7th International Conference on Spoken Language Processing

September 16-20, 2002
Denver, Colorado, USA

An Efficient Dialogue Control Method Using Decision Tree-Based Estimation of Out-of-Vocabulary Word Attributes

Yasuhiro Takahashi, Kohji Dohsaka, Kiyoaki Aikawa

NTT Corporation, Japan

This paper proposes a dialogue control method for completing a task with a short dialogue even when user utterances include out-of-vocabulary (OOV) words. When the user utterance includes an OOV word, conventional methods try to complete a task by acquiring an understanding of the OOV word through dialogue. However, completing a task does not always require that the system understands the OOV word through dialogue. Our method eliminates the dialogue for acquiring an understanding of the OOV word and tries to complete a task by estimating attributes of OOV words using a decision tree. It does not allow system utterances using those attributes, which stops the user from uttering the same OOV words. A dialogue simulation indicates that our method can complete a task with dialogues that are shorter than those conventional methods need for completing a task.

Full Paper

Bibliographic reference.  Takahashi, Yasuhiro / Dohsaka, Kohji / Aikawa, Kiyoaki (2002): "An efficient dialogue control method using decision tree-based estimation of out-of-vocabulary word attributes", In ICSLP-2002, 813-816.