Continual Adapter Tuning with Semantic Shift Compensation for Class-Incremental Learning

TL;DR

This paper introduces a parameter-efficient continual learning method using adapter tuning and prototype-based feature sampling, achieving state-of-the-art results without model expansion or sample retention.

Contribution

It proposes a novel incremental adapter tuning approach with semantic shift compensation, enhancing continual learning with pre-trained models.

Findings

Outperforms previous CIL methods on five benchmarks.

Achieves state-of-the-art performance without model expansion.

Effectively estimates semantic shift of prototypes without past samples.

Abstract

Class-incremental learning (CIL) aims to enable models to continuously learn new classes while overcoming catastrophic forgetting. The introduction of pre-trained models has brought new tuning paradigms to CIL. In this paper, we revisit different parameter-efficient tuning (PET) methods within the context of continual learning. We observe that adapter tuning demonstrates superiority over prompt-based methods, even without parameter expansion in each learning session. Motivated by this, we propose incrementally tuning the shared adapter without imposing parameter update constraints, enhancing the learning capacity of the backbone. Additionally, we employ feature sampling from stored prototypes to retrain a unified classifier, further improving its performance. We estimate the semantic shift of old prototypes without access to past samples and update stored prototypes session by session. Our proposed method eliminates model expansion and avoids retaining any image samples. It surpasses previous pre-trained model-based CIL methods and demonstrates remarkable continual learning capabilities. Experimental results on five CIL benchmarks validate the effectiveness of our approach, achieving state-of-the-art (SOTA) performance.