Building Machines that Learn and Think for Themselves: Commentary on Lake et al., Behavioral and Brain Sciences, 2017.
作者： Demis Hassabis , Matthew Botvinick , Tom Schaul , Daan Wierstra , Joel Z. Leibo
DOI:
关键词:
摘要: We agree with Lake and colleagues on their list of key ingredients for building humanlike intelligence, including the idea that model-based reasoning is essential. However, we favor an approach centers one additional ingredient: autonomy. In particular, aim toward agents can both build exploit own internal models, minimal human hand-engineering. believe centered autonomous learning has greatest chance success as scale real-world complexity, tackling domains which ready-made formal models are not available. Here survey several important examples progress been made abilities, highlight some outstanding challenges.
arxiv.org
harvard.edu参考文章(3)
Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Andrei A Rusu, Joel Veness, Marc G Bellemare, Alex Graves, Martin Riedmiller, Andreas K Fidjeland, Georg Ostrovski, Stig Petersen, Charles Beattie, Amir Sadik, Ioannis Antonoglou, Helen King, Dharshan Kumaran, Daan Wierstra, Shane Legg, Demis Hassabis, None, Human-level control through deep reinforcement learning Nature. ,vol. 518, pp. 529- 533 ,(2015) , 10.1038/NATURE14236
Matt Botvinick, Remi Munos, Joel Z Leibo, Zeb Kurth-Nelson, Dharshan Kumaran, Charles Blundell, Hubert Soyer, Jane X Wang, Dhruva Tirumala, Learning to reinforcement learn arXiv: Learning. ,(2016)
Ilya Sutskever, Pieter Abbeel, Peter L. Bartlett, Yan Duan, John Schulman, Xi Chen, RL^2: Fast Reinforcement Learning via Slow Reinforcement Learning arXiv: Artificial Intelligence. ,(2016)