Pub Date : 2021-07-21DOI: 10.1109/INDIN45523.2021.9557445
A. Pakonen
For over a decade, model checking has been successfully used to formally verify the instrumentation and control (I&C) logic design in Finnish nuclear power plant projects. One of the practical challenges is that the model checker NuSMV forces the user to abstract the way analog signals are processed in the model, which causes extra manual work, and could mask actual design issues. In this paper, we experiment with the newer tool nuXmv, which supports infinite-state modelling. Using actual models from practical industrial projects, we show that after changing the analog signal processing to be based on real number math, the analysis times are still manageable. The disadvantage is that certain useful types of formal properties are not supported by the infinite-state algorithms. We also discuss the nuclear industry specific features of I&C programming languages, which cause significant constraints on domain-specific formal verification method and tool development.
{"title":"Model-checking infinite-state nuclear safety I&C systems with nuXmv","authors":"A. Pakonen","doi":"10.1109/INDIN45523.2021.9557445","DOIUrl":"https://doi.org/10.1109/INDIN45523.2021.9557445","url":null,"abstract":"For over a decade, model checking has been successfully used to formally verify the instrumentation and control (I&C) logic design in Finnish nuclear power plant projects. One of the practical challenges is that the model checker NuSMV forces the user to abstract the way analog signals are processed in the model, which causes extra manual work, and could mask actual design issues. In this paper, we experiment with the newer tool nuXmv, which supports infinite-state modelling. Using actual models from practical industrial projects, we show that after changing the analog signal processing to be based on real number math, the analysis times are still manageable. The disadvantage is that certain useful types of formal properties are not supported by the infinite-state algorithms. We also discuss the nuclear industry specific features of I&C programming languages, which cause significant constraints on domain-specific formal verification method and tool development.","PeriodicalId":370921,"journal":{"name":"2021 IEEE 19th International Conference on Industrial Informatics (INDIN)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116508749","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 : 2021-07-21DOI: 10.1109/INDIN45523.2021.9557406
J. Zinn, B. Vogel‐Heuser, Fabian Schuhmann, Luis Alberto Cruz Salazar
The training of Deep Reinforcement Learning algorithms on robotic devices is challenging due to their large number of actuators and limited number of feasible action sequences. This paper addresses this challenge by extending and transferring existing approaches for waypoint-based exploration with Hierarchical Reinforcement Learning to the domain of robotic devices. The resulting algorithm utilizes a top-level policy, which suggests waypoints to a bottom-level policy that controls the system actuators. The waypoints can either be provided to the top-level policy as domain knowledge or be learned from scratch. The algorithm explicitly accounts for the low number of feasible waypoints and waypoint transitions that are characteristic of robotic devices. The effectiveness of the approach is evaluated on the simulation of a research demonstrator, and a separate ablation study proves the importance of its components.
{"title":"Hierarchical Reinforcement Learning for Waypoint-based Exploration in Robotic Devices","authors":"J. Zinn, B. Vogel‐Heuser, Fabian Schuhmann, Luis Alberto Cruz Salazar","doi":"10.1109/INDIN45523.2021.9557406","DOIUrl":"https://doi.org/10.1109/INDIN45523.2021.9557406","url":null,"abstract":"The training of Deep Reinforcement Learning algorithms on robotic devices is challenging due to their large number of actuators and limited number of feasible action sequences. This paper addresses this challenge by extending and transferring existing approaches for waypoint-based exploration with Hierarchical Reinforcement Learning to the domain of robotic devices. The resulting algorithm utilizes a top-level policy, which suggests waypoints to a bottom-level policy that controls the system actuators. The waypoints can either be provided to the top-level policy as domain knowledge or be learned from scratch. The algorithm explicitly accounts for the low number of feasible waypoints and waypoint transitions that are characteristic of robotic devices. The effectiveness of the approach is evaluated on the simulation of a research demonstrator, and a separate ablation study proves the importance of its components.","PeriodicalId":370921,"journal":{"name":"2021 IEEE 19th International Conference on Industrial Informatics (INDIN)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116664232","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 : 2021-07-21DOI: 10.1109/INDIN45523.2021.9557468
Tobias Striffler, H. Schotten
Deterministic communication across integrated wired and wireless networks is currently one of the big topics in research and standardization. 5G and TSN integration efforts are at the forefront of enabling the convergence of wired and wireless networks for Industry 4.0.In this paper, we investigate how synchronization and syntonization errors affect the achievable end-to-end time synchronization accuracy in integrated 5G and TSN networks. We specifically focus on the impact of the 5G System modeling a TSN transparent clock according to 3GPP Release 17.
