Set-valued classification--overview via a unified framework
Multi-class classification problem is among the most popular and well-studied statistical
frameworks. Modern multi-class datasets can be extremely ambiguous and single-output …
frameworks. Modern multi-class datasets can be extremely ambiguous and single-output …
Leveraging labeled and unlabeled data for consistent fair binary classification
We study the problem of fair binary classification using the notion of Equal Opportunity. It
requires the true positive rate to distribute equally across the sensitive groups. Within this …
requires the true positive rate to distribute equally across the sensitive groups. Within this …
Selective classification via one-sided prediction
A Gangrade, A Kag… - … Conference on Artificial …, 2021 - proceedings.mlr.press
We propose a novel method for selective classification (SC), a problem which allows a
classifier to abstain from predicting some instances, thus trading off accuracy against …
classifier to abstain from predicting some instances, thus trading off accuracy against …
Fair regression via plug-in estimator and recalibration with statistical guarantees
We study the problem of learning an optimal regression function subject to a fairness
constraint. It requires that, conditionally on the sensitive feature, the distribution of the …
constraint. It requires that, conditionally on the sensitive feature, the distribution of the …
Uncertainty in predictions of deep learning models for fine-grained classification
T Lorieul - 2020 - theses.hal.science
Deep neural networks have shown dramatic improvements in a lot of supervised
classification tasks. Such models are usually trained with the objective to ultimately minimize …
classification tasks. Such models are usually trained with the objective to ultimately minimize …
Nearest neighbor based conformal prediction
L Gyôrfi, H Walk - Annales de l'ISUP, 2019 - hal.science
In this paper we introduce a nearest neighbor based estimate of the prédiction interval with
prescribed conditional coverage probability and with small length. In the spécial case, when …
prescribed conditional coverage probability and with small length. In the spécial case, when …
Classification under ambiguity: when is average-k better than top-k?
When many labels are possible, choosing a single one can lead to low precision. A common
alternative, referred to as top-$ K $ classification, is to choose some number $ K $(commonly …
alternative, referred to as top-$ K $ classification, is to choose some number $ K $(commonly …
Regression under demographic parity constraints via unlabeled post-processing
We address the problem of performing regression while ensuring demographic parity, even
without access to sensitive attributes during inference. We present a general-purpose post …
without access to sensitive attributes during inference. We present a general-purpose post …
Nonparametric active learning for cost-sensitive classification
Cost-sensitive learning is a common type of machine learning problem where different
errors of prediction incur different costs. In this paper, we design a generic nonparametric …
errors of prediction incur different costs. In this paper, we design a generic nonparametric …
Learning fair models and representations
L Oneto - Intelligenza Artificiale, 2020 - content.iospress.com
Abstract Machine learning based systems and products are reaching society at large in
many aspects of everyday life, including financial lending, online advertising, pretrial and …
many aspects of everyday life, including financial lending, online advertising, pretrial and …