User Utterance Acquisition for Training Task-Oriented Bots: A Review of Challenges, Techniques and Opportunities

Abstract

Building conversational task-oriented bots requires large and diverse sets of annotated user utterances to learn mappings between natural language utterances and user intents. Given the complexity of human language as well as the recent advances on intent recognition (especially deep-learning-based approaches), bot developers now have faced a new challenge: efficiently and effectively collecting a large number of quality (e.g., diverse and unbiased) training samples. This article studies training user utterance acquisition along several important dimensions including cost and quality. We discuss state of the art techniques, identify open issues, and inform an outlook on future research directions.

Original language	English
Article number	9022888
Pages (from-to)	30-38
Number of pages	9
Journal	IEEE Internet Computing
Volume	24
Issue number	3
DOIs	https://doi.org/10.1109/MIC.2020.2978157
Publication status	Published - 1 May 2020
Externally published	Yes

Keywords

Task-Oriented Bots
Training Data

ASJC Scopus subject areas

Computer Networks and Communications

Access to Document

10.1109/MIC.2020.2978157

Fingerprint Dive into the research topics of 'User Utterance Acquisition for Training Task-Oriented Bots: A Review of Challenges, Techniques and Opportunities'. Together they form a unique fingerprint.

Cite this

User Utterance Acquisition for Training Task-Oriented Bots : A Review of Challenges, Techniques and Opportunities. / Yaghoub-Zadeh-Fard, Mohammad Ali; Benatallah, Boualem; Casati, Fabio; Barukh, Moshe Chai; Zamanirad, Shayan.

In: IEEE Internet Computing, Vol. 24, No. 3, 9022888, 01.05.2020, p. 30-38.

Research output: Contribution to journal › Review article › peer-review

@article{9812ec5af39d43fc80107b6edf7ae90c,

title = "User Utterance Acquisition for Training Task-Oriented Bots: A Review of Challenges, Techniques and Opportunities",

abstract = "Building conversational task-oriented bots requires large and diverse sets of annotated user utterances to learn mappings between natural language utterances and user intents. Given the complexity of human language as well as the recent advances on intent recognition (especially deep-learning-based approaches), bot developers now have faced a new challenge: efficiently and effectively collecting a large number of quality (e.g., diverse and unbiased) training samples. This article studies training user utterance acquisition along several important dimensions including cost and quality. We discuss state of the art techniques, identify open issues, and inform an outlook on future research directions.",

keywords = "Task-Oriented Bots, Training Data",

author = "Yaghoub-Zadeh-Fard, {Mohammad Ali} and Boualem Benatallah and Fabio Casati and Barukh, {Moshe Chai} and Shayan Zamanirad",

year = "2020",

month = may,

day = "1",

doi = "10.1109/MIC.2020.2978157",

language = "English",

volume = "24",

pages = "30--38",

journal = "IEEE Internet Computing",

issn = "1089-7801",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "3",

}

TY - JOUR

T1 - User Utterance Acquisition for Training Task-Oriented Bots

T2 - A Review of Challenges, Techniques and Opportunities

AU - Yaghoub-Zadeh-Fard, Mohammad Ali

AU - Benatallah, Boualem

AU - Casati, Fabio

AU - Barukh, Moshe Chai

AU - Zamanirad, Shayan

PY - 2020/5/1

Y1 - 2020/5/1

N2 - Building conversational task-oriented bots requires large and diverse sets of annotated user utterances to learn mappings between natural language utterances and user intents. Given the complexity of human language as well as the recent advances on intent recognition (especially deep-learning-based approaches), bot developers now have faced a new challenge: efficiently and effectively collecting a large number of quality (e.g., diverse and unbiased) training samples. This article studies training user utterance acquisition along several important dimensions including cost and quality. We discuss state of the art techniques, identify open issues, and inform an outlook on future research directions.

AB - Building conversational task-oriented bots requires large and diverse sets of annotated user utterances to learn mappings between natural language utterances and user intents. Given the complexity of human language as well as the recent advances on intent recognition (especially deep-learning-based approaches), bot developers now have faced a new challenge: efficiently and effectively collecting a large number of quality (e.g., diverse and unbiased) training samples. This article studies training user utterance acquisition along several important dimensions including cost and quality. We discuss state of the art techniques, identify open issues, and inform an outlook on future research directions.

KW - Task-Oriented Bots

KW - Training Data

UR - http://www.scopus.com/inward/record.url?scp=85081398235&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85081398235&partnerID=8YFLogxK

U2 - 10.1109/MIC.2020.2978157

DO - 10.1109/MIC.2020.2978157

M3 - Review article

AN - SCOPUS:85081398235

VL - 24

SP - 30

EP - 38

JO - IEEE Internet Computing

JF - IEEE Internet Computing

SN - 1089-7801

IS - 3

M1 - 9022888