Pub Date : 2023-06-01DOI: 10.1109/COMPSAC57700.2023.00072
Wesley B. R. Herculano, Everton L. G. Alves, Melina Mongiovi
Maintenance tasks often rely on failing test cases, highlighting the importance of well-designed tests for their success. While automatically generated tests can provide higher code coverage and detect faults, it is unclear whether they can be effective in guiding maintenance tasks or if developers fully accept them. In our recent work, we presented the results of a series of empirical studies that evaluated the practical support of generated tests. Our studies with 126 developers showed that automatically generated tests can effectively identify faults during maintenance tasks. Developers were equally effective in creating bug fixes when using manually-written, Evosuite, and Randoop tests. However, developers perceived generated tests as not well-designed and preferred refactored versions of Randoop tests. We plan to enhance Evosuite tests and propose an approach/tool that assesses the quality of generated tests and automatically enhances them. Our research may impact the design and use of generated tests in the context of maintenance tasks.
{"title":"Assessing and Improving the Quality of Generated Tests in the Context of Maintenance Tasks","authors":"Wesley B. R. Herculano, Everton L. G. Alves, Melina Mongiovi","doi":"10.1109/COMPSAC57700.2023.00072","DOIUrl":"https://doi.org/10.1109/COMPSAC57700.2023.00072","url":null,"abstract":"Maintenance tasks often rely on failing test cases, highlighting the importance of well-designed tests for their success. While automatically generated tests can provide higher code coverage and detect faults, it is unclear whether they can be effective in guiding maintenance tasks or if developers fully accept them. In our recent work, we presented the results of a series of empirical studies that evaluated the practical support of generated tests. Our studies with 126 developers showed that automatically generated tests can effectively identify faults during maintenance tasks. Developers were equally effective in creating bug fixes when using manually-written, Evosuite, and Randoop tests. However, developers perceived generated tests as not well-designed and preferred refactored versions of Randoop tests. We plan to enhance Evosuite tests and propose an approach/tool that assesses the quality of generated tests and automatically enhances them. Our research may impact the design and use of generated tests in the context of maintenance tasks.","PeriodicalId":296288,"journal":{"name":"2023 IEEE 47th Annual Computers, Software, and Applications Conference (COMPSAC)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121709928","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.1109/COMPSAC57700.2023.00213
Martin Brown, Md Abdullah Khan, Dominic Thomas, Yong Pei, M. Nandan
Early detection of and intervention in behavioral health cases, including mental health, is crucial to prevent harm to one’s self and others. Police reports generated by officers on duty or in response to 911 calls remain an untapped resource for identifying such incidents. To expedite the detection process, we propose a workflow that involves collaboration between experts to manually annotate cases and correct model predictions. This approach can improve both initial annotation and model performance. Therefore, we advocate for the incorporation of manual annotations from experts, natural language processing (NLP), active learning, and advanced machine learning techniques to detect behavioral health cases within police reports. The experimentation suggests that a CNN-LSTM model achieves the best performance with an accuracy of 86.67% and an F1-score of 0.82 in detecting behavioral health issues.
