Presents the first online learning-to-defer algorithm with regret bounds O((n + n_e) T^{2/3}) generally and O((n + n_e) sqrt(T)) under low noise for multiclass classification with varying experts.
SIAM journal on computing , volume=
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
years
2026 5verdicts
UNVERDICTED 5representative citing papers
With opponent-action feedback in zero-sum games, an efficient algorithm achieves near-optimal t^{-1/2} last-iterate convergence in duality gap with high probability.
A robust variant of binary search achieves regret O(C + log T) for dynamic pricing with known corruption C and O(C + log² T) when unknown.
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.
Learnability of adversarial noisy bandits is characterized by the convexified generalized maximin volume for oblivious adversaries and for adaptive adversaries when the arm space is countable.
citing papers explorer
-
Online Learning-to-Defer with Varying Experts
Presents the first online learning-to-defer algorithm with regret bounds O((n + n_e) T^{2/3}) generally and O((n + n_e) sqrt(T)) under low noise for multiclass classification with varying experts.
-
Near-Optimal Last-Iterate Convergence for Zero-Sum Games with Bandit Feedback and Opponent Actions
With opponent-action feedback in zero-sum games, an efficient algorithm achieves near-optimal t^{-1/2} last-iterate convergence in duality gap with high probability.
-
Toward Optimal Regret in Robust Pricing: Decoupling Corruption and Time
A robust variant of binary search achieves regret O(C + log T) for dynamic pricing with known corruption C and O(C + log² T) when unknown.
-
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.
-
On Characterizing Learnability for Adversarial Noisy Bandits
Learnability of adversarial noisy bandits is characterized by the convexified generalized maximin volume for oblivious adversaries and for adaptive adversaries when the arm space is countable.