Timezone: »

 
Poster
Gamma-Models: Generative Temporal Difference Learning for Infinite-Horizon Prediction
Michael Janner · Igor Mordatch · Sergey Levine

Tue Dec 08 09:00 AM -- 11:00 AM (PST) @ Poster Session 1 #544

We introduce the gamma-model, a predictive model of environment dynamics with an infinite, probabilistic horizon. Replacing standard single-step models with gamma-models leads to generalizations of the procedures that form the foundation of model-based control, including the model rollout and model-based value estimation. The gamma-model, trained with a generative reinterpretation of temporal difference learning, is a natural continuous analogue of the successor representation and a hybrid between model-free and model-based mechanisms. Like a value function, it contains information about the long-term future; like a standard predictive model, it is independent of task reward. We instantiate the gamma-model as both a generative adversarial network and normalizing flow, discuss how its training reflects an inescapable tradeoff between training-time and testing-time compounding errors, and empirically investigate its utility for prediction and control.

Author Information

Michael Janner (UC Berkeley)
Igor Mordatch (Google)
Sergey Levine (UC Berkeley)

More from the Same Authors