X post

Chelsea Finn shares paper on generative value classifiers: a new Q-function parameterization without action conditioning

Chelsea Finn shared Alexander Swerdlow's paper, which proposes generative value classifiers, a way of parameterizing Q-functions that does not require action conditioning. The authors believe this type of method can be combined with large-scale pretrained policies and can learn and generalize better from richer output signals. The paper link is alexanderswerdlow.github.io/gvc/.

Part of the event: Generative Value Classifier: Q-function parameterization without action conditioning

This page provides a summary and source link only.