS J. Org. Chem. J. Phys. Chem. J. Phys. Chem. A J. Phys. Chem. B J. Phys. Chem. C J. Phys. Chem. Lett. Analyst Anal. Methods Biomater. Sci. Catal. Sci. Technol. Chem. Commun. Chem. Soc. Rev. CHEM EDUC RES PRACT CRYSTENGCOMM Dalton Trans. Energy Environ. Sci. ENVIRON SCI-NANO ENVIRON SCI-PROC IMP ENVIRON SCI-WAT RES Faraday Discuss. Food Funct. Green Chem. Inorg. Chem. Front. Integr. Biol. J. Anal. At. Spectrom. J. Mater. Chem. A J. Mater. Chem. B J. Mater. Chem. C Lab Chip Mater. Chem. Front. Mater. Horiz. MEDCHEMCOMM Metallomics Mol. Biosyst. Mol. Syst. Des. Eng. Nanoscale Nanoscale Horiz. Nat. Prod. Rep. New J. Chem. Org. Biomol. Chem. Org. Chem. Front. PHOTOCH PHOTOBIO SCI PCCP Polym. Chem.

﹀