Sub-network Discovery and Soft-masking for Continual Learning of Mixed Tasks

TL;DR

This paper introduces a novel continual learning method that isolates task knowledge through subnetwork discovery and employs soft-masking to prevent forgetting and promote knowledge transfer across mixed task types.

Contribution

It presents a new approach combining subnetwork discovery and soft-masking to effectively address both catastrophic forgetting and knowledge transfer in continual learning of mixed tasks.

Findings

Outperforms strong baselines across diverse tasks

Effectively prevents catastrophic forgetting

Enhances knowledge transfer in heterogeneous tasks

Abstract

Continual learning (CL) has two main objectives: preventing catastrophic forgetting (CF) and encouraging knowledge transfer (KT). The existing literature mainly focused on overcoming CF. Some work has also been done on KT when the tasks are similar. To our knowledge, only one method has been proposed to learn a sequence of mixed tasks. However, these techniques still suffer from CF and/or limited KT. This paper proposes a new CL method to achieve both. It overcomes CF by isolating the knowledge of each task via discovering a subnetwork for it. A soft-masking mechanism is also proposed to preserve the previous knowledge and to enable the new task to leverage the past knowledge to achieve KT. Experiments using classification, generation, information extraction, and their mixture (i.e., heterogeneous tasks) show that the proposed method consistently outperforms strong baselines.