Beyond Prompting: An Efficient Embedding Framework for Open-Domain Question Answering
This work addresses efficiency and performance issues in open-domain question answering for users of large language models, representing an incremental improvement over existing retriever-reader pipelines.
The paper tackled the problem of high computational overhead and suboptimal retrieval in open-domain question answering by proposing EmbQA, an embedding-level framework that refines query representations and introduces exploratory embeddings, achieving substantial improvements in accuracy and efficiency across multiple benchmarks.
Large language models have recently pushed open domain question answering (ODQA) to new frontiers. However, prevailing retriever-reader pipelines often depend on multiple rounds of prompt level instructions, leading to high computational overhead, instability, and suboptimal retrieval coverage. In this paper, we propose EmbQA, an embedding-level framework that alleviates these shortcomings by enhancing both the retriever and the reader. Specifically, we refine query representations via lightweight linear layers under an unsupervised contrastive learning objective, thereby reordering retrieved passages to highlight those most likely to contain correct answers. Additionally, we introduce an exploratory embedding that broadens the model's latent semantic space to diversify candidate generation and employs an entropy-based selection mechanism to choose the most confident answer automatically. Extensive experiments across three open-source LLMs, three retrieval methods, and four ODQA benchmarks demonstrate that EmbQA substantially outperforms recent baselines in both accuracy and efficiency.