Hierarchical Inductive Transfer for Continual Dialogue Learning

TL;DR

This paper introduces a hierarchical inductive transfer framework that enhances continual dialogue learning by using adapter modules, reducing knowledge interference, and maintaining performance with minimal additional parameters.

Contribution

It proposes a novel hierarchical transfer method with adapter modules for efficient continual dialogue learning, addressing knowledge interference and capacity issues.

Findings

Achieves comparable performance with lower model capacity

Effectively alleviates knowledge interference among tasks

Enables continual learning with minimal parameter increase

Abstract

Pre-trained models have achieved excellent performance on the dialogue task. However, for the continual increase of online chit-chat scenarios, directly fine-tuning these models for each of the new tasks not only explodes the capacity of the dialogue system on the embedded devices but also causes knowledge forgetting on pre-trained models and knowledge interference among diverse dialogue tasks. In this work, we propose a hierarchical inductive transfer framework to learn and deploy the dialogue skills continually and efficiently. First, we introduce the adapter module into pre-trained models for learning new dialogue tasks. As the only trainable module, it is beneficial for the dialogue system on the embedded devices to acquire new dialogue skills with negligible additional parameters. Then, for alleviating knowledge interference between tasks yet benefiting the regularization between them, we further design hierarchical inductive transfer that enables new tasks to use general knowledge in the base adapter without being misled by diverse knowledge in task-specific adapters. Empirical evaluation and analysis indicate that our framework obtains comparable performance under deployment-friendly model capacity.