Answering Unanswered Questions through Semantic Reformulations in Spoken QA
Faustini, Pedro, Chen, Zhiyu, Fetahu, Besnik, Rokhlenko, Oleg, Malmasi, Shervin
–arXiv.org Artificial Intelligence
Spoken Question Answering (QA) is a key feature of voice assistants, usually backed by multiple QA systems. Users ask questions via spontaneous speech which can contain disfluencies, errors, and informal syntax or phrasing. This is a major challenge in QA, causing unanswered questions or irrelevant answers, and leading to bad user experiences. We analyze failed QA requests to identify core challenges: lexical gaps, proposition types, complex syntactic structure, and high specificity. We propose a Semantic Question Reformulation (SURF) model offering three linguistically-grounded operations (repair, syntactic reshaping, generalization) to rewrite questions to facilitate answering. Offline evaluation on 1M unanswered questions from a leading voice assistant shows that SURF significantly improves answer rates: up to 24% of previously unanswered questions obtain relevant answers (75%). Live deployment shows positive impact for millions of customers with unanswered questions; explicit relevance feedback shows high user satisfaction.
arXiv.org Artificial Intelligence
Jun-3-2023
- Country:
- Oceania > Australia
- New South Wales > Sydney (0.14)
- North America
- Canada (0.04)
- United States
- New Jersey (0.04)
- Utah (0.04)
- Washington > King County
- Seattle (0.04)
- Pennsylvania
- Philadelphia County > Philadelphia (0.04)
- Allegheny County > Pittsburgh (0.04)
- Minnesota > Hennepin County
- Minneapolis (0.14)
- California
- Los Angeles County > Los Angeles (0.14)
- San Francisco County > San Francisco (0.04)
- Europe
- Germany (0.04)
- Ireland (0.04)
- United Kingdom > Scotland
- City of Aberdeen > Aberdeen (0.04)
- Italy > Tuscany
- Florence (0.04)
- France > Provence-Alpes-Côte d'Azur
- Bouches-du-Rhône > Marseille (0.04)
- Belgium > Brussels-Capital Region
- Brussels (0.04)
- Asia
- Myanmar > Tanintharyi Region
- Dawei (0.04)
- Middle East
- UAE > Abu Dhabi Emirate
- Abu Dhabi (0.14)
- Republic of Türkiye > Batman Province
- Batman (0.04)
- UAE > Abu Dhabi Emirate
- Japan > Hokkaidō
- Hokkaidō Prefecture > Sapporo (0.04)
- India > Maharashtra
- Mumbai (0.04)
- Myanmar > Tanintharyi Region
- Oceania > Australia
- Genre:
- Research Report > New Finding (0.93)
- Industry:
- Technology:
- Information Technology > Artificial Intelligence
- Machine Learning (1.00)
- Speech > Speech Recognition (0.89)
- Natural Language > Question Answering (0.88)
- Information Technology > Artificial Intelligence