Learning Personalized User Preference from Cold Start in Multi-turn Conversations

TL;DR

This paper introduces TAI, a system that learns user preferences from cold start in multi-turn conversations by leveraging BERT models, achieving high accuracy and being adopted in production.

Contribution

The paper presents a novel teachable conversation system that automatically learns and adapts to user preferences during live interactions using advanced NLP models.

Findings

Achieves 0.9122 turn-level accuracy on out-of-sample data

Successfully adopted in production environment

Effectively manages dialogue flows for preference learning

Abstract

This paper presents a novel teachable conversation interaction system that is capable of learning users preferences from cold start by gradually adapting to personal preferences. In particular, the TAI system is able to automatically identify and label user preference in live interactions, manage dialogue flows for interactive teaching sessions, and reuse learned preference for preference elicitation. We develop the TAI system by leveraging BERT encoder models to encode both dialogue and relevant context information, and build action prediction (AP), argument filling (AF) and named entity recognition (NER) models to understand the teaching session. We adopt a seeker-provider interaction loop mechanism to generate diverse dialogues from cold-start. TAI is capable of learning user preference, which achieves 0.9122 turn level accuracy on out-of-sample dataset, and has been successfully adopted in production.