{"title":"Detection of Behavioral Health Cases from Sensitive Police Officer Narratives","authors":"Martin Brown, Md Abdullah Khan, Dominic Thomas, Yong Pei, M. Nandan","doi":"10.1109/COMPSAC57700.2023.00213","DOIUrl":"https://doi.org/10.1109/COMPSAC57700.2023.00213","url":null,"abstract":"Early detection of and intervention in behavioral health cases, including mental health, is crucial to prevent harm to one’s self and others. Police reports generated by officers on duty or in response to 911 calls remain an untapped resource for identifying such incidents. To expedite the detection process, we propose a workflow that involves collaboration between experts to manually annotate cases and correct model predictions. This approach can improve both initial annotation and model performance. Therefore, we advocate for the incorporation of manual annotations from experts, natural language processing (NLP), active learning, and advanced machine learning techniques to detect behavioral health cases within police reports. The experimentation suggests that a CNN-LSTM model achieves the best performance with an accuracy of 86.67% and an F1-score of 0.82 in detecting behavioral health issues.","PeriodicalId":296288,"journal":{"name":"2023 IEEE 47th Annual Computers, Software, and Applications Conference (COMPSAC)","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122376623","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.1109/COMPSAC57700.2023.00053
Chi Mai Nguyen, Phat Thai, Duy Khang Lam, Van Tuan Nguyen
We live in an age of information overload. Manual information processing is increasingly overwhelmed with the enormous amount of information created by the explosive growth of news portals and online social networks. Such a situation calls for an automatic system that can support the process of handling, analyzing, and filtering information, especially information from online sources. In this work, we proposed a text analysis system that automatically collects, extracts, and analyses information from public-source-text documents such as news portals and social media networks. The proposed system can handle both long and short-text documents. It also has real-time features and is not restricted by any input data domain. The system can be used in different domains, such as scientific research, marketing, and security-related domains. Moreover, the system is engineered in modules and is flexible. Each module is an independent micro-service that can be used as a separate standalone application. The system is also extensible since new modules can be added easily. Index Terms—text analysis system, data mining, natural language processing
{"title":"A Real-Time Text Analysis System","authors":"Chi Mai Nguyen, Phat Thai, Duy Khang Lam, Van Tuan Nguyen","doi":"10.1109/COMPSAC57700.2023.00053","DOIUrl":"https://doi.org/10.1109/COMPSAC57700.2023.00053","url":null,"abstract":"We live in an age of information overload. Manual information processing is increasingly overwhelmed with the enormous amount of information created by the explosive growth of news portals and online social networks. Such a situation calls for an automatic system that can support the process of handling, analyzing, and filtering information, especially information from online sources. In this work, we proposed a text analysis system that automatically collects, extracts, and analyses information from public-source-text documents such as news portals and social media networks. The proposed system can handle both long and short-text documents. It also has real-time features and is not restricted by any input data domain. The system can be used in different domains, such as scientific research, marketing, and security-related domains. Moreover, the system is engineered in modules and is flexible. Each module is an independent micro-service that can be used as a separate standalone application. The system is also extensible since new modules can be added easily. Index Terms—text analysis system, data mining, natural language processing","PeriodicalId":296288,"journal":{"name":"2023 IEEE 47th Annual Computers, Software, and Applications Conference (COMPSAC)","volume":"70 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130055860","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.1109/COMPSAC57700.2023.00284
Md Mostafizur Rahman, Aiasha Siddika Arshi, Md. Golam Moula Mehedi Hasan, Sumayia Farzana Mishu, H. Shahriar, Fan Wu
This survey paper provides an overview of the current state of AI attacks and risks for AI security and privacy as artificial intelligence becomes more prevalent in various applications and services. The risks associated with AI attacks and security breaches are becoming increasingly apparent and cause many financial and social losses. This paper will categorize the different types of attacks on AI models, including adversarial attacks, model inversion attacks, poisoning attacks, data poisoning attacks, data extraction attacks, and membership inference attacks. The paper also emphasizes the importance of developing secure and robust AI models to ensure the privacy and security of sensitive data. Through a systematic literature review, this survey paper comprehensively analyzes the current state of AI attacks and risks for AI security and privacy and detection techniques.
{"title":"Security Risk and Attacks in AI: A Survey of Security and Privacy","authors":"Md Mostafizur Rahman, Aiasha Siddika Arshi, Md. Golam Moula Mehedi Hasan, Sumayia Farzana Mishu, H. Shahriar, Fan Wu","doi":"10.1109/COMPSAC57700.2023.00284","DOIUrl":"https://doi.org/10.1109/COMPSAC57700.2023.00284","url":null,"abstract":"This survey paper provides an overview of the current state of AI attacks and risks for AI security and privacy as artificial intelligence becomes more prevalent in various applications and services. The risks associated with AI attacks and security breaches are becoming increasingly apparent and cause many financial and social losses. This paper will categorize the different types of attacks on AI models, including adversarial attacks, model inversion attacks, poisoning attacks, data poisoning attacks, data extraction attacks, and membership inference attacks. The paper also emphasizes the importance of developing secure and robust AI models to ensure the privacy and security of sensitive data. Through a systematic literature review, this survey paper comprehensively analyzes the current state of AI attacks and risks for AI security and privacy and detection techniques.","PeriodicalId":296288,"journal":{"name":"2023 IEEE 47th Annual Computers, Software, and Applications Conference (COMPSAC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131039387","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.1109/COMPSAC57700.2023.00204
Joylal Das, Sulalitha Bowala, R. Thulasiram, A. Thavaneswaran
Constructing resilient portfolios is of crucial and utmost importance to investment management. This study compares traditional and data-driven models for building resilient portfolios and analyzes their performance for stocks (S&P 500) and highly volatile cryptocurrency markets. The study investigates the performance of traditional models, such as mean-variance and constrained optimization, and a recently proposed data-driven resilient portfolio optimization model for stocks. Moreover, the study analyzes these methods with evolving S&P CME bitcoin futures index and the Crypto20 index. These analyses highlight the need for further investigation into traditional and data-driven approaches for resilient portfolio optimization, including higher-order moments, particularly under varying market conditions. This study provides valuable insights for investors and portfolio managers aiming to build resilient portfolios that could be used in different market environments.
