P2L-CA: An Effective Parameter Tuning Framework for Rehearsal-Free Multi-Label Class-Incremental Learning

TL;DR

P2L-CA is a novel, parameter-efficient framework for multi-label class-incremental learning that improves recognition accuracy and domain adaptation without extensive fine-tuning or memory buffers.

Contribution

It introduces a Prompt-to-Label module with class-specific prompts and linguistic priors, along with a lightweight Continuous Adapter, to enhance multi-label incremental learning.

Findings

Outperforms state-of-the-art methods on MS-COCO and PASCAL VOC

Requires minimal trainable parameters and no memory buffers

Demonstrates strong generalization in CIL scenarios

Abstract

Multi-label Class-Incremental Learning aims to continuously recognize novel categories in complex scenes where multiple objects co-occur. However, existing approaches often incur high computational costs due to full-parameter fine-tuning and substantial storage overhead from memory buffers, or they struggle to address feature confusion and domain discrepancies adequately. To overcome these limitations, we introduce P2L-CA, a parameter-efficient framework that integrates a Prompt-to-Label module with a Continuous Adapter module. The P2L module leverages class-specific prompts to disentangle multi-label representations while incorporating linguistic priors to enforce stable semantic-visual alignment. Meanwhile, the CA module employs lightweight adapters to mitigate domain gaps between pre-trained models and downstream tasks, thereby enhancing model plasticity. Extensive experiments across standard and challenging MLCIL settings on MS-COCO and PASCAL VOC show that P2L-CA not only achieves substantial improvements over state-of-the-art methods but also demonstrates strong generalization in CIL scenarios, all while requiring minimal trainable parameters and eliminating the need for memory buffers.