Mitigating Interference in the Knowledge Continuum through Attention-Guided Incremental Learning

TL;DR

This paper introduces AGILE, a novel rehearsal-based continual learning method that uses attention-guided mechanisms to reduce interference between tasks, thereby improving performance and scalability in class-incremental learning scenarios.

Contribution

AGILE is the first to incorporate compact task attention with learnable projection vectors to effectively mitigate task interference in rehearsal-based continual learning.

Findings

AGILE outperforms existing rehearsal methods in multiple CL scenarios.

It scales efficiently to many tasks with minimal overhead.

AGILE reduces task-recency bias and improves generalization.

Abstract

Continual learning (CL) remains a significant challenge for deep neural networks, as it is prone to forgetting previously acquired knowledge. Several approaches have been proposed in the literature, such as experience rehearsal, regularization, and parameter isolation, to address this problem. Although almost zero forgetting can be achieved in task-incremental learning, class-incremental learning remains highly challenging due to the problem of inter-task class separation. Limited access to previous task data makes it difficult to discriminate between classes of current and previous tasks. To address this issue, we propose `Attention-Guided Incremental Learning' (AGILE), a novel rehearsal-based CL approach that incorporates compact task attention to effectively reduce interference between tasks. AGILE utilizes lightweight, learnable task projection vectors to transform the latent representations of a shared task attention module toward task distribution. Through extensive empirical evaluation, we show that AGILE significantly improves generalization performance by mitigating task interference and outperforming rehearsal-based approaches in several CL scenarios. Furthermore, AGILE can scale well to a large number of tasks with minimal overhead while remaining well-calibrated with reduced task-recency bias.