{"title":"Resilient Portfolio Optimization using Traditional and Data-Driven Models for Cryptocurrencies and Stocks","authors":"Joylal Das, Sulalitha Bowala, R. Thulasiram, A. Thavaneswaran","doi":"10.1109/COMPSAC57700.2023.00204","DOIUrl":"https://doi.org/10.1109/COMPSAC57700.2023.00204","url":null,"abstract":"Constructing resilient portfolios is of crucial and utmost importance to investment management. This study compares traditional and data-driven models for building resilient portfolios and analyzes their performance for stocks (S&P 500) and highly volatile cryptocurrency markets. The study investigates the performance of traditional models, such as mean-variance and constrained optimization, and a recently proposed data-driven resilient portfolio optimization model for stocks. Moreover, the study analyzes these methods with evolving S&P CME bitcoin futures index and the Crypto20 index. These analyses highlight the need for further investigation into traditional and data-driven approaches for resilient portfolio optimization, including higher-order moments, particularly under varying market conditions. This study provides valuable insights for investors and portfolio managers aiming to build resilient portfolios that could be used in different market environments.","PeriodicalId":296288,"journal":{"name":"2023 IEEE 47th Annual Computers, Software, and Applications Conference (COMPSAC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128612196","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.1109/COMPSAC57700.2023.00018
Anoop Bhagyanath, K. Schneider
In traditional von Neumann processors, the central register file is an inherent limiting factor in exploiting the instruction-level parallelism (ILP) of programs. To alleviate this problem, many processors follow a hybrid von Neumann/dataflow computing model in which specific instruction sequences are executed in dataflow order by communicating intermediate values directly from producer processing units (PUs) to consumer PUs without using a central register file. However, the intermediate values often reside in local registers of the PUs, which requires a synchronization of the data transports that still limits the exploitation of the ILP.To avoid the use of a central register file and the need for any synchronization between PUs, some newer architectures suggest first-in-first-out (FIFO) buffers instead of local registers at the input and output ports of the PUs. Since values are produced and consumed, and are thus never overwritten (as in registers), the compiler must determine the required number of copies of each value. Furthermore, it is necessary to control the number of copies of values to develop buffer size aware compilation methods. However, the number of variable uses in a sequential program may depend on the future execution. This paper presents transformations for ‘balancing’ a given program, i.e., transforming the program so that for all points in the program, the number of future uses of all variables can be accurately determined in order to allocate the required buffer sizes in the later compilation phases. The classical space-time trade-off is demonstrated by the experimental results which show an improvement of the processor performance with increasing buffer sizes and vice versa. More importantly, the experimental results demonstrate the potential of buffered hybrid dataflow architectures for a scalable use of ILP.
