Joint Passage Ranking for Diverse Multi-Answer Retrieval
This addresses the under-explored problem of retrieving diverse passages for multi-answer questions, which is incremental as it builds on prior retrieval methods but introduces a novel joint modeling approach.
The paper tackles the problem of multi-answer retrieval, where passages must cover multiple distinct answers for a question, by introducing JPR, a joint passage retrieval model that uses an autoregressive reranker and tree-decoding for diversity. JPR achieves significantly better answer coverage on three datasets and, when combined with question answering, enables larger models with fewer passages, establishing a new state-of-the-art.
We study multi-answer retrieval, an under-explored problem that requires retrieving passages to cover multiple distinct answers for a given question. This task requires joint modeling of retrieved passages, as models should not repeatedly retrieve passages containing the same answer at the cost of missing a different valid answer. In this paper, we introduce JPR, the first joint passage retrieval model for multi-answer retrieval. JPR makes use of an autoregressive reranker that selects a sequence of passages, each conditioned on previously selected passages. JPR is trained to select passages that cover new answers at each timestep and uses a tree-decoding algorithm to enable flexibility in the degree of diversity. Compared to prior approaches, JPR achieves significantly better answer coverage on three multi-answer datasets. When combined with downstream question answering, the improved retrieval enables larger answer generation models since they need to consider fewer passages, establishing a new state-of-the-art.