NC unit secretary
2020-10-08 13:27:26

This is a discussion channel for “Deep Reinforcement Learning from AlphaGo to AlphaStar” by Dr. David Silver (DeepMind). A link to the talk is the following. Please do not share the link to anyone outside of this slack workspace. Access Pass code can be found at the announcement channel.   URL: https://vimeo.com/471271876 (31 minutes)

olewow
2020-10-08 23:39:55

Dear Dr. Silver, at first I hope you are fine. Thank you for the great presentation (and for your excellent videolectures which introduced me to RL ). I would like to ask what is your general view on meta-reinforcement learning and how RL and Deep RL can be utilized to general-purpose methods in other ways. What do you think would or should be the relative fields which will grow next years?

👍 P Sakagiannis, Bisser Raytchev
olewow
2020-10-08 23:43:23
Image from Android
😂 Mario Aburto, Mauro Comi, P Sakagiannis, Taiki Miyagawa, olewow, Harish Gunasekaran, SHI Mingyang, Kan Sakamoto, Alex Carvalho, Ryuichiro Hataya, Leon WAN, Naoto Yoshida, Anupama Rajkumar, Kanon Maruyama, Shohei Hido (Preferred Networks)
uchibe
2020-10-10 16:36:48

Thank you for your interesting talk. If my understanding is correct, AlphaFold does not use reinforcement learning. What is the major difficulty in applying RL to the problem of protein structure prediction?

👍 olewow, Tatsuya Matsushima
Kenji Doya
2020-10-10 21:26:20

Thank you for your clear demonstration of how first principle of reinforcement learning can scale with deep learning. One question is how you selected the network architecture and learning parameters to achieve such extremely high performance.

👍 Tatsuya Matsushima
Robert Mueller
2020-10-10 21:39:45

Thank you for your talk! Brain inspired architectures have significantly advanced subfields of AI (CNNs for computer vision, recurrent networks for language). Is there a brain inspired model architecture advancing RL to a similar extent (and if not what are properties such an RL specific architecture should possess)?

Shin'ya Nishida
2020-10-11 18:53:06

Thank you. Do you think it possible for future RL to find the best strategy for human social actions (e.g., what we should say to the supporters of the current US president to make our world better), and what will be the problems we should solve to achieve that goal?

👍 Anupama Rajkumar, Takeshi Yamada