PAQ: 65 Million Probably-Asked Questions and What You Can Do With Them

• Published in: Transactions of the Association for Computational Linguistics Vol. 9; pp. 1098 - 1115
• Main Authors: Lewis, Patrick, Wu, Yuxiang, Liu, Linqing, Minervini, Pasquale, Küttler, Heinrich, Piktus, Aleksandra, Stenetorp, Pontus, Riedel, Sebastian
• Format: Journal Article
• Language: English
• Published: One Rogers Street, Cambridge, MA 02142-1209, USA MIT Press 07.10.2021; MIT Press Journals, The; The MIT Press
• Subjects: Accuracy; Computational linguistics; Question answer sequences; Questions; Testing time
• ISSN: 2307-387X; 2307-387X
• DOI: 10.1162/tacl_a_00415

Open-domain Question Answering models that directly leverage question-answer (QA) pairs, such as closed-book QA (CBQA) models and QA-pair retrievers, show promise in terms of speed and memory compared with conventional models which retrieve and read from text corpora. QA-pair retrievers also offer interpretable answers, a high degree of control, and are trivial to update at test time with new knowledge. However, these models fall short of the accuracy of retrieve-and-read systems, as substantially less knowledge is covered by the available QA-pairs relative to text corpora like Wikipedia. To facilitate improved QA-pair models, we introduce (PAQ), a very large resource of 65M automatically generated QA-pairs. We introduce a new QA-pair retriever, RePAQ, to complement PAQ. We find that PAQ and test questions, enabling RePAQ to match the accuracy of recent retrieve-and-read models, whilst being significantly faster. Using PAQ, we train CBQA models which outperform comparable baselines by 5%, but trail RePAQ by over 15%, indicating the effectiveness of explicit retrieval. RePAQ can be configured for size (under 500MB) or speed (over 1K questions per second) while retaining high accuracy. Lastly, we demonstrate RePAQ’s strength at , abstaining from answering when it is likely to be incorrect. This enables RePAQ to “back-off” to a more expensive state-of-the-art model, leading to a combined system which is both more accurate and 2x faster than the state-of-the-art model alone.