Machine learning with a reject option: A survey

K Hendrickx, L Perini, D Van der Plas, W Meert… - Machine Learning, 2024‏ - Springer
Abstract Machine learning models always make a prediction, even when it is likely to be
inaccurate. This behavior should be avoided in many decision support applications, where …

Beyond perturbations: Learning guarantees with arbitrary adversarial test examples

S Goldwasser, AT Kalai, Y Kalai… - Advances in Neural …, 2020‏ - proceedings.neurips.cc
We present a transductive learning algorithm that takes as input training examples from a
distribution P and arbitrary (unlabeled) test examples, possibly chosen by an adversary. This …

Learning to defer in content moderation: The human-ai interplay

T Lykouris, W Weng - arxiv preprint arxiv:2402.12237, 2024‏ - arxiv.org
Successful content moderation in online platforms relies on a human-AI collaboration
approach. A typical heuristic estimates the expected harmfulness of a post and uses fixed …

Adversarial resilience in sequential prediction via abstention

S Goel, S Hanneke, S Moran… - Advances in Neural …, 2023‏ - proceedings.neurips.cc
We study the problem of sequential prediction in the stochastic setting with an adversary that
is allowed to inject clean-label adversarial (or out-of-distribution) examples. Algorithms …

Online learning with sublinear best-action queries

M Russo, A Celli, R Colini Baldeschi… - Advances in …, 2025‏ - proceedings.neurips.cc
In online learning, a decision maker repeatedly selects one of a set of actions, with the goal
of minimizing the overall loss incurred. Following the recent line of research on algorithms …

Online learning with abstention

C Cortes, G DeSalvo, C Gentile… - … on machine learning, 2018‏ - proceedings.mlr.press
We present an extensive study of a key problem in online learning where the learner can opt
to abstain from making a prediction, at a certain cost. In the adversarial setting, we show how …

Partially interpretable models with guarantees on coverage and accuracy

N Frost, Z Lipton, Y Mansour… - … on algorithmic learning …, 2024‏ - proceedings.mlr.press
Simple, sufficient explanations furnished by short decision lists can be useful for guiding
stakeholder actions. Unfortunately, this transparency can come at the expense of the higher …

The extended littlestone's dimension for learning with mistakes and abstentions

C Zhang, K Chaudhuri - Conference on learning theory, 2016‏ - proceedings.mlr.press
This paper studies classification with an abstention option in the online setting. In this
setting, examples arrive sequentially, the learner is given a hypothesis class\mathcalH, and …

Online decision mediation

D Jarrett, A Hüyük… - Advances in Neural …, 2022‏ - proceedings.neurips.cc
Consider learning a decision support assistant to serve as an intermediary between (oracle)
expert behavior and (imperfect) human behavior: At each time, the algorithm observes an …

Active online learning with hidden shifting domains

Y Chen, H Luo, T Ma, C Zhang - International Conference on …, 2021‏ - proceedings.mlr.press
Online machine learning systems need to adapt to domain shifts. Meanwhile, acquiring label
at every timestep is expensive. We propose a surprisingly simple algorithm that adaptively …