[ICLR] Boosting RL Solvability Demo 4/23 @12p, Room 204C

[Marisa Klee]

RL training breaks down on hard tasks where pass@k = 0. How do we get useful learning signals for tasks which typically remain unsolvable through RL?

Satyaki Chakraborty is heading to ICLR in Rio with the AGI Lab and will be discussing how we are making reinforcement learning more scalable and broadly applicable.
Topic: Boosting RL Solvability
When: Thursday, April 23rd @ 12pm 
Where: Room hashtag#204C in the Riocentro Convention and Event Center in Rio de Janeiro, Brazil

You can find more information here.

See you there!

Marisa Klee

AGI Lab