Prompt Customization for Continual Learning

TL;DR

This paper introduces a prompt customization approach for continual learning that generates and modulates prompts adaptively, improving performance over existing methods across multiple benchmarks.

Contribution

The paper proposes a novel prompt customization method with generation and modulation modules, addressing noise issues in prompt selection for continual learning.

Findings

Achieved up to 16.2% performance improvement over SOTA methods.

Effective across class, domain, and task-agnostic incremental learning settings.

Demonstrated consistent gains on four benchmark datasets.

Abstract

Contemporary continual learning approaches typically select prompts from a pool, which function as supplementary inputs to a pre-trained model. However, this strategy is hindered by the inherent noise of its selection approach when handling increasing tasks. In response to these challenges, we reformulate the prompting approach for continual learning and propose the prompt customization (PC) method. PC mainly comprises a prompt generation module (PGM) and a prompt modulation module (PMM). In contrast to conventional methods that employ hard prompt selection, PGM assigns different coefficients to prompts from a fixed-sized pool of prompts and generates tailored prompts. Moreover, PMM further modulates the prompts by adaptively assigning weights according to the correlations between input data and corresponding prompts. We evaluate our method on four benchmark datasets for three diverse settings, including the class, domain, and task-agnostic incremental learning tasks. Experimental results demonstrate consistent improvement (by up to 16.2\%), yielded by the proposed method, over the state-of-the-art (SOTA) techniques.