咨询与建议

看过本文的还看了

相关文献

该作者的其他文献

文献详情 >A Domain-Agnostic Approach for... 收藏
arXiv

A Domain-Agnostic Approach for Characterization of Lifelong Learning Systems

作     者:Baker, Megan M. New, Alexander Aguilar-Simon, Mario Al-Halah, Ziad Arnold, Sébastien M.R. Ben-Iwhiwhu, Ese Brna, Andrew P. Brooks, Ethan Brown, Ryan C. Daniels, Zachary Daram, Anurag Delattre, Fabien Dellana, Ryan Eaton, Eric Fu, Haotian Grauman, Kristen Hostetler, Jesse Iqbal, Shariq Kent, Cassandra Ketz, Nicholas Kolouri, Soheil Konidaris, George Kudithipudi, Dhireesha Learned-Miller, Erik Lee, Seungwon Littman, Michael L. Madireddy, Sandeep Mendez, Jorge A. Nguyen, Eric Q. Piatko, Christine Pilly, Praveen K. Raghavan, Aswin Rahman, Abrar Ramakrishnan, Santhosh Kumar Ratzlaff, Neale Soltoggio, Andrea Stone, Peter Sur, Indranil Tang, Zhipeng Tiwari, Saket Vedder, Kyle Wang, Felix Xu, Zifan Yanguas-Gil, Angel Yedidsion, Harel Yu, Shangqun Vallabha, Gautam K. 

作者机构:Johns Hopkins University Applied Physics Laboratory 11100 Johns Hopkins Rd. LaurelMD20723 United States Teledyne Scientific Company Intelligent Systems Laboratory 19 T.W. Alexander Drive NC27709 United States Department of Computer Science University of Texas at Austin AustinTX United States Department of Computer Science University of Southern California Los AngelesCA United States Department of Computer Science Loughborough University Loughborough United Kingdom Department of Electrical Engineering and Computer Science University of Michigan Ann ArborMI United States SRI International 201 Washington Rd PrincetonNJ United States University of Texas at San Antonio San AntonioTX United States Department of Computer Science University of Massachusetts Amherst AmherstMA United States Sandia National Laboratories AlbuquerqueNM United States Department of Computer and Information Science University of Pennsylvania PhiladelphiaPA United States Department of Computer Science Brown University ProvidenceRI United States Information and Systems Sciences Laboratory HRL Laboratories 3011 Malibu Canyon Road MalibuCA90265 United States Department of Computer Science Vanderbilt University NashvilleTN United States Argonne National Laboratory 9700 S Cass Ave LemontIL United States 

出 版 物:《arXiv》 (arXiv)

年 卷 期:2023年

核心收录:

主  题:Reinforcement learning 

摘      要:Despite the advancement of machine learning techniques in recent years, state-of-the-art systems lack robustness to real world events, where the input distributions and tasks encountered by the deployed systems will not be limited to the original training context, and systems will instead need to adapt to novel distributions and tasks while deployed. This critical gap may be addressed through the development of Lifelong Learning systems that are capable of 1) Continuous Learning, 2) Transfer and Adaptation, and 3) Scalability. Unfortunately, efforts to improve these capabilities are typically treated as distinct areas of research that are assessed independently, without regard to the impact of each separate capability on other aspects of the system. We instead propose a holistic approach, using a suite of metrics and an evaluation framework to assess Lifelong Learning in a principled way that is agnostic to specific domains or system techniques. Through five case studies, we show that this suite of metrics can inform the development of varied and complex Lifelong Learning systems. We highlight how the proposed suite of metrics quantifies performance trade-offs present during Lifelong Learning system development - both the widely discussed Stability-Plasticity dilemma and the newly proposed relationship between Sample Efficient and Robust Learning. Further, we make recommendations for the formulation and use of metrics to guide the continuing development of Lifelong Learning systems and assess their progress in the future. Copyright © 2023, The Authors. All rights reserved.

读者评论 与其他读者分享你的观点

用户名:未登录
我的评分