Online IL overcomes an information-theoretic bottleneck that offline IL faces in non-realizable settings even at horizon 1, under a new structural characterization of reward-relative misspecification.
arXiv preprint arXiv:2312.16730 , year=
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 4years
2026 4verdicts
UNVERDICTED 4roles
background 1polarities
background 1representative citing papers
The paper establishes the first tilde O(epsilon^{-1}) upper bounds and matching lower bounds for forward-KL-regularized offline contextual bandits under single-policy concentrability in both tabular and general function approximation settings.
A modular reduction from budget-constrained contextual bandits with adversarial contexts to unconstrained bandits via surrogate rewards, yielding improved guarantees and an efficient algorithm based on SquareCB.
A projection-based algorithm for COCO achieves O(log T) regret and O(log T) CCV for strongly convex losses and O(sqrt(T)) for convex losses by leveraging self-contracted curves.
citing papers explorer
-
When Does Online Imitation Learning Help in LLM Post-Training? The Role of (Non-)Realizability Beyond Horizon
Online IL overcomes an information-theoretic bottleneck that offline IL faces in non-realizable settings even at horizon 1, under a new structural characterization of reward-relative misspecification.
-
Fast Rates for Offline Contextual Bandits with Forward-KL Regularization under Single-Policy Concentrability
The paper establishes the first tilde O(epsilon^{-1}) upper bounds and matching lower bounds for forward-KL-regularized offline contextual bandits under single-policy concentrability in both tabular and general function approximation settings.
-
Constrained Contextual Bandits with Adversarial Contexts
A modular reduction from budget-constrained contextual bandits with adversarial contexts to unconstrained bandits via surrogate rewards, yielding improved guarantees and an efficient algorithm based on SquareCB.
-
Improved Guarantees for Constrained Online Convex Optimization via Self-Contraction
A projection-based algorithm for COCO achieves O(log T) regret and O(log T) CCV for strongly convex losses and O(sqrt(T)) for convex losses by leveraging self-contracted curves.