{"title":"The 5G Transparent Clock: Synchronization Errors in Integrated 5G-TSN Industrial Networks","authors":"Tobias Striffler, H. Schotten","doi":"10.1109/INDIN45523.2021.9557468","DOIUrl":"https://doi.org/10.1109/INDIN45523.2021.9557468","url":null,"abstract":"Deterministic communication across integrated wired and wireless networks is currently one of the big topics in research and standardization. 5G and TSN integration efforts are at the forefront of enabling the convergence of wired and wireless networks for Industry 4.0.In this paper, we investigate how synchronization and syntonization errors affect the achievable end-to-end time synchronization accuracy in integrated 5G and TSN networks. We specifically focus on the impact of the 5G System modeling a TSN transparent clock according to 3GPP Release 17.","PeriodicalId":370921,"journal":{"name":"2021 IEEE 19th International Conference on Industrial Informatics (INDIN)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129880298","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 : 2021-07-21DOI: 10.1109/INDIN45523.2021.9557359
S. Sinha, P. Franciosa, D. Ceglarek
The paper proposes a novel approach, Object Shape Error Correction (OSEC), to determine corrective action in order to mitigate root cause(s) (RCs) of dimensional and geometric product shape errors. It leverages Deep Deterministic Policy Gradient (DDPG) algorithm to learn optimal process parameters update policies based on high dimensional state estimates of multi-station assembly systems (MAS). These policies can be interpreted in engineering terms as sequential corrective adjustments of process parameters that are necessary to mitigate RCs of product shape errors. The approach has the capability to estimate adjustments of process parameters related to fixturing and joining while simultaneously accounting for (i) RC uncertainty estimation, (ii) Key Performance Indicator (KPI) improvement, (iii) MAS design architecture; and, (iv) MAS inherent stochasticity. In addition, the OSEC methodology leverages a reward function parameterized by user interpretable functional coefficients for optimal tradeoff involving various corrections requirements. Benchmarking using an industrial, automotive cross-member assembly system demonstrates a 40% increase in the effectiveness of corrective actions when compared to current approaches.
{"title":"Object Shape Error Correction using Deep Reinforcement Learning for Multi-Station Assembly Systems","authors":"S. Sinha, P. Franciosa, D. Ceglarek","doi":"10.1109/INDIN45523.2021.9557359","DOIUrl":"https://doi.org/10.1109/INDIN45523.2021.9557359","url":null,"abstract":"The paper proposes a novel approach, Object Shape Error Correction (OSEC), to determine corrective action in order to mitigate root cause(s) (RCs) of dimensional and geometric product shape errors. It leverages Deep Deterministic Policy Gradient (DDPG) algorithm to learn optimal process parameters update policies based on high dimensional state estimates of multi-station assembly systems (MAS). These policies can be interpreted in engineering terms as sequential corrective adjustments of process parameters that are necessary to mitigate RCs of product shape errors. The approach has the capability to estimate adjustments of process parameters related to fixturing and joining while simultaneously accounting for (i) RC uncertainty estimation, (ii) Key Performance Indicator (KPI) improvement, (iii) MAS design architecture; and, (iv) MAS inherent stochasticity. In addition, the OSEC methodology leverages a reward function parameterized by user interpretable functional coefficients for optimal tradeoff involving various corrections requirements. Benchmarking using an industrial, automotive cross-member assembly system demonstrates a 40% increase in the effectiveness of corrective actions when compared to current approaches.","PeriodicalId":370921,"journal":{"name":"2021 IEEE 19th International Conference on Industrial Informatics (INDIN)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128622587","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 : 2021-07-21DOI: 10.1109/INDIN45523.2021.9557446
Riccardo Colelli, Chiara Foglietta, Roberto Fusacchia, S. Panzieri, F. Pascucci
Critical Infrastructures (CIs) such as power grid, water and gas distribution are controlled by Industrial Control Systems (ICS). Sensors and actuators of a physical plant are managed by the ICS. Data and commands transmitted over the network from the Programmable Logic Controllers (PLCs) are saved and parsed within the Historian. Generally, this architecture guarantees to check for any process anomalies that may occur due to component failures and cyber attacks. The other use of this data allows activities such as forensic analysis. To secure the network is also crucial to protect the communication between devices. A cyber attack on the log devices could jeopardize any forensic analysis be it for maintenance, or discovering an attack trail. In this paper is proposed a strategy to secure plant operational data recorded in the Historian and data exchange in the network. An integrity checking mechanism, in combination with blockchain, is used to ensure data integrity. Data redundancy is achieved by applying an efficient replication mechanism and enables data recovery after an attack.
