Multi-Domain Multi-Task Rehearsal for Lifelong Learning

TL;DR

This paper introduces a multi-domain multi-task rehearsal method for lifelong learning that mitigates catastrophic forgetting by addressing domain shift and task isolation through parallel training, angular margin loss, and episodic distillation.

Contribution

It proposes a novel multi-domain multi-task rehearsal framework with a two-level angular margin loss and episodic distillation to better preserve old knowledge during lifelong learning.

Findings

Effective mitigation of domain shift in lifelong learning.

Improved retention of old task knowledge compared to baseline methods.

Validated on benchmark datasets with positive results.

Abstract

Rehearsal, seeking to remind the model by storing old knowledge in lifelong learning, is one of the most effective ways to mitigate catastrophic forgetting, i.e., biased forgetting of previous knowledge when moving to new tasks. However, the old tasks of the most previous rehearsal-based methods suffer from the unpredictable domain shift when training the new task. This is because these methods always ignore two significant factors. First, the Data Imbalance between the new task and old tasks that makes the domain of old tasks prone to shift. Second, the Task Isolation among all tasks will make the domain shift toward unpredictable directions; To address the unpredictable domain shift, in this paper, we propose Multi-Domain Multi-Task (MDMT) rehearsal to train the old tasks and new task parallelly and equally to break the isolation among tasks. Specifically, a two-level angular margin loss is proposed to encourage the intra-class/task compactness and inter-class/task discrepancy, which keeps the model from domain chaos. In addition, to further address domain shift of the old tasks, we propose an optional episodic distillation loss on the memory to anchor the knowledge for each old task. Experiments on benchmark datasets validate the proposed approach can effectively mitigate the unpredictable domain shift.