Adversarial Semantic Collisions
Lists
Only you see this
Abstract
We study semantic collisions: texts that are semantically unrelated but judged as similar by NLP models. We develop gradient-based approaches for generating semantic collisions and demonstrate that state-of-the-art models for many tasks which rely on analyzing the meaning and similarity of texts-- including paraphrase identification, document retrieval, response suggestion, and extractive summarization-- are vulnerable to semantic collisions. For example, given a target query, inserting a crafted collision into an irrelevant document can shift its retrieval rank from 1000 to top 3. We show how to generate semantic collisions that evade perplexity-based filtering and discuss other potential mitigations. Our code is available at https://github.com/csong27/collision-bert.
BibTeX
@inproceedings{song2020adversarial,
author = {Congzheng Song and Alexander M. Rush and Vitaly Shmatikov},
title = {{Adversarial Semantic Collisions}},
booktitle = {Conference on Empirical Methods in Natural Language Processing},
year = {2020},
doi = {10.18653/v1/2020.emnlp-main.344},
}