With time, we hope these design principles can help inform best practices for how to build capable RL agents without reward tampering incentives.
— DeepMind (@DeepMindAI) August 14, 2019
This builds upon our previous work on understanding agent incentives with causal influence diagrams: https://t.co/WBc7fUJCbf
from Twitter https://twitter.com/DeepMindAI
August 14, 2019 at 08:47AM
via IFTTT
No comments:
Post a Comment