FANDA: A Novel Approach to Perform Follow-Up Query Analysis


  • Qian Liu Beihang University
  • Bei Chen Microsoft Research
  • Jian-Guang Lou Microsoft
  • Ge Jin Peking University
  • Dongmei Zhang Microsoft Research



Recent work on Natural Language Interfaces to Databases (NLIDB) has attracted considerable attention. NLIDB allow users to search databases using natural language instead of SQL-like query languages. While saving the users from having to learn query languages, multi-turn interaction with NLIDB usually involves multiple queries where contextual information is vital to understand the users’ query intents. In this paper, we address a typical contextual understanding problem, termed as follow-up query analysis. In spite of its ubiquity, follow-up query analysis has not been well studied due to two primary obstacles: the multifarious nature of follow-up query scenarios and the lack of high-quality datasets. Our work summarizes typical follow-up query scenarios and provides a new FollowUp dataset with 1000 query triples on 120 tables. Moreover, we propose a novel approach FANDA, which takes into account the structures of queries and employs a ranking model with weakly supervised max-margin learning. The experimental results on FollowUp demonstrate the superiority of FANDA over multiple baselines across multiple metrics.




How to Cite

Liu, Q., Chen, B., Lou, J.-G., Jin, G., & Zhang, D. (2019). FANDA: A Novel Approach to Perform Follow-Up Query Analysis. Proceedings of the AAAI Conference on Artificial Intelligence, 33(01), 6770-6777.



AAAI Technical Track: Natural Language Processing