Introduces the first pessimism-free offline learning guarantee for KL-regularized games with fast O(1/n) sample complexity via smoothness and skew-symmetry stability.
21 C.2 Proof of Lemma 7 Proof
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.GT 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Beyond Pessimism: Offline Learning in KL-regularized Games
Introduces the first pessimism-free offline learning guarantee for KL-regularized games with fast O(1/n) sample complexity via smoothness and skew-symmetry stability.