Pre-training With Large Language Model-based Document Expansion For Dense Passage Retrieval

Ma Guangyuan, Wu Xing, Wang Peng, Lin Zijia, Hu Songlin. Arxiv 2023

[Paper]    
Pretraining Methods RAG Training Techniques

In this paper, we systematically study the potential of pre-training with Large Language Model(LLM)-based document expansion for dense passage retrieval. Concretely, we leverage the capabilities of LLMs for document expansion, i.e. query generation, and effectively transfer expanded knowledge to retrievers using pre-training strategies tailored for passage retrieval. These strategies include contrastive learning and bottlenecked query generation. Furthermore, we incorporate a curriculum learning strategy to reduce the reliance on LLM inferences. Experimental results demonstrate that pre-training with LLM-based document expansion significantly boosts the retrieval performance on large-scale web-search tasks. Our work shows strong zero-shot and out-of-domain retrieval abilities, making it more widely applicable for retrieval when initializing with no human-labeled data.

Similar Work