Zero-Shot Learning for Natural Language Understanding Using Domain-Independent Sequential Structure and Question Types

Kugatsu Sadamitsu, Yukinori Homma, Ryuichiro Higashinaka, Yoshihiro Matsuo

Natural language understanding (NLU) is an important module of spoken dialogue systems. One of the difficulties when it comes to adapting NLU to new domains is the high cost of constructing new training data for each domain. To reduce this cost, we propose a zero-shot learning of NLU that takes into account the sequential structures of sentences together with general question types across different domains. Experimental results show that our methods achieve higher accuracy than baseline methods in two completely different domains (insurance and sightseeing).

doi: 10.21437/Interspeech.2017-638

Cite as: Sadamitsu, K., Homma, Y., Higashinaka, R., Matsuo, Y. (2017) Zero-Shot Learning for Natural Language Understanding Using Domain-Independent Sequential Structure and Question Types. Proc. Interspeech 2017, 3306-3310, doi: 10.21437/Interspeech.2017-638

@inproceedings{sadamitsu17_interspeech,
  author={Kugatsu Sadamitsu and Yukinori Homma and Ryuichiro Higashinaka and Yoshihiro Matsuo},
  title={{Zero-Shot Learning for Natural Language Understanding Using Domain-Independent Sequential Structure and Question Types}},
  year=2017,
  booktitle={Proc. Interspeech 2017},
  pages={3306--3310},
  doi={10.21437/Interspeech.2017-638}
}