{"title":"Program Balancing in Compilation for Buffered Hybrid Dataflow Processors","authors":"Anoop Bhagyanath, K. Schneider","doi":"10.1109/COMPSAC57700.2023.00018","DOIUrl":"https://doi.org/10.1109/COMPSAC57700.2023.00018","url":null,"abstract":"In traditional von Neumann processors, the central register file is an inherent limiting factor in exploiting the instruction-level parallelism (ILP) of programs. To alleviate this problem, many processors follow a hybrid von Neumann/dataflow computing model in which specific instruction sequences are executed in dataflow order by communicating intermediate values directly from producer processing units (PUs) to consumer PUs without using a central register file. However, the intermediate values often reside in local registers of the PUs, which requires a synchronization of the data transports that still limits the exploitation of the ILP.To avoid the use of a central register file and the need for any synchronization between PUs, some newer architectures suggest first-in-first-out (FIFO) buffers instead of local registers at the input and output ports of the PUs. Since values are produced and consumed, and are thus never overwritten (as in registers), the compiler must determine the required number of copies of each value. Furthermore, it is necessary to control the number of copies of values to develop buffer size aware compilation methods. However, the number of variable uses in a sequential program may depend on the future execution. This paper presents transformations for ‘balancing’ a given program, i.e., transforming the program so that for all points in the program, the number of future uses of all variables can be accurately determined in order to allocate the required buffer sizes in the later compilation phases. The classical space-time trade-off is demonstrated by the experimental results which show an improvement of the processor performance with increasing buffer sizes and vice versa. More importantly, the experimental results demonstrate the potential of buffered hybrid dataflow architectures for a scalable use of ILP.","PeriodicalId":296288,"journal":{"name":"2023 IEEE 47th Annual Computers, Software, and Applications Conference (COMPSAC)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126970702","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.1109/COMPSAC57700.2023.00134
S. Reddivari, James K. Orr, Reyansh Reddy
Blockchain-oriented software (BOS) is emerging rapidly, however, tailored software engineering techniques to assist software testing is one of the important and necessary phases of software development that plays a vital role in software quality assurance. Blockchain-oriented software testing (BOST) is another area that is still emerging. This paper conducts a preliminary literature review to assess the type of research published, where this research can be found, and the main themes of existing research in BOST. This study also introduces a framework for conducting an empirical study on BOS available on GitHub and proposes future work to expand the knowledge in this domain.
{"title":"Blockchain-Oriented Software Testing: A Preliminary Literature Review","authors":"S. Reddivari, James K. Orr, Reyansh Reddy","doi":"10.1109/COMPSAC57700.2023.00134","DOIUrl":"https://doi.org/10.1109/COMPSAC57700.2023.00134","url":null,"abstract":"Blockchain-oriented software (BOS) is emerging rapidly, however, tailored software engineering techniques to assist software testing is one of the important and necessary phases of software development that plays a vital role in software quality assurance. Blockchain-oriented software testing (BOST) is another area that is still emerging. This paper conducts a preliminary literature review to assess the type of research published, where this research can be found, and the main themes of existing research in BOST. This study also introduces a framework for conducting an empirical study on BOS available on GitHub and proposes future work to expand the knowledge in this domain.","PeriodicalId":296288,"journal":{"name":"2023 IEEE 47th Annual Computers, Software, and Applications Conference (COMPSAC)","volume":"115 4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114262782","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.1109/COMPSAC57700.2023.00017
Niccolò De Bona, L. Santoro, D. Brunelli, D. Fontanelli
Multi-robot patrolling for dynamic coverage in flat environments is proposed, through a systematic simulative analysis between the Greedy Bayesian Strategy and the Expected Reactive algorithm based on the expected idleness. The two approaches are compared against unreliable communications, communication and sensing range, and number of conflicts. In addition, we introduce a new weighting-term for the regions close to a quantity of interest detected by robots, decreasing the passing-time for those regions. Combining the proposed control strategy and a traditional distributed and recursive Weighted Least Square estimation algorithm, the swarm is capable to compute the quantity of interest position with a desired target uncertainty. Extensive simulations and comparisons are reported.
