Private Counterfactual Retrieval With Immutable Features

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Meel, Shreya, Dissanayake, Pasan, Nomeir, Mohamed, Dutta, Sanghamitra, Ulukus, Sennur
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910700064997376
author Meel, Shreya
Dissanayake, Pasan
Nomeir, Mohamed
Dutta, Sanghamitra
Ulukus, Sennur
author_facet Meel, Shreya
Dissanayake, Pasan
Nomeir, Mohamed
Dutta, Sanghamitra
Ulukus, Sennur
contents In a classification task, counterfactual explanations provide the minimum change needed for an input to be classified into a favorable class. We consider the problem of privately retrieving the exact closest counterfactual from a database of accepted samples while enforcing that certain features of the input sample cannot be changed, i.e., they are \emph{immutable}. An applicant (user) whose feature vector is rejected by a machine learning model wants to retrieve the sample closest to them in the database without altering a private subset of their features, which constitutes the immutable set. While doing this, the user should keep their feature vector, immutable set and the resulting counterfactual index information-theoretically private from the institution. We refer to this as immutable private counterfactual retrieval (I-PCR) problem which generalizes PCR to a more practical setting. In this paper, we propose two I-PCR schemes by leveraging techniques from private information retrieval (PIR) and characterize their communication costs. Further, we quantify the information that the user learns about the database and compare it for the proposed schemes.
format Preprint
id arxiv_https___arxiv_org_abs_2411_10429
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Private Counterfactual Retrieval With Immutable Features
Meel, Shreya
Dissanayake, Pasan
Nomeir, Mohamed
Dutta, Sanghamitra
Ulukus, Sennur
Information Theory
Cryptography and Security
Machine Learning
Signal Processing
In a classification task, counterfactual explanations provide the minimum change needed for an input to be classified into a favorable class. We consider the problem of privately retrieving the exact closest counterfactual from a database of accepted samples while enforcing that certain features of the input sample cannot be changed, i.e., they are \emph{immutable}. An applicant (user) whose feature vector is rejected by a machine learning model wants to retrieve the sample closest to them in the database without altering a private subset of their features, which constitutes the immutable set. While doing this, the user should keep their feature vector, immutable set and the resulting counterfactual index information-theoretically private from the institution. We refer to this as immutable private counterfactual retrieval (I-PCR) problem which generalizes PCR to a more practical setting. In this paper, we propose two I-PCR schemes by leveraging techniques from private information retrieval (PIR) and characterize their communication costs. Further, we quantify the information that the user learns about the database and compare it for the proposed schemes.
title Private Counterfactual Retrieval With Immutable Features
topic Information Theory
Cryptography and Security
Machine Learning
Signal Processing
url https://arxiv.org/abs/2411.10429