Recurrent Knowledge Identification and Fusion for Language Model Continual Learning
This addresses the challenge of deploying LLMs in dynamic environments without costly retraining, though it appears incremental as it builds on existing model ensemble and merging methods.
The paper tackles the problem of catastrophic forgetting and knowledge transfer in continual learning for large language models by introducing Recurrent-KIF, a framework that dynamically estimates parameter importance distributions, resulting in effective mitigation of forgetting and enhanced transfer as demonstrated in experiments with models up to 13B parameters.
Continual learning (CL) is crucial for deploying large language models (LLMs) in dynamic real-world environments without costly retraining. While recent model ensemble and model merging methods guided by parameter importance have gained popularity, they often struggle to balance knowledge transfer and forgetting, mainly due to the reliance on static importance estimates during sequential training. In this paper, we present Recurrent-KIF, a novel CL framework for Recurrent Knowledge Identification and Fusion, which enables dynamic estimation of parameter importance distributions to enhance knowledge transfer. Inspired by human continual learning, Recurrent-KIF employs an inner loop that rapidly adapts to new tasks while identifying important parameters, coupled with an outer loop that globally manages the fusion of new and historical knowledge through redundant knowledge pruning and key knowledge merging. These inner-outer loops iteratively perform multiple rounds of fusion, allowing Recurrent-KIF to leverage intermediate training information and adaptively adjust fusion strategies based on evolving importance distributions. Extensive experiments on two CL benchmarks with various model sizes (from 770M to 13B) demonstrate that Recurrent-KIF effectively mitigates catastrophic forgetting and enhances knowledge transfer.