Memory Replay with Data Compression for Continual Learning

TL;DR

This paper introduces a memory replay method with data compression for continual learning, effectively increasing stored old data and reducing memory costs while maintaining performance.

Contribution

It proposes a novel data compression approach using DPPs to optimize quality and quantity of stored samples, improving continual learning performance.

Findings

Significant performance boost over strong baselines.

Efficient data compression with optimized quality selection.

Validated across class-incremental learning and autonomous driving scenarios.

Abstract

Continual learning needs to overcome catastrophic forgetting of the past. Memory replay of representative old training samples has been shown as an effective solution, and achieves the state-of-the-art (SOTA) performance. However, existing work is mainly built on a small memory buffer containing a few original data, which cannot fully characterize the old data distribution. In this work, we propose memory replay with data compression (MRDC) to reduce the storage cost of old training samples and thus increase their amount that can be stored in the memory buffer. Observing that the trade-off between the quality and quantity of compressed data is highly nontrivial for the efficacy of memory replay, we propose a novel method based on determinantal point processes (DPPs) to efficiently determine an appropriate compression quality for currently-arrived training samples. In this way, using a naive data compression algorithm with a properly selected quality can largely boost recent strong baselines by saving more compressed data in a limited storage space. We extensively validate this across several benchmarks of class-incremental learning and in a realistic scenario of object detection for autonomous driving.