Compacting, Picking and Growing for Unforgetting Continual Learning

TL;DR

This paper introduces a scalable continual learning method that combines model compression, critical weight selection, and network expansion to prevent forgetting and maintain compactness across multiple tasks.

Contribution

It presents a novel incremental learning approach integrating compression, selection, and expansion, enabling scalable, compact, and effective continual learning.

Findings

Prevents forgetting of previous tasks

Maintains model compactness during learning

Achieves better performance than training tasks independently

Abstract

Continual lifelong learning is essential to many applications. In this paper, we propose a simple but effective approach to continual deep learning. Our approach leverages the principles of deep model compression, critical weights selection, and progressive networks expansion. By enforcing their integration in an iterative manner, we introduce an incremental learning method that is scalable to the number of sequential tasks in a continual learning process. Our approach is easy to implement and owns several favorable characteristics. First, it can avoid forgetting (i.e., learn new tasks while remembering all previous tasks). Second, it allows model expansion but can maintain the model compactness when handling sequential tasks. Besides, through our compaction and selection/expansion mechanism, we show that the knowledge accumulated through learning previous tasks is helpful to build a better model for the new tasks compared to training the models independently with tasks. Experimental results show that our approach can incrementally learn a deep model tackling multiple tasks without forgetting, while the model compactness is maintained with the performance more satisfiable than individual task training.