arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.02880cs.AI
收藏

探索探索:在统一环境中比较儿童和RL代理

Exploring Exploration: Comparing Children with RL Agents in Unified Environments

University of California, Berkeley(加州大学伯克利分校);DeepMind;Massachusetts Institute of Technology(麻省理工学院)

Eliza Kosoy, Jasmine Collins, David M. Chan, Sandy Huang, Deepak Pathak, Pulkit Agrawal, John Canny, Alison Gopnik, Jessica B. Hamrick

摘要:

发展心理学的研究一致表明,儿童彻底而有效地探索世界,这种探索使他们能够学习。反过来,这种早期学习支持更强大的概括和智能行为在以后的生活。虽然在开发机器学习的探索方法方面已经做了很多工作,但人工智能还没有达到人类同行设定的高标准。在这项工作中,我们建议使用DeepMind实验室(Beattie等人,2016)作为一个平台,直接比较儿童和代理的行为,并开发新的探索技术。我们概述了两个正在进行的实验,以证明直接比较的有效性,并概述了一些开放的研究问题,我们认为可以使用这种方法进行测试。

英文摘要:

Research in developmental psychology consistently shows that children explore the world thoroughly and efficiently and that this exploration allows them to learn. In turn, this early learning supports more robust generalization and intelligent behavior later in life. While much work has gone into developing methods for exploration in machine learning, artificial agents have not yet reached the high standard set by their human counterparts. In this work we propose using DeepMind Lab (Beattie et al., 2016) as a platform to directly compare child and agent behaviors and to develop new exploration techniques. We outline two ongoing experiments to demonstrate the effectiveness of a direct comparison, and outline a number of open research questions that we believe can be tested using this methodology.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