IAI MovieBot 2.0 aims to enhance user-facing experiments with trainable neural components, transparent user modeling, and improved research infrastructure.
The proposed KERL framework leverages knowledge graphs and pre-trained language models to enhance entity representation learning, enabling more informed recommendations and informative responses in conversational recommender systems.
SAPIENT, a novel framework for multi-turn conversational recommendation, leverages Monte Carlo Tree Search (MCTS) and a self-training loop to enable strategic and non-myopic conversational planning, outperforming state-of-the-art baselines.
Combining Large Language Models (LLMs) with Conversational Recommender Systems (CRSs) can significantly improve their individual performance in understanding and responding to user needs within e-commerce pre-sales dialogues.