Subtracting the state value from action values scales gradient updates to downweight frequent state-action pairs, helping agents learn causal state representations and generalize out-of-trajectory.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Breaking Habits: On the Role of the Advantage Function in Learning Causal State Representations
Subtracting the state value from action values scales gradient updates to downweight frequent state-action pairs, helping agents learn causal state representations and generalize out-of-trajectory.