{"title":"Adaptive Expected Reactive algorithm for Heterogeneous Patrolling Systems based on Target Uncertainty","authors":"Niccolò De Bona, L. Santoro, D. Brunelli, D. Fontanelli","doi":"10.1109/COMPSAC57700.2023.00017","DOIUrl":"https://doi.org/10.1109/COMPSAC57700.2023.00017","url":null,"abstract":"Multi-robot patrolling for dynamic coverage in flat environments is proposed, through a systematic simulative analysis between the Greedy Bayesian Strategy and the Expected Reactive algorithm based on the expected idleness. The two approaches are compared against unreliable communications, communication and sensing range, and number of conflicts. In addition, we introduce a new weighting-term for the regions close to a quantity of interest detected by robots, decreasing the passing-time for those regions. Combining the proposed control strategy and a traditional distributed and recursive Weighted Least Square estimation algorithm, the swarm is capable to compute the quantity of interest position with a desired target uncertainty. Extensive simulations and comparisons are reported.","PeriodicalId":296288,"journal":{"name":"2023 IEEE 47th Annual Computers, Software, and Applications Conference (COMPSAC)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114408011","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.1109/COMPSAC57700.2023.00139
Yong Jin, M. Tomoishi
Authentication and authorization are essential functionalities in Internet of Things (IoT) systems in the areas of device monitoring, access control, data sharing, and privacy preservation. The current authentication and authorization technologies are mainly based on ID/Password (credential), public key cryptography collaborating with multi-factor approaches, and manual basis static access control for data resources. However, they have disadvantages in terms of identity/credential management, privacy preservation, and the mandatory requirement of Trusted Third Party (TTP) Certificate Authorities (which are not cryptographically secure) for certificate verifications. The objective of this fast abstract is to present a novel idea for named-entity-based TTP-free authentication and authorization architecture in IoT systems. The proposed architecture is based on the collaboration of blockchain technology and DNS-based Authentication of Named Entity (DANE) protocol to provide secure and privacy-preserved authentication and authorization functionalities with considering effective device identity management and secure data access control in IoT systems.
{"title":"A Named-Entity-based TTP-free Authentication and Authorization Architecture for IoT Systems","authors":"Yong Jin, M. Tomoishi","doi":"10.1109/COMPSAC57700.2023.00139","DOIUrl":"https://doi.org/10.1109/COMPSAC57700.2023.00139","url":null,"abstract":"Authentication and authorization are essential functionalities in Internet of Things (IoT) systems in the areas of device monitoring, access control, data sharing, and privacy preservation. The current authentication and authorization technologies are mainly based on ID/Password (credential), public key cryptography collaborating with multi-factor approaches, and manual basis static access control for data resources. However, they have disadvantages in terms of identity/credential management, privacy preservation, and the mandatory requirement of Trusted Third Party (TTP) Certificate Authorities (which are not cryptographically secure) for certificate verifications. The objective of this fast abstract is to present a novel idea for named-entity-based TTP-free authentication and authorization architecture in IoT systems. The proposed architecture is based on the collaboration of blockchain technology and DNS-based Authentication of Named Entity (DANE) protocol to provide secure and privacy-preserved authentication and authorization functionalities with considering effective device identity management and secure data access control in IoT systems.","PeriodicalId":296288,"journal":{"name":"2023 IEEE 47th Annual Computers, Software, and Applications Conference (COMPSAC)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122838729","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-01DOI: 10.1109/COMPSAC57700.2023.00045
D.P.H. Napolitano, Luca Cagliero
Shapley Values are established concepts used to explain local and global contribution of individual features to the prediction of AI models. Currently, global Shapley-based explainers do not consider the co-occurrences of feature-value pairs in the analyzed data. This paper proposes a novel approach to leverage the High-Utility Itemset Mining framework to jointly consider Shapley-based feature-level contributions and feature-value pair co-occurrences. The results achieved on benchmark datasets show that the extracted patterns provide actionable knowledge, complementary to those of global Shapley Values.
{"title":"GX-HUI: Global Explanations of AI Models based on High-Utility Itemsets","authors":"D.P.H. Napolitano, Luca Cagliero","doi":"10.1109/COMPSAC57700.2023.00045","DOIUrl":"https://doi.org/10.1109/COMPSAC57700.2023.00045","url":null,"abstract":"Shapley Values are established concepts used to explain local and global contribution of individual features to the prediction of AI models. Currently, global Shapley-based explainers do not consider the co-occurrences of feature-value pairs in the analyzed data. This paper proposes a novel approach to leverage the High-Utility Itemset Mining framework to jointly consider Shapley-based feature-level contributions and feature-value pair co-occurrences. The results achieved on benchmark datasets show that the extracted patterns provide actionable knowledge, complementary to those of global Shapley Values.","PeriodicalId":296288,"journal":{"name":"2023 IEEE 47th Annual Computers, Software, and Applications Conference (COMPSAC)","volume":"359 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122754129","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：