Multi-Step Reasoning Over Unstructured Text with Beam Dense Retrieval
This addresses the problem of complex question answering requiring multi-step reasoning over unstructured text for researchers and practitioners in NLP, representing an incremental improvement by adapting dense retrieval methods.
The paper tackles multi-hop question answering by proposing BeamDR, a multi-step retrieval approach that iteratively forms evidence chains through beam search in dense representations, achieving competitive performance with state-of-the-art systems without relying on semi-structured information.
Complex question answering often requires finding a reasoning chain that consists of multiple evidence pieces. Current approaches incorporate the strengths of structured knowledge and unstructured text, assuming text corpora is semi-structured. Building on dense retrieval methods, we propose a new multi-step retrieval approach (BeamDR) that iteratively forms an evidence chain through beam search in dense representations. When evaluated on multi-hop question answering, BeamDR is competitive to state-of-the-art systems, without using any semi-structured information. Through query composition in dense space, BeamDR captures the implicit relationships between evidence in the reasoning chain. The code is available at https://github.com/ henryzhao5852/BeamDR.