Bayesian Optimization with LLM-Based Acquisition Functions for Natural Language Preference Elicitation

arXiv:2405.00981v1 Announce Type: cross
Abstract: Designing preference elicitation (PE) methodologies that can quickly ascertain a user's top item preferences in a cold-start setting is a key challenge for building effective and personalized conversational recommendation (ConvRec) systems. While large language models (LLMs) constitute a novel technology that enables fully natural language (NL) PE dialogues, we hypothesize that monolithic LLM NL-PE approaches lack the multi-turn, decision-theoretic reasoning required to effectively balance the NL exploration and exploitation of user preferences towards an arbitrary …

abstract acquisition arxiv bayesian building challenge conversational cs.ai cs.cl designing functions key language language models large language large language models llm llms natural natural language novel optimization personalized recommendation systems technology type while