<ul><li>Retrieval-augmented generation (RAG) systems, which incorporate external knowledge, are vulnerable to corpus poisoning attacks.</li><li>Existing methods for crafting adversarial passages are slow and computationally expensive.</li><li>A new method called Dynamic Importance-Guided Genetic Algorithm (DIGA) is proposed to efficiently generate adversarial passages.</li><li>DIGA achieves superior efficiency and scalability, with comparable or better attack success rates.</li></ul>

Tricking Retrievers with Influential Tokens: An Efficient Black-Box Corpus Poisoning Attack

Discover more