TY - GEN
T1 - A non-Task-oriented dialogue system controlling the utterance length
AU - Isoshima, Kazuki
AU - Hagiwara, Masafumi
PY - 2019/5/15
Y1 - 2019/5/15
N2 - In this paper, we propose a non-task-oriented dialogue system controlling the utterance length. The dialogue system can be classified into a task-oriented dialogue system or a non-task-oriented dialogue system. Recently, demand for the non-task-oriented dialogue system is increasing. The utterance length is one of the important information in a dialogue system. In general, our utterance length tends to be long when we are speakers. On the other hand, the length of our utterance tends to be short when we are listeners. In addition, the utterance length differs from person to person, so we change our utterance length for friendly communication. The effect of the utterance length has never considered in dialogue systems using Encoder-Decoder model. Therefore, we propose an utterance length estimator (ULE) and an index of the utterance length. ULE is a neural network which learns the utterance length by training data of dialogue. The index of the utterance length is the parameter considers user's personality and it is calculated during dialogue. Our dialogue system decides the length of system's utterance by ULE and index of the utterance length, and generates output sequences by using a neural encoder-decoder controlling output length. Experimental results show our system can decide the appropriate length of the utterance and makes users more satisfied than the conventional method.
AB - In this paper, we propose a non-task-oriented dialogue system controlling the utterance length. The dialogue system can be classified into a task-oriented dialogue system or a non-task-oriented dialogue system. Recently, demand for the non-task-oriented dialogue system is increasing. The utterance length is one of the important information in a dialogue system. In general, our utterance length tends to be long when we are speakers. On the other hand, the length of our utterance tends to be short when we are listeners. In addition, the utterance length differs from person to person, so we change our utterance length for friendly communication. The effect of the utterance length has never considered in dialogue systems using Encoder-Decoder model. Therefore, we propose an utterance length estimator (ULE) and an index of the utterance length. ULE is a neural network which learns the utterance length by training data of dialogue. The index of the utterance length is the parameter considers user's personality and it is calculated during dialogue. Our dialogue system decides the length of system's utterance by ULE and index of the utterance length, and generates output sequences by using a neural encoder-decoder controlling output length. Experimental results show our system can decide the appropriate length of the utterance and makes users more satisfied than the conventional method.
KW - Dialogue System
KW - Encoder-decoder model
KW - Neural network
KW - The utterance length
UR - http://www.scopus.com/inward/record.url?scp=85067119358&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85067119358&partnerID=8YFLogxK
U2 - 10.1109/SCIS-ISIS.2018.00140
DO - 10.1109/SCIS-ISIS.2018.00140
M3 - Conference contribution
AN - SCOPUS:85067119358
T3 - Proceedings - 2018 Joint 10th International Conference on Soft Computing and Intelligent Systems and 19th International Symposium on Advanced Intelligent Systems, SCIS-ISIS 2018
SP - 849
EP - 854
BT - Proceedings - 2018 Joint 10th International Conference on Soft Computing and Intelligent Systems and 19th International Symposium on Advanced Intelligent Systems, SCIS-ISIS 2018
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - Joint 10th International Conference on Soft Computing and Intelligent Systems and 19th International Symposium on Advanced Intelligent Systems, SCIS-ISIS 2018
Y2 - 5 December 2018 through 8 December 2018
ER -