Making Alice Appear Like Bob: A Probabilistic Preference Obfuscation Method For Implicit Feedback Recommendation Models

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Escobedo, Gustavo, Moscati, Marta, Muellner, Peter, Kopeinik, Simone, Kowald, Dominik, Lex, Elisabeth, Schedl, Markus
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866913531429912576
author Escobedo, Gustavo
Moscati, Marta
Muellner, Peter
Kopeinik, Simone
Kowald, Dominik
Lex, Elisabeth
Schedl, Markus
author_facet Escobedo, Gustavo
Moscati, Marta
Muellner, Peter
Kopeinik, Simone
Kowald, Dominik
Lex, Elisabeth
Schedl, Markus
contents Users' interaction or preference data used in recommender systems carry the risk of unintentionally revealing users' private attributes (e.g., gender or race). This risk becomes particularly concerning when the training data contains user preferences that can be used to infer these attributes, especially if they align with common stereotypes. This major privacy issue allows malicious attackers or other third parties to infer users' protected attributes. Previous efforts to address this issue have added or removed parts of users' preferences prior to or during model training to improve privacy, which often leads to decreases in recommendation accuracy. In this work, we introduce SBO, a novel probabilistic obfuscation method for user preference data designed to improve the accuracy--privacy trade-off for such recommendation scenarios. We apply SBO to three state-of-the-art recommendation models (i.e., BPR, MultVAE, and LightGCN) and two popular datasets (i.e., MovieLens-1M and LFM-2B). Our experiments reveal that SBO outperforms comparable approaches with respect to the accuracy--privacy trade-off. Specifically, we can reduce the leakage of users' protected attributes while maintaining on-par recommendation accuracy.
format Preprint
id arxiv_https___arxiv_org_abs_2406_11505
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Making Alice Appear Like Bob: A Probabilistic Preference Obfuscation Method For Implicit Feedback Recommendation Models
Escobedo, Gustavo
Moscati, Marta
Muellner, Peter
Kopeinik, Simone
Kowald, Dominik
Lex, Elisabeth
Schedl, Markus
Information Retrieval
Users' interaction or preference data used in recommender systems carry the risk of unintentionally revealing users' private attributes (e.g., gender or race). This risk becomes particularly concerning when the training data contains user preferences that can be used to infer these attributes, especially if they align with common stereotypes. This major privacy issue allows malicious attackers or other third parties to infer users' protected attributes. Previous efforts to address this issue have added or removed parts of users' preferences prior to or during model training to improve privacy, which often leads to decreases in recommendation accuracy. In this work, we introduce SBO, a novel probabilistic obfuscation method for user preference data designed to improve the accuracy--privacy trade-off for such recommendation scenarios. We apply SBO to three state-of-the-art recommendation models (i.e., BPR, MultVAE, and LightGCN) and two popular datasets (i.e., MovieLens-1M and LFM-2B). Our experiments reveal that SBO outperforms comparable approaches with respect to the accuracy--privacy trade-off. Specifically, we can reduce the leakage of users' protected attributes while maintaining on-par recommendation accuracy.
title Making Alice Appear Like Bob: A Probabilistic Preference Obfuscation Method For Implicit Feedback Recommendation Models
topic Information Retrieval
url https://arxiv.org/abs/2406.11505