MoR: Better Handling Diverse Queries with a Mixture of Sparse, Dense, and Human Retrievers
This addresses the challenge of handling diverse queries in retrieval-augmented generation, offering a dynamic and efficient solution without manual selection.
The paper tackles the problem of fixed retriever selection in retrieval-augmented generation by proposing a zero-shot mixture of sparse, dense, and human retrievers, achieving average performance improvements of +10.8% over individual retrievers and +3.9% over larger models.
Retrieval-augmented Generation (RAG) is powerful, but its effectiveness hinges on which retrievers we use and how. Different retrievers offer distinct, often complementary signals: BM25 captures lexical matches; dense retrievers, semantic similarity. Yet in practice, we typically fix a single retriever based on heuristics, which fails to generalize across diverse information needs. Can we dynamically select and integrate multiple retrievers for each individual query, without the need for manual selection? In our work, we validate this intuition with quantitative analysis and introduce mixture of retrievers: a zero-shot, weighted combination of heterogeneous retrievers. Extensive experiments show that such mixtures are effective and efficient: Despite totaling just 0.8B parameters, this mixture outperforms every individual retriever and even larger 7B models by +10.8% and +3.9% on average, respectively. Further analysis also shows that this mixture framework can help incorporate specialized non-oracle human information sources as retrievers to achieve good collaboration, with a 58.9% relative performance improvement over simulated humans alone.