Adversarial Text Rewriting for Text-aware Recommender Systems

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Oh, Sejoon, Verma, Gaurav, Kumar, Srijan
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866916342335012864
author Oh, Sejoon
Verma, Gaurav
Kumar, Srijan
author_facet Oh, Sejoon
Verma, Gaurav
Kumar, Srijan
contents Text-aware recommender systems incorporate rich textual features, such as titles and descriptions, to generate item recommendations for users. The use of textual features helps mitigate cold-start problems, and thus, such recommender systems have attracted increased attention. However, we argue that the dependency on item descriptions makes the recommender system vulnerable to manipulation by adversarial sellers on e-commerce platforms. In this paper, we explore the possibility of such manipulation by proposing a new text rewriting framework to attack text-aware recommender systems. We show that the rewriting attack can be exploited by sellers to unfairly uprank their products, even though the adversarially rewritten descriptions are perceived as realistic by human evaluators. Methodologically, we investigate two different variations to carry out text rewriting attacks: (1) two-phase fine-tuning for greater attack performance, and (2) in-context learning for higher text rewriting quality. Experiments spanning 3 different datasets and 4 existing approaches demonstrate that recommender systems exhibit vulnerability against the proposed text rewriting attack. Our work adds to the existing literature around the robustness of recommender systems, while highlighting a new dimension of vulnerability in the age of large-scale automated text generation.
format Preprint
id arxiv_https___arxiv_org_abs_2408_00312
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Adversarial Text Rewriting for Text-aware Recommender Systems
Oh, Sejoon
Verma, Gaurav
Kumar, Srijan
Information Retrieval
Cryptography and Security
Machine Learning
Social and Information Networks
Text-aware recommender systems incorporate rich textual features, such as titles and descriptions, to generate item recommendations for users. The use of textual features helps mitigate cold-start problems, and thus, such recommender systems have attracted increased attention. However, we argue that the dependency on item descriptions makes the recommender system vulnerable to manipulation by adversarial sellers on e-commerce platforms. In this paper, we explore the possibility of such manipulation by proposing a new text rewriting framework to attack text-aware recommender systems. We show that the rewriting attack can be exploited by sellers to unfairly uprank their products, even though the adversarially rewritten descriptions are perceived as realistic by human evaluators. Methodologically, we investigate two different variations to carry out text rewriting attacks: (1) two-phase fine-tuning for greater attack performance, and (2) in-context learning for higher text rewriting quality. Experiments spanning 3 different datasets and 4 existing approaches demonstrate that recommender systems exhibit vulnerability against the proposed text rewriting attack. Our work adds to the existing literature around the robustness of recommender systems, while highlighting a new dimension of vulnerability in the age of large-scale automated text generation.
title Adversarial Text Rewriting for Text-aware Recommender Systems
topic Information Retrieval
Cryptography and Security
Machine Learning
Social and Information Networks
url https://arxiv.org/abs/2408.00312