{"status":"public","type":"preprint","external_id":{"arxiv":["2111.05850"]},"date_updated":"2022-04-12T12:01:23Z","language":[{"iso":"eng"}],"date_created":"2022-04-12T11:57:15Z","department":[{"_id":"34"},{"_id":"7"},{"_id":"26"}],"project":[{"_id":"1","name":"SFB 901: SFB 901"},{"_id":"3","name":"SFB 901 - B: SFB 901 - Project Area B"},{"name":"SFB 901 - B2: SFB 901 - Subproject B2","_id":"10"}],"user_id":"38209","year":"2021","author":[{"first_name":"Tanja","last_name":"Tornede","full_name":"Tornede, Tanja","id":"40795"},{"first_name":"Alexander","full_name":"Tornede, Alexander","id":"38209","last_name":"Tornede"},{"first_name":"Jonas Manuel","orcid":"0000-0002-1231-4985","id":"43980","full_name":"Hanselle, Jonas Manuel","last_name":"Hanselle"},{"last_name":"Wever","id":"33176","full_name":"Wever, Marcel Dominik","orcid":" https://orcid.org/0000-0001-9782-6818","first_name":"Marcel Dominik"},{"first_name":"Felix","full_name":"Mohr, Felix","last_name":"Mohr"},{"full_name":"Hüllermeier, Eyke","id":"48129","last_name":"Hüllermeier","first_name":"Eyke"}],"title":"Towards Green Automated Machine Learning: Status Quo and Future Directions","abstract":[{"text":"Automated machine learning (AutoML) strives for the automatic configuration\r\nof machine learning algorithms and their composition into an overall (software)\r\nsolution - a machine learning pipeline - tailored to the learning task\r\n(dataset) at hand. Over the last decade, AutoML has developed into an\r\nindependent research field with hundreds of contributions. While AutoML offers\r\nmany prospects, it is also known to be quite resource-intensive, which is one\r\nof its major points of criticism. The primary cause for a high resource\r\nconsumption is that many approaches rely on the (costly) evaluation of many\r\nmachine learning pipelines while searching for good candidates. This problem is\r\namplified in the context of research on AutoML methods, due to large scale\r\nexperiments conducted with many datasets and approaches, each of them being run\r\nwith several repetitions to rule out random effects. In the spirit of recent\r\nwork on Green AI, this paper is written in an attempt to raise the awareness of\r\nAutoML researchers for the problem and to elaborate on possible remedies. To\r\nthis end, we identify four categories of actions the community may take towards\r\nmore sustainable research on AutoML, i.e. Green AutoML: design of AutoML\r\nsystems, benchmarking, transparency and research incentives.","lang":"eng"}],"publication":"arXiv:2111.05850","citation":{"chicago":"Tornede, Tanja, Alexander Tornede, Jonas Manuel Hanselle, Marcel Dominik Wever, Felix Mohr, and Eyke Hüllermeier. “Towards Green Automated Machine Learning: Status Quo and Future Directions.” <i>ArXiv:2111.05850</i>, 2021.","short":"T. Tornede, A. Tornede, J.M. Hanselle, M.D. Wever, F. Mohr, E. Hüllermeier, ArXiv:2111.05850 (2021).","ama":"Tornede T, Tornede A, Hanselle JM, Wever MD, Mohr F, Hüllermeier E. Towards Green Automated Machine Learning: Status Quo and Future Directions. <i>arXiv:211105850</i>. Published online 2021.","mla":"Tornede, Tanja, et al. “Towards Green Automated Machine Learning: Status Quo and Future Directions.” <i>ArXiv:2111.05850</i>, 2021.","bibtex":"@article{Tornede_Tornede_Hanselle_Wever_Mohr_Hüllermeier_2021, title={Towards Green Automated Machine Learning: Status Quo and Future Directions}, journal={arXiv:2111.05850}, author={Tornede, Tanja and Tornede, Alexander and Hanselle, Jonas Manuel and Wever, Marcel Dominik and Mohr, Felix and Hüllermeier, Eyke}, year={2021} }","apa":"Tornede, T., Tornede, A., Hanselle, J. M., Wever, M. D., Mohr, F., &#38; Hüllermeier, E. (2021). Towards Green Automated Machine Learning: Status Quo and Future Directions. In <i>arXiv:2111.05850</i>.","ieee":"T. Tornede, A. Tornede, J. M. Hanselle, M. D. Wever, F. Mohr, and E. Hüllermeier, “Towards Green Automated Machine Learning: Status Quo and Future Directions,” <i>arXiv:2111.05850</i>. 2021."},"_id":"30866"}