Adversarial Hubness in Multi-Modal Retrieval

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Tingwei, Suya, Fnu, Jha, Rishi, Zhang, Collin, Shmatikov, Vitaly
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914565931925504
author Zhang, Tingwei
Suya, Fnu
Jha, Rishi
Zhang, Collin
Shmatikov, Vitaly
author_facet Zhang, Tingwei
Suya, Fnu
Jha, Rishi
Zhang, Collin
Shmatikov, Vitaly
contents Hubness is a phenomenon in high-dimensional vector spaces where a point from the natural distribution is unusually close to many other points. This is a well-known problem in information retrieval that causes some items to accidentally (and incorrectly) appear relevant to many queries. In this paper, we investigate how attackers can exploit hubness to turn any image or audio input in a multi-modal retrieval system into an adversarial hub. Adversarial hubs can be used to inject universal adversarial content (e.g., spam) that will be retrieved in response to thousands of different queries, and also for targeted attacks on queries related to specific, attacker-chosen concepts. We present a method for creating adversarial hubs and evaluate the resulting hubs on benchmark multi-modal retrieval datasets and an image-to-image retrieval system implemented by Pinecone, a popular vector database. For example, in text-caption-to-image retrieval, a single adversarial hub, generated using 100 random queries, is retrieved as the top-1 most relevant image for more than 21,000 out of 25,000 test queries (by contrast, the most common natural hub is the top-1 response to only 102 queries), demonstrating the strong generalization capabilities of adversarial hubs. We also investigate whether techniques for mitigating natural hubness can also mitigate adversarial hubs, and show that they are not effective against hubs that target queries related to specific concepts.
format Preprint
id arxiv_https___arxiv_org_abs_2412_14113
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Adversarial Hubness in Multi-Modal Retrieval
Zhang, Tingwei
Suya, Fnu
Jha, Rishi
Zhang, Collin
Shmatikov, Vitaly
Cryptography and Security
Information Retrieval
Hubness is a phenomenon in high-dimensional vector spaces where a point from the natural distribution is unusually close to many other points. This is a well-known problem in information retrieval that causes some items to accidentally (and incorrectly) appear relevant to many queries. In this paper, we investigate how attackers can exploit hubness to turn any image or audio input in a multi-modal retrieval system into an adversarial hub. Adversarial hubs can be used to inject universal adversarial content (e.g., spam) that will be retrieved in response to thousands of different queries, and also for targeted attacks on queries related to specific, attacker-chosen concepts. We present a method for creating adversarial hubs and evaluate the resulting hubs on benchmark multi-modal retrieval datasets and an image-to-image retrieval system implemented by Pinecone, a popular vector database. For example, in text-caption-to-image retrieval, a single adversarial hub, generated using 100 random queries, is retrieved as the top-1 most relevant image for more than 21,000 out of 25,000 test queries (by contrast, the most common natural hub is the top-1 response to only 102 queries), demonstrating the strong generalization capabilities of adversarial hubs. We also investigate whether techniques for mitigating natural hubness can also mitigate adversarial hubs, and show that they are not effective against hubs that target queries related to specific concepts.
title Adversarial Hubness in Multi-Modal Retrieval
topic Cryptography and Security
Information Retrieval
url https://arxiv.org/abs/2412.14113