arXiv research

Paper presents robust confidence sequences for means with known moment bounds and arbitrary corruption.

A new method optimizes robustness measures under input uncertainty using randomized Gaussian process upper confidence bound.

New framework tightens certified robustness gaps in machine learning models.

WS-KDE provides robust confidence bounds for stochastic functions.

Paper aims to ensure reliable detection of out-of-distribution data with certifiable worst-case guarantees.

Paper improves confidence intervals and variance estimation for deep learning models.

Machine learning (ML) classification is increasingly used in safety-critical systems. Protecting ML classifiers from adversarial examples is crucial. We propose that the main threat is that of an attacker perturbing a confidently classified input to produce a confident misclassification. To protect against this we devi…

This paper analyzes statistical properties of the Robust Satisficing model.

This paper proposes a distributionally robust approach to logistic regression. We use the Wasserstein distance to construct a ball in the space of probability distributions centered at the uniform distribution on the training samples. If the radius of this ball is chosen judiciously, we can guarantee that it contains t…

CP improves robustness against distribution shift using physics-informed structural causal models.

New flexible confidence sequences for robust statistical inference.

Deep neural networks bring in impressive accuracy in various applications, but the success often relies on the heavy network architecture. Taking well-trained heavy networks as teachers, classical teacher-student learning paradigm aims to learn a student network that is lightweight yet accurate. In this way, a portable…

New algorithm robust to outliers in Bayesian Optimization.

Bayes-assisted confidence sequences improve efficiency for bounded means.

Classifiers used in the wild, in particular for safety-critical systems, should not only have good generalization properties but also should know when they don't know, in particular make low confidence predictions far away from the training data. We show that ReLU type neural networks which yield a piecewise linear cla…

Standard results in stochastic convex optimization bound the number of samples that an algorithm needs to generate a point with small function value in expectation. More nuanced high probability guarantees are rare, and typically either rely on "light-tail" noise assumptions or exhibit worse sample complexity. In this …

Adversarial training yields robust models against a specific threat model, e.g., $L_\infty$ adversarial examples. Typically robustness does not generalize to previously unseen threat models, e.g., other $L_p$ norms, or larger perturbations. Our confidence-calibrated adversarial training (CCAT) tackles this problem by b…

Paper solves robust convex problems with heavy-tailed noise.

Optimism about the poorly understood states and actions is the main driving force of exploration for many provably-efficient reinforcement learning algorithms. We propose optimism in the face of sensible value functions (OFVF)- a novel data-driven Bayesian algorithm to constructing Plausibility sets for MDPs to explore…

Recent work has shown that state-of-the-art classifiers are quite brittle, in the sense that a small adversarial change of an originally with high confidence correctly classified input leads to a wrong classification again with high confidence. This raises concerns that such classifiers are vulnerable to attacks and ca…

We develop an efficient algorithm to find confidence ellipsoids with volume guarantees in high dimensions.

This paper proposes a DGP approach with UCBs for point target tracking over WSNs.

In this paper we study leveraging confidence information induced by adversarial training to reinforce adversarial robustness of a given adversarially trained model. A natural measure of confidence is $\|F({\bf x})\|_\infty$ (i.e. how confident $F$ is about its prediction?). We start by analyzing an adversarial training…

Robust MDPs (RMDPs) can be used to compute policies with provable worst-case guarantees in reinforcement learning. The quality and robustness of an RMDP solution are determined by the ambiguity set---the set of plausible transition probabilities---which is usually constructed as a multi-dimensional confidence region. E…

New tighter confidence bounds for sequential kernel regression.

PAC-MCTS addresses biased search in LLM-guided planning by dynamically pruning.

We propose a bootstrap-based robust high-confidence level upper bound (Robust H-CLUB) for assessing the risks of large portfolios. The proposed approach exploits rank-based and quantile-based estimators, and can be viewed as a robust extension of the H-CLUB method (Fan et al., 2015). Such an extension allows us to hand…

New algorithms balance collaboration and adversarial behavior in linear bandits.

A new method for evaluating and selecting policies in contextual bandits improves confidence intervals and policy quality.

This paper improves offline contextual bandits using distributional robustness.

A new differentiable UCB algorithm for linear bandits learns adaptive confidence bounds.

We introduce a probabilistic robustness measure for Bayesian Neural Networks (BNNs), defined as the probability that, given a test point, there exists a point within a bounded set such that the BNN prediction differs between the two. Such a measure can be used, for instance, to quantify the probability of the existence…

FROB model improves robustness and reliable confidence for few-shot OoD detection.

Method estimates treatment effect bounds in sample selection models.

A new algorithm reduces regret in cooperative multi-agent bandits with heavy-tailed data.

Adversarial training based on the minimax formulation is necessary for obtaining adversarial robustness of trained models. However, it is conservative or even pessimistic so that it sometimes hurts the natural generalization. In this paper, we raise a fundamental question---do we have to trade off natural generalizatio…

Sharp bounds on ATE with unmeasured confounders, valid even when misspecified.

New method improves certified robustness for classifier confidence.

SelectMix improves deep learning robustness against noisy labels.

The paper honors Lai's contributions to multi-armed bandits and establishes new regret bounds.

Paper improves neural network robustness analysis for safety-critical systems.

We derive PAC-Bayesian learning guarantees for heavy-tailed losses, and obtain a novel optimal Gibbs posterior which enjoys finite-sample excess risk bounds at logarithmic confidence. Our core technique itself makes use of PAC-Bayesian inequalities in order to derive a robust risk estimator, which by design is easy to …

The paper extends confidence sequences for infinite variance data.

Improved confidence bounds for linear logistic model with applications to bandits.

Optimizes target value in stochastic black box functions.

Research aims to improve confidence intervals for RKHS elements in online learning.

New method reduces confidence interval sizes for causal inference.

Near-optimal confidence intervals for bounded data.