Chase D. Carthen, Christopher Lewis, Vinh D. Le, A. Tavakkoli, Frederick Harris, S. Dascalu
{"title":"星期四:支持AutoML的网络平台","authors":"Chase D. Carthen, Christopher Lewis, Vinh D. Le, A. Tavakkoli, Frederick Harris, S. Dascalu","doi":"10.1145/3543895.3543940","DOIUrl":null,"url":null,"abstract":"THURSDAY is a web platform that aids users in building machine learning models by providing easily accessible tools to either create models manually, or through the use of automated machine learning (AutoML) libraries like AutoKeras. As part of THURSDAY’s key innovations, users are given the opportunity to configure and run multiple machine learning models. The results of these model executions can then be compared with built-in performance metrics. Finally, THURSDAY allows users to analyze hyperparameter changes, as well as the changes created by AutoML libraries, in order to provide a vital tool that aids in the revision of existing models. To meet the high volume demands of machine learning, THURDAY adopted a microservice-based design pattern that supports containerization, orchestration, and scalabability. In this paper, the design, implementation, and impact of the THURSDAY system is explored in detail. In order to evaluate the capability of THURSDAY, its core functionality is compared against similar platforms that provide machine learning support.","PeriodicalId":191129,"journal":{"name":"Proceedings of the 9th International Conference on Applied Computing & Information Technology","volume":"9 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"THURSDAY: A Web Platform to Support AutoML\",\"authors\":\"Chase D. Carthen, Christopher Lewis, Vinh D. Le, A. Tavakkoli, Frederick Harris, S. Dascalu\",\"doi\":\"10.1145/3543895.3543940\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"THURSDAY is a web platform that aids users in building machine learning models by providing easily accessible tools to either create models manually, or through the use of automated machine learning (AutoML) libraries like AutoKeras. As part of THURSDAY’s key innovations, users are given the opportunity to configure and run multiple machine learning models. The results of these model executions can then be compared with built-in performance metrics. Finally, THURSDAY allows users to analyze hyperparameter changes, as well as the changes created by AutoML libraries, in order to provide a vital tool that aids in the revision of existing models. To meet the high volume demands of machine learning, THURDAY adopted a microservice-based design pattern that supports containerization, orchestration, and scalabability. In this paper, the design, implementation, and impact of the THURSDAY system is explored in detail. In order to evaluate the capability of THURSDAY, its core functionality is compared against similar platforms that provide machine learning support.\",\"PeriodicalId\":191129,\"journal\":{\"name\":\"Proceedings of the 9th International Conference on Applied Computing & Information Technology\",\"volume\":\"9 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2022-05-25\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Proceedings of the 9th International Conference on Applied Computing & Information Technology\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1145/3543895.3543940\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 9th International Conference on Applied Computing & Information Technology","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3543895.3543940","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
THURSDAY is a web platform that aids users in building machine learning models by providing easily accessible tools to either create models manually, or through the use of automated machine learning (AutoML) libraries like AutoKeras. As part of THURSDAY’s key innovations, users are given the opportunity to configure and run multiple machine learning models. The results of these model executions can then be compared with built-in performance metrics. Finally, THURSDAY allows users to analyze hyperparameter changes, as well as the changes created by AutoML libraries, in order to provide a vital tool that aids in the revision of existing models. To meet the high volume demands of machine learning, THURDAY adopted a microservice-based design pattern that supports containerization, orchestration, and scalabability. In this paper, the design, implementation, and impact of the THURSDAY system is explored in detail. In order to evaluate the capability of THURSDAY, its core functionality is compared against similar platforms that provide machine learning support.