Question-answering aspect classification with multi-attention representation

Hanqian Wu; Mumu Liu; Jingjing Wang; Jue Xie; Shoushan Li

doi:10.1007/978-3-030-01012-6_7

Question-answering aspect classification with multi-attention representation

Hanqian Wu
, Mumu Liu
, Jingjing Wang
, Jue Xie
, Shoushan Li

Department of Human Centred Computing

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research

3 Citations (Scopus)

Abstract

In e-commerce platforms, the question-answering style reviews are emerging, which usually contains much aspect-related information about products. In this paper, Question-answering (QA) aspect classification is a new task that aims to identify the aspect category of a given QA text pair. According to characteristics of QA-style reviews, we draw up annotation guidelines and build a high-consistency annotated corpus for QA aspect classification. Then, we propose a recurrent neural network based on multi-attention representation to tackle this new task. Specifically, we firstly segment the answer text into clauses, and then leverage the multi-attention representation layer to match the question text with clauses inside answer text and generate multiple attention representations of the question text, which extends feature information of the question text. The experimental results demonstrate that our method for QA aspect classification, which is based on multi-attention representation, can make the most of useful information in answer texts and perform better than some strong baselines in QA aspect classification.

Original language	English
Title of host publication	24th China Conference, CCIR 2018 Guilin, China, September 27–29, 2018 Proceedings
Editors	Xianxian Li, Chenliang Li, Tie-Yan Liu, Jiafeng Guo, Shichao Zhang
Place of Publication	Cham Switzerland
Publisher	Springer
Pages	78-89
Number of pages	12
ISBN (Electronic)	9783030010126
ISBN (Print)	9783030010119
DOIs	https://doi.org/10.1007/978-3-030-01012-6_7
Publication status	Published - 2018
Event	China Conference on Information Retrieval 2018 - Guilin, China Duration: 27 Sept 2018 → 29 Sept 2018 Conference number: 24th https://link.springer.com/book/10.1007/978-3-030-01012-6 (Proceedings)

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer
Volume	11168
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	China Conference on Information Retrieval 2018
Abbreviated title	CCIR 2018
Country/Territory	China
City	Guilin
Period	27/09/18 → 29/09/18
Internet address	https://link.springer.com/book/10.1007/978-3-030-01012-6 (Proceedings)

Keywords

Aspect classification
Attention mechanism
Question answering

Access to Document

10.1007/978-3-030-01012-6_7

Cite this

Wu, H., Liu, M., Wang, J., Xie, J., & Li, S. (2018). Question-answering aspect classification with multi-attention representation. In X. Li, C. Li, T.-Y. Liu, J. Guo, & S. Zhang (Eds.), 24th China Conference, CCIR 2018 Guilin, China, September 27–29, 2018 Proceedings (pp. 78-89). (Lecture Notes in Computer Science; Vol. 11168). Springer. https://doi.org/10.1007/978-3-030-01012-6_7

Wu, Hanqian ; Liu, Mumu ; Wang, Jingjing et al. / Question-answering aspect classification with multi-attention representation. 24th China Conference, CCIR 2018 Guilin, China, September 27–29, 2018 Proceedings. editor / Xianxian Li ; Chenliang Li ; Tie-Yan Liu ; Jiafeng Guo ; Shichao Zhang. Cham Switzerland : Springer, 2018. pp. 78-89 (Lecture Notes in Computer Science).

@inproceedings{82ce18ed81c64635b4122296c6a0885a,

title = "Question-answering aspect classification with multi-attention representation",

abstract = "In e-commerce platforms, the question-answering style reviews are emerging, which usually contains much aspect-related information about products. In this paper, Question-answering (QA) aspect classification is a new task that aims to identify the aspect category of a given QA text pair. According to characteristics of QA-style reviews, we draw up annotation guidelines and build a high-consistency annotated corpus for QA aspect classification. Then, we propose a recurrent neural network based on multi-attention representation to tackle this new task. Specifically, we firstly segment the answer text into clauses, and then leverage the multi-attention representation layer to match the question text with clauses inside answer text and generate multiple attention representations of the question text, which extends feature information of the question text. The experimental results demonstrate that our method for QA aspect classification, which is based on multi-attention representation, can make the most of useful information in answer texts and perform better than some strong baselines in QA aspect classification.",

keywords = "Aspect classification, Attention mechanism, Question answering",

author = "Hanqian Wu and Mumu Liu and Jingjing Wang and Jue Xie and Shoushan Li",