{"title":"Blockchain application in simulated environment for Cyber-Physical Systems Security","authors":"Riccardo Colelli, Chiara Foglietta, Roberto Fusacchia, S. Panzieri, F. Pascucci","doi":"10.1109/INDIN45523.2021.9557446","DOIUrl":"https://doi.org/10.1109/INDIN45523.2021.9557446","url":null,"abstract":"Critical Infrastructures (CIs) such as power grid, water and gas distribution are controlled by Industrial Control Systems (ICS). Sensors and actuators of a physical plant are managed by the ICS. Data and commands transmitted over the network from the Programmable Logic Controllers (PLCs) are saved and parsed within the Historian. Generally, this architecture guarantees to check for any process anomalies that may occur due to component failures and cyber attacks. The other use of this data allows activities such as forensic analysis. To secure the network is also crucial to protect the communication between devices. A cyber attack on the log devices could jeopardize any forensic analysis be it for maintenance, or discovering an attack trail. In this paper is proposed a strategy to secure plant operational data recorded in the Historian and data exchange in the network. An integrity checking mechanism, in combination with blockchain, is used to ensure data integrity. Data redundancy is achieved by applying an efficient replication mechanism and enables data recovery after an attack.","PeriodicalId":370921,"journal":{"name":"2021 IEEE 19th International Conference on Industrial Informatics (INDIN)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129674933","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 : 2021-07-21DOI: 10.1109/INDIN45523.2021.9557429
Ricardo Silva Peres, José Barata
The advent of Industry 4.0 has made it crucial to improve the accessibility of higher education and to ensure that the future generation of engineers is able to acquire interdisciplinary competences to face the challenges of the data-driven era, including hands-on experience with real scenarios. We present an approach for teaching Cyber-Physical Production Systems remotely to graduate students, along with a discussion of sample projects, recommendations and the lessons learned in the effort to break down geographical barriers to education, reduce the cost associated with material and mitigate the impact of unforeseen disruptions such as the one caused by the pandemic scenario of COVID-19 in 2020. A combination of physical learning factory demonstrators, digital twin and simulation scenarios was developed to provide students with the resources to remotely implement an end-to-end Cyber-Physical Production System, along with its integration with other key technologies of Industry 4.0. A case study at the NOVA University of Lisbon showed that average attendance improved by 26.6%, retention in lab component improved by 12.9% and lab grades improved on average by 7.33% compared to on-site iterations of the same course in the two previous years.
{"title":"Remote E-Learning for Cyber-Physical Production Systems in Higher Education","authors":"Ricardo Silva Peres, José Barata","doi":"10.1109/INDIN45523.2021.9557429","DOIUrl":"https://doi.org/10.1109/INDIN45523.2021.9557429","url":null,"abstract":"The advent of Industry 4.0 has made it crucial to improve the accessibility of higher education and to ensure that the future generation of engineers is able to acquire interdisciplinary competences to face the challenges of the data-driven era, including hands-on experience with real scenarios. We present an approach for teaching Cyber-Physical Production Systems remotely to graduate students, along with a discussion of sample projects, recommendations and the lessons learned in the effort to break down geographical barriers to education, reduce the cost associated with material and mitigate the impact of unforeseen disruptions such as the one caused by the pandemic scenario of COVID-19 in 2020. A combination of physical learning factory demonstrators, digital twin and simulation scenarios was developed to provide students with the resources to remotely implement an end-to-end Cyber-Physical Production System, along with its integration with other key technologies of Industry 4.0. A case study at the NOVA University of Lisbon showed that average attendance improved by 26.6%, retention in lab component improved by 12.9% and lab grades improved on average by 7.33% compared to on-site iterations of the same course in the two previous years.","PeriodicalId":370921,"journal":{"name":"2021 IEEE 19th International Conference on Industrial Informatics (INDIN)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128909086","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 : 2021-07-21DOI: 10.1109/INDIN45523.2021.9557388
Dapeng Zhang, Zhiwei Gao
The great success of deep neural network (DNN) in image field stimulates its application in fault detection and diagnose. However due to the limitation of system security, it is impossible to obtain complete fault data as the training database for neural network, so that it is challenging to identify a fault that never occurred before. In this paper, an ensemble approach is proposed to adapt to a new fault by adding output branches of the neural network. Firstly, the time series are transferred to numerous imaging matrixes. The intrinsic characteristics of the matrixes are then extracted using deep neural network which are used to judge whether it is a new fault according to the distance criterion. For a new fault, the DNN will retrain by transferring learning in order to reduce the computation and training time. The effectiveness of the algorithm is demonstrated by a numerical simulation example based on a wind turbine benchmark model.
