Season 2 Ep. 1 Sergey Levine explains the challenges of real world robotics

In Episode One of Season Two, Host Pieter Abbeel is joined by guest (and close collaborator) Sergey Levine, professor at UC Berkeley, EECS. Sergey discusses the early years of his career, how Andrew Ng influenced him to become interested in machine learning, his current projects, and his lab's recent accomplishments.
The conversation concludes with Sergey's view on the dangers of machines not being intelligent enough and his advice for students seeking a career in robotic.
What's in this episode:
00:00:00 Introduction
00:02:07 Sergey's PhD work
00:04:55 Andrew Ng's influence
00:05:46 Defining supervised learning and reinforcement learning (RL)
00:10:03 Sergey's predictions for RL in the future
00:11:29 Sergey's switch from graphics to robotics
00:13:58 Robots learning in the real world and Sergey's current research
00:32:48 How to keep collecting useful data
00:43:31 Offline model-based optimization
00:47:12 The concept of singularity
00:54:10 Managing AI responsibly
Links:
Sergey's Twitter: / svlevine
EECS lab: rail.eecs.berkeley.edu/
Sergey's Website: people.eecs.berkeley.edu/~svl...
Subscribe and listen to our podcast today:
Apple: apple.co/3vRqBlV
Spotify: spoti.fi/3d0PEwf
Amazon: amzn.to/2SWdZfP
Google: bit.ly/3d52Uju
Acast: bit.ly/3A0uqIO
Host: Pieter Abbeel
Executive Producers: Alice Patel & Henry Tobias Jones
Audio Production: Kieron Matthew Banerji
Video Production: Bo Obradovic

[@masoodmortazavi]

These interviews are also extremely important from a historical point of view on how the research evolved in the past decade. Nice. . . . . "The Confusing Case of the Pink Stapler" . . . could be the name of a book on the subject.

[@NoStackEngineer]

I'm very excited for what this season has in store, I can't thank you enough for spending your time on this! Also, excellent choice of guest to start things off. Perhaps I will get the pleasure of meeting the both of you in the lab one day :)

[@peterdavidfagan]

Great episode 🤖 to start season 2, thank you for this and looking forward to continued innovation in the field of robotics and rl 🙌.

[@TheRobotBrainsPodcast]

Our pleasure!

[@FiDelZarlar]

Extremely interesting subject!

[@sercharnaud8690]

Thank you so much, this is really inspiring 💙

[@Dobafresh]

Great video! Thanks for making conversations like these available. Can I make a small suggestion for improving future videos, though? Can you mute your mic while the guest is talking? It is very distracting to hear "mhmm" every few seconds throughout the video.

[@pieterabbeel2107]

nice suggestion!

[@scienceisfuture]

“Attention is all you need.”
Come on, kid. Just focus on Sergey’s voice. Is that difficult for you? Even a robot 🤖 could do that hahahahahaha

[@tnfru]

never realized how much of a fanboy i am of you two :X

[@NikeyvS]

Which algorithms is he specificially referring to at th-cam.com/video/vMEdchIjzfE/w-d-xo.html ?
Model based RL algorithms? Or SAC? Which ones?

[@ninatko]

Sorry but I cannot agree about RL vs SL, in RL you too do prediction on value or ev of action. But is true that the learning processes are not the same. Ye RL has advantage here that the prediction space is much bigger, nevertheless the whole value funcion stuff is just doomed to ever work imo