Timezone: »

Learning Transferable Skills
Marwan Mattar · Arthur Juliani · Danny Lange · Matthew Crosby · Benjamin Beyret

Sat Dec 14 08:00 AM -- 06:00 PM (PST) @ West 211 - 214
Event URL: https://www.skillsworkshop.ai/ »

After spending several decades on the margin of AI, reinforcement learning has recently emerged as a powerful framework for developing intelligent systems that can solve complex tasks in real-world environments. This has had a tremendous impact on a wide range of tasks ranging from playing games such as Go and StarCraft to learning dexterity. However, one attribute of intelligence that still eludes modern learning systems is generalizability. Until very recently, the majority of reinforcement learning research involved training and testing algorithms on the same, sometimes deterministic, environment. This has resulted in algorithms that learn policies that typically perform poorly when deployed in environments that differ, even slightly, from those they were trained on. Even more importantly, the paradigm of task-specific training results in learning systems that scale poorly to a large number of (even interrelated) tasks.

Recently there has been an enduring interest in developing learning systems that can learn transferable skills. This could mean robustness to changing environment dynamics, the ability to quickly adapt to environment and task variations or the ability to learn to perform multiple tasks at once (or any combination thereof). This interest has also resulted in a number of new data sets and challenges (e.g. Obstacle Tower Environment, Animal-AI, CoinRun) and an urgency to standardize the metrics and evaluation protocols to better assess the generalization abilities of novel algorithms. We expect this area to continue to increase in popularity and importance, but this can only happen if we manage to build consensus on which approaches are promising, and, equally important, how to test them.

The workshop will include a mix of invited speakers, accepted papers (oral and poster sessions) and a panel discussion. The workshop welcomes both theoretical and applied research, in addition to novel data sets and evaluation protocols.

Sat 9:00 a.m. - 9:15 a.m.
Opening Remarks (Announcement)
Marwan Mattar, Arthur Juliani, Matthew Crosby, Benjamin Beyret, Danny Lange
Sat 9:15 a.m. - 10:00 a.m.
Challenges of Deep RL in Complex Environments (Invited Talk)
Raia Hadsell
Sat 10:00 a.m. - 10:30 a.m.
Coffee Break (Break)
Sat 10:30 a.m. - 11:20 a.m.
Environments and Data Sets (Invited Talk)
Karl Cobbe, Gianni De Fabritiis, Denys Makoviichuk
Sat 11:20 a.m. - 12:05 p.m.
Vladlen Koltun (Intel) (Invited Talk)
Vladlen Koltun
Sat 12:05 p.m. - 1:30 p.m.
Lunch (Break)
Sat 1:30 p.m. - 2:15 p.m.
Innate Bodies, Innate Brains, and Innate World Models (Invited Talk)
David Ha
Sat 2:15 p.m. - 3:15 p.m.
Oral Presentations (Talk)
Janith Petangoda, Sergio Pascual-Diaz, Jordi Grau-Moya, Raphaël Marinier, Olivier Pietquin, Alexei Efros, Phillip Isola, Trevor Darrell, Chris Lu, Deepak Pathak, Johan Ferret
Sat 3:15 p.m. - 4:15 p.m.
Poster Presentations (Poster Session)
Rahul Mehta, Andrew Lampinen, Binghong Chen, Sergio Pascual-Diaz, Jordi Grau-Moya, Aldo Faisal, Jonathan Tompson, Yiren Lu, Khimya Khetarpal, Martin Klissarov, Pierre-Luc Bacon, Doina Precup, Thanard Kurutach, Aviv Tamar, Pieter Abbeel, Jinke He, Max Igl, Shimon Whiteson, Wendelin Boehmer, Raphaël Marinier, Olivier Pietquin, Karol Hausman, Sergey Levine, Chelsea Finn, Tianhe (Kevin) Yu, Lisa Lee, Benjamin Eysenbach, Emilio Parisotto, Eric Xing, Ruslan Salakhutdinov, Hongyu Ren, Anima Anandkumar, Deepak Pathak, Chris Lu, Trevor Darrell, Alexei Efros, Phillip Isola, Feng Liu, Bo Han, Gang Niu, Masashi Sugiyama, Saurabh Kumar, Janith Petangoda, Johan Ferret, Jay McClelland, Kara Liu, Animesh Garg, Rob Lange
Sat 4:15 p.m. - 5:00 p.m.
Multi-Task Reinforcement Learning and Generalization (Invited Talk)
Katja Hofmann
Sat 5:00 p.m. - 5:45 p.m.
Solving Rubik’s Cube with a Robot Hand (Invited Talk)
Wojciech Zaremba
Sat 5:45 p.m. - 6:00 p.m.
Closing Remarks (Announcement)
Marwan Mattar, Arthur Juliani, Matthew Crosby, Benjamin Beyret, Danny Lange

Author Information

Marwan Mattar (Unity Technologies)
Arthur Juliani (Unity Technologies)
Danny Lange (Unity Technologies)
Matthew Crosby (Imperial College London)
Benjamin Beyret (Imperial College London)

I am currently developing the [Animal-AI Olympics, ](http://www.animalaiolympics.com/) which offers researchers a platform for testing AI algorithms on various tasks inspired by the animal cognition literature.

More from the Same Authors