UP-NRPA: User Portrait based Nested Rollout Policy Adaptation for Planning with Large Language Models in Goal-oriented Dialogue Systems
Researchers have introduced UP-NRPA, a novel online framework designed to enhance dialogue policy planning in goal-oriented systems by adapting to individual user characteristics. Unlike traditional methods that rely on offline reinforcement learning for user groups, UP-NRPA dynamically customizes dialogue strategies using real-time user feedback, including personality and preferences. This adaptive approach achieved a 100% success rate on collaborative and non-collaborative dialogue benchmarks, significantly boosting the sale-to-list ratio by 56.41% in negotiation tasks. AI
IMPACT Enhances dialogue systems' ability to personalize interactions, potentially improving user satisfaction and task completion rates.