EXtremely PRIvate supervised Learning - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2021

EXtremely PRIvate supervised Learning

Abstract

This paper presents a new approach called ExPriL for learning from extremely private data. Iteratively, the learner supplies a candidate hypothesis and the data curator only releases the marginals of the error incurred by the hypothesis on the privately-held target data. Using the marginals as supervisory signal, the goal is to learn a hypothesis that fits this target data as best as possible. The privacy of the mechanism is provably enforced, assuming that the overall number of iterations is known in advance.
Fichier principal
Vignette du fichier
CAp2021_paper_33.pdf (449.95 Ko) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-03620873 , version 1 (27-03-2022)

Identifiers

  • HAL Id : hal-03620873 , version 1

Cite

Armand Lacombe, Saumya Jetley, Michèle Sebag. EXtremely PRIvate supervised Learning. Conférence d'APprentssage - CAP, Jun 2021, St-Etienne, France. ⟨hal-03620873⟩
66 View
64 Download

Share

Gmail Mastodon Facebook X LinkedIn More