note = "Funding Information: Acknowledgements. This work is supported in part by Industrial Prospective Project of Jiangsu Technology Department under Grant No. BE2017081 and the National Natural Science Foundation of China under Grant No. 61572129. Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2018.; China Conference on Information Retrieval 2018, CCIR 2018 ; Conference date: 27-09-2018 Through 29-09-2018",

year = "2018",

doi = "10.1007/978-3-030-01012-6\_7",

language = "English",

isbn = "9783030010119",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

pages = "78--89",

editor = "Xianxian Li and Chenliang Li and Tie-Yan Liu and Jiafeng Guo and Shichao Zhang",

booktitle = "24th China Conference, CCIR 2018 Guilin, China, September 27–29, 2018 Proceedings",

address = "Switzerland",

url = "https://link.springer.com/book/10.1007/978-3-030-01012-6",

}

Wu, H, Liu, M, Wang, J, Xie, J & Li, S 2018, Question-answering aspect classification with multi-attention representation. in X Li, C Li, T-Y Liu, J Guo & S Zhang (eds), 24th China Conference, CCIR 2018 Guilin, China, September 27–29, 2018 Proceedings. Lecture Notes in Computer Science, vol. 11168, Springer, Cham Switzerland, pp. 78-89, China Conference on Information Retrieval 2018, Guilin, China, 27/09/18. https://doi.org/10.1007/978-3-030-01012-6_7

Question-answering aspect classification with multi-attention representation. / Wu, Hanqian; Liu, Mumu; Wang, Jingjing et al.
24th China Conference, CCIR 2018 Guilin, China, September 27–29, 2018 Proceedings. ed. / Xianxian Li; Chenliang Li; Tie-Yan Liu; Jiafeng Guo; Shichao Zhang. Cham Switzerland: Springer, 2018. p. 78-89 (Lecture Notes in Computer Science; Vol. 11168).

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research

TY - GEN

T1 - Question-answering aspect classification with multi-attention representation

AU - Wu, Hanqian

AU - Liu, Mumu

AU - Wang, Jingjing

AU - Xie, Jue

AU - Li, Shoushan

N1 - Conference code: 24th

PY - 2018

Y1 - 2018

N2 - In e-commerce platforms, the question-answering style reviews are emerging, which usually contains much aspect-related information about products. In this paper, Question-answering (QA) aspect classification is a new task that aims to identify the aspect category of a given QA text pair. According to characteristics of QA-style reviews, we draw up annotation guidelines and build a high-consistency annotated corpus for QA aspect classification. Then, we propose a recurrent neural network based on multi-attention representation to tackle this new task. Specifically, we firstly segment the answer text into clauses, and then leverage the multi-attention representation layer to match the question text with clauses inside answer text and generate multiple attention representations of the question text, which extends feature information of the question text. The experimental results demonstrate that our method for QA aspect classification, which is based on multi-attention representation, can make the most of useful information in answer texts and perform better than some strong baselines in QA aspect classification.

AB - In e-commerce platforms, the question-answering style reviews are emerging, which usually contains much aspect-related information about products. In this paper, Question-answering (QA) aspect classification is a new task that aims to identify the aspect category of a given QA text pair. According to characteristics of QA-style reviews, we draw up annotation guidelines and build a high-consistency annotated corpus for QA aspect classification. Then, we propose a recurrent neural network based on multi-attention representation to tackle this new task. Specifically, we firstly segment the answer text into clauses, and then leverage the multi-attention representation layer to match the question text with clauses inside answer text and generate multiple attention representations of the question text, which extends feature information of the question text. The experimental results demonstrate that our method for QA aspect classification, which is based on multi-attention representation, can make the most of useful information in answer texts and perform better than some strong baselines in QA aspect classification.

KW - Aspect classification

KW - Attention mechanism

KW - Question answering

UR - https://www.scopus.com/pages/publications/85057261329

U2 - 10.1007/978-3-030-01012-6_7

DO - 10.1007/978-3-030-01012-6_7

M3 - Conference Paper

AN - SCOPUS:85057261329

SN - 9783030010119

T3 - Lecture Notes in Computer Science

SP - 78

EP - 89

BT - 24th China Conference, CCIR 2018 Guilin, China, September 27–29, 2018 Proceedings

A2 - Li, Xianxian

A2 - Li, Chenliang

A2 - Liu, Tie-Yan

A2 - Guo, Jiafeng

A2 - Zhang, Shichao

PB - Springer

CY - Cham Switzerland

T2 - China Conference on Information Retrieval 2018

Y2 - 27 September 2018 through 29 September 2018

ER -

Question-answering aspect classification with multi-attention representation

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Cite this