Brief

Researchers have developed SCI-Defense, a novel framework designed to counter manipulation attacks targeting LLM-based ranking systems. These attacks, known as Generative Engine Optimization (GEO), involve adversaries injecting misleading signals into product descriptions to artificially inflate their rankings. SCI-Defense integrates Perplexity detection, Semantic Integrity Scoring, and Inter-Candidate Detection to identify and block these manipulations. Evaluations on Amazon product descriptions and MS MARCO web passages demonstrated SCI-Defense's high precision and recall against various attack types, outperforming existing defense mechanisms. AI

Researchers have developed a new method called Layer-wise Token Compression (LTC) to improve the efficiency of transformer-based document reranking models used in information retrieval. Unlike previous token compression techniques that only applied to the initial embedding layer, LTC adapts token pooling at intermediate transformer layers. This approach has shown significant speedups, increasing inference queries per second by up to 25% for passage ranking and 116% for document ranking, while maintaining ranking quality. The method is also adaptable to long-context listwise reranking and may even act as a beneficial regularizer for long-document tasks. AI