Cognitive Personalized Search Integrating Large Language Models with an Efficient Memory Mechanism

TL;DR

CoPS tackles data sparsity and personalization challenges in search by fusing large language models with an external cognitive memory mechanism that mirrors human memory (sensory, working, long-term). The framework computes a personalized score using $p(d|q,H)=\mathcal{R}(U_{q,H},d)$ by leveraging query rewriting, profile retrieval, and user modeling, with three complementary rankers (term-based, vector-based, and LLM-based). It demonstrates strong zero-shot performance on AOL and a large commercial dataset, approaching the effectiveness of fine-tuned personalized models while emphasizing privacy and efficiency through memory-driven retrieval and local deployment. The work suggests practical implications for privacy-preserving, user-centric information retrieval and points to future improvements in personalized GPT tuning and more capable LLMs.

Abstract

Traditional search engines usually provide identical search results for all users, overlooking individual preferences. To counter this limitation, personalized search has been developed to re-rank results based on user preferences derived from query logs. Deep learning-based personalized search methods have shown promise, but they rely heavily on abundant training data, making them susceptible to data sparsity challenges. This paper proposes a Cognitive Personalized Search (CoPS) model, which integrates Large Language Models (LLMs) with a cognitive memory mechanism inspired by human cognition. CoPS employs LLMs to enhance user modeling and user search experience. The cognitive memory mechanism comprises sensory memory for quick sensory responses, working memory for sophisticated cognitive responses, and long-term memory for storing historical interactions. CoPS handles new queries using a three-step approach: identifying re-finding behaviors, constructing user profiles with relevant historical information, and ranking documents based on personalized query intent. Experiments show that CoPS outperforms baseline models in zero-shot scenarios.

Figures (5)

• Figure 1: The memory mechanism of the human brain.
• Figure 2: The overview of CoPS. The system initially engages the sensory memory to identify re-finding behaviors, thus generating a sensory response if identified. Otherwise, the working memory collaborates with an LLM to accumulate personalized cues related to the query. After user modeling, a ranker is employed to re-rank the results based on user interests.
• Figure 3: Performance of different history lengths.
• Figure 4: Performance on different query sets.
• Figure 5: Analysis of fine-tuning and inference efficiency on different models.