KnowLA: Enhancing Parameter-efficient Finetuning with Knowledgeable Adaptation
This work addresses the problem of efficiently adapting LLMs to downstream tasks for AI practitioners, representing an incremental improvement by combining existing techniques like LoRA with knowledge integration.
The paper tackled improving parameter-efficient finetuning (PEFT) for large language models by integrating knowledge graph embeddings, resulting in enhanced effectiveness and robustness across six benchmarks with two LLMs and three knowledge graphs.
Parameter-efficient finetuning (PEFT) is a key technique for adapting large language models (LLMs) to downstream tasks. In this paper, we study leveraging knowledge graph embeddings to improve the effectiveness of PEFT. We propose a knowledgeable adaptation method called KnowLA. It inserts an adaptation layer into an LLM to integrate the embeddings of entities appearing in the input text. The adaptation layer is trained in combination with LoRA on instruction data. Experiments on six benchmarks with two popular LLMs and three knowledge graphs demonstrate the effectiveness and robustness of KnowLA. We show that \modelname can help activate the relevant parameterized knowledge in an LLM to answer a question without changing its parameters or input prompts.