{"title":"An Ensemble Approach for Fault Diagnosis via Continuous Learning","authors":"Dapeng Zhang, Zhiwei Gao","doi":"10.1109/INDIN45523.2021.9557388","DOIUrl":"https://doi.org/10.1109/INDIN45523.2021.9557388","url":null,"abstract":"The great success of deep neural network (DNN) in image field stimulates its application in fault detection and diagnose. However due to the limitation of system security, it is impossible to obtain complete fault data as the training database for neural network, so that it is challenging to identify a fault that never occurred before. In this paper, an ensemble approach is proposed to adapt to a new fault by adding output branches of the neural network. Firstly, the time series are transferred to numerous imaging matrixes. The intrinsic characteristics of the matrixes are then extracted using deep neural network which are used to judge whether it is a new fault according to the distance criterion. For a new fault, the DNN will retrain by transferring learning in order to reduce the computation and training time. The effectiveness of the algorithm is demonstrated by a numerical simulation example based on a wind turbine benchmark model.","PeriodicalId":370921,"journal":{"name":"2021 IEEE 19th International Conference on Industrial Informatics (INDIN)","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125463097","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 : 2021-07-21DOI: 10.1109/INDIN45523.2021.9557485
M. E. I. Martínez, J. Antonino-Daviu, C. Platero, L. Dunai, J. Conejero, P. F. Córdoba
In this work, the application of multifractal spectrum of higher order cumulants slice and bicoherence of stator current signals is proposed as a way to detect field winding faults in wound field synchronous motors. These signals are analyzed both under starting and under steady-state regimes. Likewise, a quantitative indicator based on the summation of the first three log cumulants of the scaling exponents obtained from the multifractal analysis is proposed. In addition, a comparative study is carried out during starting and at steady-state, obtaining satisfactory results that prove the potential of the proposed methodology for its implementation in real applications.
{"title":"Multifractal Spectrum and Higher Order Statistics for the Detection of Field Winding Faults in Wound Field Synchronous Motors","authors":"M. E. I. Martínez, J. Antonino-Daviu, C. Platero, L. Dunai, J. Conejero, P. F. Córdoba","doi":"10.1109/INDIN45523.2021.9557485","DOIUrl":"https://doi.org/10.1109/INDIN45523.2021.9557485","url":null,"abstract":"In this work, the application of multifractal spectrum of higher order cumulants slice and bicoherence of stator current signals is proposed as a way to detect field winding faults in wound field synchronous motors. These signals are analyzed both under starting and under steady-state regimes. Likewise, a quantitative indicator based on the summation of the first three log cumulants of the scaling exponents obtained from the multifractal analysis is proposed. In addition, a comparative study is carried out during starting and at steady-state, obtaining satisfactory results that prove the potential of the proposed methodology for its implementation in real applications.","PeriodicalId":370921,"journal":{"name":"2021 IEEE 19th International Conference on Industrial Informatics (INDIN)","volume":"262 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123021479","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 : 2021-07-21DOI: 10.1109/INDIN45523.2021.9557374
Victor Pazmino Betancourt, Bo Liu, Jürgen Becker
Innovations and novel applications in the area of the Industrial Internet of Things (IIoT) are driven by the technical possibilities of digitalization and edge computing. This leads to rapid advancements and enormous time pressure in the development and operation of new functionalities. Edge computing systems with self-x functionalities are able to react independently to changes in operation and thus mitigate this time pressure problem. The autonomous response during the operation of the self-x system must nevertheless remain compliant with the original system design requirements. A distributed edge computing system has complex requirements in different components and at different levels of the system. This leads to a major challenge when describing these requirements and constraints in such a way that they can be automatically checked and fulfilled during operation. This paper proposes a model-based description of policies that is used as a basis for reallocation of services during operation. The approach was tested and evaluated using an IIoT use case of a camera-based monitoring system for smart construction sites. Our results show that, based on the policy description, it is possible to automatically compute the reallocation when changes occur in the system, without any intervention from the developer. With this self-x capability, the system can remain in operation longer. Overall, this helps to reduce time pressure in the development, deployment and maintenance of new innovations and applications in the field of the Industrial Internet of Things.
