Improving Conversational Recommendation with Contextual Adaptation of External Recommenders and LLM-Based Reranking

Chuang Li, Weida Liang, Hengchang Hu, See-Kiong Ng, Min-Yen Kan, Haizhou Li, Yang Deng

March 2026

CARE uses conversational, contextual, and entity-level inputs for recommendation.

Abstract

CARE integrates external recommender systems with LLMs for conversational recommendation. External recommenders provide entity-level candidate items while LLMs perform contextual reranking, improving recommendation accuracy on ReDial and INSPIRED.

Type

Conference paper

Publication

Proceedings of the 48th European Conference on Information Retrieval (ECIR 2026)

Improving Conversational Recommendation with Contextual Adaptation of External Recommenders and LLM-Based Reranking

Abstract

Chuang Li

ISEP Doctoral Alumnus (Mar ‘25). Thesis: Towards Holistic and Practical Conversational Recommender Systems.

Hengchang Hu

Doctoral Alumnus (Oct ‘24). Thesis: Going beyond ID-based Recommender Systems by Exploiting External Knowledge.

Min-Yen Kan

Associate Professor