QAEncoder: Towards Aligned Representation Learning in Question Answering Systems
This addresses a bottleneck in QA systems for users needing accurate retrieval, though it appears incremental as a novel method for a known problem.
The paper tackles the gap between user queries and relevant documents in retrieval-augmented generation QA systems by introducing QAEncoder, a training-free method that estimates query expectations in embedding space and uses document fingerprints, achieving alignment across diverse datasets and models with zero additional storage or latency.
Modern QA systems entail retrieval-augmented generation (RAG) for accurate and trustworthy responses. However, the inherent gap between user queries and relevant documents hinders precise matching. We introduce QAEncoder, a training-free approach to bridge this gap. Specifically, QAEncoder estimates the expectation of potential queries in the embedding space as a robust surrogate for the document embedding, and attaches document fingerprints to effectively distinguish these embeddings. Extensive experiments across diverse datasets, languages, and embedding models confirmed QAEncoder's alignment capability, which offers a simple-yet-effective solution with zero additional index storage, retrieval latency, training costs, or catastrophic forgetting and hallucination issues. The repository is publicly available at https://github.com/IAAR-Shanghai/QAEncoder.