Huber-robust confidence sequences
Confidence sequences are confidence intervals that can be sequentially tracked, and are
valid at arbitrary data-dependent stopping times. This paper presents confidence sequences …
valid at arbitrary data-dependent stopping times. This paper presents confidence sequences …
Piecewise stationary bandits under risk criteria
Piecewise stationary stochastic multi-armed bandits have been extensively explored in the
risk-neutral and sub-Gaussian setting. In this work, we consider a multi-armed bandit …
risk-neutral and sub-Gaussian setting. In this work, we consider a multi-armed bandit …
Empirical Risk Minimization for Losses without Variance
This paper considers an empirical risk minimization problem under heavy-tailed settings,
where data does not have finite variance, but only has $ p $-th moment with $ p\in (1, 2) …
where data does not have finite variance, but only has $ p $-th moment with $ p\in (1, 2) …
Adaptive Robust Confidence Intervals
Y Luo, C Gao - arXiv preprint arXiv:2410.22647, 2024 - arxiv.org
This paper studies the construction of adaptive confidence intervals under Huber's
contamination model when the contamination proportion is unknown. For the robust …
contamination model when the contamination proportion is unknown. For the robust …
Robust Score Matching
Proposed in Hyv\" arinen (2005), score matching is a parameter estimation procedure that
does not require computation of distributional normalizing constants. In this work we utilize …
does not require computation of distributional normalizing constants. In this work we utilize …
Locally Private and Robust Multi-Armed Bandits
X Zhou, W Zhang - The Thirty-eighth Annual Conference on Neural … - openreview.net
We study the interplay between local differential privacy (LDP) and robustness to Huber
corruption and possibly heavy-tailed rewards in the context of multi-armed bandits (MABs) …
corruption and possibly heavy-tailed rewards in the context of multi-armed bandits (MABs) …