In-context Continual Learning Assisted by an External Continual Learner

TL;DR

This paper introduces InCA, a scalable continual learning method that combines an external learner with in-context learning, effectively preventing catastrophic forgetting and managing prompt length issues in large language models.

Contribution

InCA is the first approach to integrate an external continual learner with in-context learning, enabling scalable continual learning without parameter updates or catastrophic forgetting.

Findings

InCA outperforms existing continual learning methods in experiments.

InCA maintains high performance with shorter prompts.

InCA effectively prevents catastrophic forgetting.

Abstract

Existing continual learning (CL) methods mainly rely on fine-tuning or adapting large language models (LLMs). They still suffer from catastrophic forgetting (CF). Little work has been done to exploit in-context learning (ICL) to leverage the extensive knowledge within LLMs for CL without updating any parameters. However, incrementally learning each new task in ICL necessitates adding training examples from each class of the task to the prompt, which hampers scalability as the prompt length increases. This issue not only leads to excessively long prompts that exceed the input token limit of the underlying LLM but also degrades the model's performance due to the overextended context. To address this, we introduce InCA, a novel approach that integrates an external continual learner (ECL) with ICL to enable scalable CL without CF. The ECL is built incrementally to pre-select a small subset of likely classes for each test instance. By restricting the ICL prompt to only these selected classes, InCA prevents prompt lengths from becoming excessively long, while maintaining high performance. Experimental results demonstrate that InCA significantly outperforms existing CL baselines, achieving substantial performance gains.