Podcasts Technology Data Science at Home What is wrong with reinforcement learning? (Ep. 82)

What is wrong with reinforcement learning? (Ep. 82)

Published: Oct. 15, 2019, 3:09 a.m.

b'Join the discussion on our Discord server\\n\\xa0\\nAfter reinforcement learning agents doing great at playing Atari video games, Alpha Go, doing financial trading, dealing with language modeling, let me tell you the real story here.In this episode I want to shine some light on reinforcement learning (RL) and the limitations that every practitioner should consider before taking certain directions.\\xa0RL seems to work so well! What is wrong with it?\\n\\xa0\\n\\nAre you a listener of\\xa0Data Science at Home\\xa0podcast? A reader of the Amethix Blog?\\xa0Or did you subscribe to the\\xa0Artificial Intelligence at your fingertips\\xa0newsletter? In any case let\\u2019s stay in touch!\\xa0https://amethix.com/survey/\\n\\n\\xa0\\n\\xa0\\nReferences\\nEmergence of Locomotion Behaviours in Rich Environments\\xa0https://arxiv.org/abs/1707.02286\\nRainbow: Combining Improvements in Deep Reinforcement Learning\\xa0https://arxiv.org/abs/1710.02298\\nAlphaGo Zero: Starting from scratch\\xa0https://deepmind.com/blog/article/alphago-zero-starting-scratch'