Continual Prompt Tuning for Dialog State Tracking

TL;DR

This paper introduces Continual Prompt Tuning, a parameter-efficient method for dialog state tracking that prevents forgetting and promotes knowledge transfer across tasks by learning prompt embeddings while freezing the main model.

Contribution

It proposes a novel continual learning framework using prompt tuning with techniques for knowledge transfer and memory replay, addressing catastrophic forgetting in dialog systems.

Findings

Outperforms state-of-the-art baselines in continual dialog state tracking

Effectively prevents catastrophic forgetting

Enables knowledge transfer between tasks

Abstract

A desirable dialog system should be able to continually learn new skills without forgetting old ones, and thereby adapt to new domains or tasks in its life cycle. However, continually training a model often leads to a well-known catastrophic forgetting issue. In this paper, we present Continual Prompt Tuning, a parameter-efficient framework that not only avoids forgetting but also enables knowledge transfer between tasks. To avoid forgetting, we only learn and store a few prompt tokens' embeddings for each task while freezing the backbone pre-trained model. To achieve bi-directional knowledge transfer among tasks, we propose several techniques (continual prompt initialization, query fusion, and memory replay) to transfer knowledge from preceding tasks and a memory-guided technique to transfer knowledge from subsequent tasks. Extensive experiments demonstrate the effectiveness and efficiency of our proposed method on continual learning for dialog state tracking, compared with state-of-the-art baselines.