{"title":"Towards Policy-based Task Self-Reallocation in Dynamic Edge Computing Systems","authors":"Victor Pazmino Betancourt, Bo Liu, Jürgen Becker","doi":"10.1109/INDIN45523.2021.9557374","DOIUrl":"https://doi.org/10.1109/INDIN45523.2021.9557374","url":null,"abstract":"Innovations and novel applications in the area of the Industrial Internet of Things (IIoT) are driven by the technical possibilities of digitalization and edge computing. This leads to rapid advancements and enormous time pressure in the development and operation of new functionalities. Edge computing systems with self-x functionalities are able to react independently to changes in operation and thus mitigate this time pressure problem. The autonomous response during the operation of the self-x system must nevertheless remain compliant with the original system design requirements. A distributed edge computing system has complex requirements in different components and at different levels of the system. This leads to a major challenge when describing these requirements and constraints in such a way that they can be automatically checked and fulfilled during operation. This paper proposes a model-based description of policies that is used as a basis for reallocation of services during operation. The approach was tested and evaluated using an IIoT use case of a camera-based monitoring system for smart construction sites. Our results show that, based on the policy description, it is possible to automatically compute the reallocation when changes occur in the system, without any intervention from the developer. With this self-x capability, the system can remain in operation longer. Overall, this helps to reduce time pressure in the development, deployment and maintenance of new innovations and applications in the field of the Industrial Internet of Things.","PeriodicalId":370921,"journal":{"name":"2021 IEEE 19th International Conference on Industrial Informatics (INDIN)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114788051","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 : 2021-07-21DOI: 10.1109/INDIN45523.2021.9557392
Yimin Yang, Min Wu
Model explainability has become an important objective when developing machine learning algorithms, especially in highly regulated industries. However, it is difficult to achieve both prediction accuracy and intrinsic explainability as the two objectives usually conflict with each other. Recent development regarding Explainable Neural Network, or xNN, shed some lights on resolving the trade-off between accuracy and explainability for neural network. In this paper, we propose an xNN approach to develop or improve logistic regressions, which can be useful in credit risk modeling and money-laundering or fraud detection. Our data experiment shows that the proposed xNN model keeps the flexibility of pursuing high prediction accuracy while attaining improved explainability.
{"title":"Explainable Machine Learning for Improving Logistic Regression Models","authors":"Yimin Yang, Min Wu","doi":"10.1109/INDIN45523.2021.9557392","DOIUrl":"https://doi.org/10.1109/INDIN45523.2021.9557392","url":null,"abstract":"Model explainability has become an important objective when developing machine learning algorithms, especially in highly regulated industries. However, it is difficult to achieve both prediction accuracy and intrinsic explainability as the two objectives usually conflict with each other. Recent development regarding Explainable Neural Network, or xNN, shed some lights on resolving the trade-off between accuracy and explainability for neural network. In this paper, we propose an xNN approach to develop or improve logistic regressions, which can be useful in credit risk modeling and money-laundering or fraud detection. Our data experiment shows that the proposed xNN model keeps the flexibility of pursuing high prediction accuracy while attaining improved explainability.","PeriodicalId":370921,"journal":{"name":"2021 IEEE 19th International Conference on Industrial Informatics (INDIN)","volume":"80 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134215885","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：