Reinforcement Learning

Ravid Shwartz-Ziv's research on world models, predictive representations, learned dynamics, and training agents in imagination.

Training agents in learned world models: how model errors and rollout budgets shape policy learning.