1-PAGER: One Pass Answer Generation and Evidence Retrieval
This work addresses the need for efficient and interpretable attributed generation in NLP, though it is incremental as it builds on existing sequence-to-sequence paradigms without surpassing more expensive multi-document systems.
The paper tackles the problem of answering questions and retrieving evidence simultaneously by introducing 1-PAGER, a system that uses a single Transformer model to partition a corpus and generate answers in one pass, achieving competitive retrieval and answer accuracy compared to retrieve-and-read alternatives and outperforming closed-book models.
We present 1-Pager the first system that answers a question and retrieves evidence using a single Transformer-based model and decoding process. 1-Pager incrementally partitions the retrieval corpus using constrained decoding to select a document and answer string, and we show that this is competitive with comparable retrieve-and-read alternatives according to both retrieval and answer accuracy metrics. 1-Pager also outperforms the equivalent closed-book question answering model, by grounding predictions in an evidence corpus. While 1-Pager is not yet on-par with more expensive systems that read many more documents before generating an answer, we argue that it provides an important step toward attributed generation by folding retrieval into the sequence-to-sequence paradigm that is currently dominant in NLP. We also show that the search paths used to partition the corpus are easy to read and understand, paving a way forward for interpretable neural retrieval.