Attention-Based Dialog State Tracking for Conversational Interview Coaching

Ming Hsiang Su, Chung Hsien Wu, Kun Yi Huang, Chu Kwang Chen

研究成果: Conference contribution

1 引文 (Scopus)

摘要

This study proposes an approach to dialog state tracking (DST) in a conversational interview coaching system. For the interview coaching task, the semantic slots, used mostly in traditional dialog systems, are difficult to define manually. This study adopts the topic profile of the response from the interviewee as the dialog state representation. In addition, as the response generally consists of several sentences, the summary vector obtained from a long short-term memory neural network (LSTM) is likely to contain noisy information from many irrelevant sentences. This study proposes a sentence attention mechanism combining the sentence attention weights from a convolutional neural tensor network (CNTN) and the topic profile by selectively focusing on significant sentences for attention-based dialog state tracking. This study collected 260 interview dialogs consisting of 3,016 dialog turns for performance evaluation. A five-fold cross validation scheme was employed and the results show that the proposed method outperformed the semantic slot-based baseline method.

原文English
主出版物標題2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings
發行者Institute of Electrical and Electronics Engineers Inc.
頁面6144-6148
頁數5
2018-April
ISBN(列印)9781538646588
DOIs
出版狀態Published - 2018 九月 10
事件2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Calgary, Canada
持續時間: 2018 四月 152018 四月 20

Other

Other2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018
國家Canada
城市Calgary
期間18-04-1518-04-20

指紋

Semantics
Tensors
Neural networks
Long short-term memory

All Science Journal Classification (ASJC) codes

  • Software
  • Signal Processing
  • Electrical and Electronic Engineering

引用此文

Su, M. H., Wu, C. H., Huang, K. Y., & Chen, C. K. (2018). Attention-Based Dialog State Tracking for Conversational Interview Coaching. 於 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings (卷 2018-April, 頁 6144-6148). [8461494] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICASSP.2018.8461494
Su, Ming Hsiang ; Wu, Chung Hsien ; Huang, Kun Yi ; Chen, Chu Kwang. / Attention-Based Dialog State Tracking for Conversational Interview Coaching. 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. 卷 2018-April Institute of Electrical and Electronics Engineers Inc., 2018. 頁 6144-6148
@inproceedings{5e827ffe011e46e3a0997685e583036e,
title = "Attention-Based Dialog State Tracking for Conversational Interview Coaching",
abstract = "This study proposes an approach to dialog state tracking (DST) in a conversational interview coaching system. For the interview coaching task, the semantic slots, used mostly in traditional dialog systems, are difficult to define manually. This study adopts the topic profile of the response from the interviewee as the dialog state representation. In addition, as the response generally consists of several sentences, the summary vector obtained from a long short-term memory neural network (LSTM) is likely to contain noisy information from many irrelevant sentences. This study proposes a sentence attention mechanism combining the sentence attention weights from a convolutional neural tensor network (CNTN) and the topic profile by selectively focusing on significant sentences for attention-based dialog state tracking. This study collected 260 interview dialogs consisting of 3,016 dialog turns for performance evaluation. A five-fold cross validation scheme was employed and the results show that the proposed method outperformed the semantic slot-based baseline method.",
author = "Su, {Ming Hsiang} and Wu, {Chung Hsien} and Huang, {Kun Yi} and Chen, {Chu Kwang}",
year = "2018",
month = "9",
day = "10",
doi = "10.1109/ICASSP.2018.8461494",
language = "English",
isbn = "9781538646588",
volume = "2018-April",
pages = "6144--6148",
booktitle = "2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

Su, MH, Wu, CH, Huang, KY & Chen, CK 2018, Attention-Based Dialog State Tracking for Conversational Interview Coaching. 於 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. 卷 2018-April, 8461494, Institute of Electrical and Electronics Engineers Inc., 頁 6144-6148, 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018, Calgary, Canada, 18-04-15. https://doi.org/10.1109/ICASSP.2018.8461494

Attention-Based Dialog State Tracking for Conversational Interview Coaching. / Su, Ming Hsiang; Wu, Chung Hsien; Huang, Kun Yi; Chen, Chu Kwang.

2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. 卷 2018-April Institute of Electrical and Electronics Engineers Inc., 2018. p. 6144-6148 8461494.

研究成果: Conference contribution

TY - GEN

T1 - Attention-Based Dialog State Tracking for Conversational Interview Coaching

AU - Su, Ming Hsiang

AU - Wu, Chung Hsien

AU - Huang, Kun Yi

AU - Chen, Chu Kwang

PY - 2018/9/10

Y1 - 2018/9/10

N2 - This study proposes an approach to dialog state tracking (DST) in a conversational interview coaching system. For the interview coaching task, the semantic slots, used mostly in traditional dialog systems, are difficult to define manually. This study adopts the topic profile of the response from the interviewee as the dialog state representation. In addition, as the response generally consists of several sentences, the summary vector obtained from a long short-term memory neural network (LSTM) is likely to contain noisy information from many irrelevant sentences. This study proposes a sentence attention mechanism combining the sentence attention weights from a convolutional neural tensor network (CNTN) and the topic profile by selectively focusing on significant sentences for attention-based dialog state tracking. This study collected 260 interview dialogs consisting of 3,016 dialog turns for performance evaluation. A five-fold cross validation scheme was employed and the results show that the proposed method outperformed the semantic slot-based baseline method.

AB - This study proposes an approach to dialog state tracking (DST) in a conversational interview coaching system. For the interview coaching task, the semantic slots, used mostly in traditional dialog systems, are difficult to define manually. This study adopts the topic profile of the response from the interviewee as the dialog state representation. In addition, as the response generally consists of several sentences, the summary vector obtained from a long short-term memory neural network (LSTM) is likely to contain noisy information from many irrelevant sentences. This study proposes a sentence attention mechanism combining the sentence attention weights from a convolutional neural tensor network (CNTN) and the topic profile by selectively focusing on significant sentences for attention-based dialog state tracking. This study collected 260 interview dialogs consisting of 3,016 dialog turns for performance evaluation. A five-fold cross validation scheme was employed and the results show that the proposed method outperformed the semantic slot-based baseline method.

UR - http://www.scopus.com/inward/record.url?scp=85054233036&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85054233036&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2018.8461494

DO - 10.1109/ICASSP.2018.8461494

M3 - Conference contribution

AN - SCOPUS:85054233036

SN - 9781538646588

VL - 2018-April

SP - 6144

EP - 6148

BT - 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Su MH, Wu CH, Huang KY, Chen CK. Attention-Based Dialog State Tracking for Conversational Interview Coaching. 於 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. 卷 2018-April. Institute of Electrical and Electronics Engineers Inc. 2018. p. 6144-6148. 8461494 https://doi.org/10.1109/ICASSP.2018.8461494