Modeling dynamic pairwise attention for crime classification over legal articles

Pengfei Wang, Ze Yang, Shuzi Niu, Yongfeng Zhang, Lei Zhang, Shaozhang Niu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

In juridical field, judges usually need to consult several relevant cases to determine the specific articles that the evidence violated, which is a task that is time consuming and needs extensive professional knowledge. In this paper, we focus on how to save the manual efforts and make the conviction process more efficient. Specifically, we treat the evidences as documents, and articles as labels, thus the conviction process can be cast as a multi-label classification problem. However, the challenge in this specific scenario lies in two aspects. One is that the number of articles that evidences violated is dynamic, which we denote as the label dynamic problem. The other is that most articles are violated by only a few of the evidences, which we denote as the label imbalance problem. Previous methods usually learn the multi-label classification model and the label thresholds independently, and may ignore the label imbalance problem. To tackle with both challenges, we propose a unified D ynamic P airwise A ttention M odel (DPAM for short) in this paper. Specifically, DPAM adopts the multi-task learning paradigm to learn the multi-label classifier and the threshold predictor jointly, and thus DPAM can improve the generalization performance by leveraging the information learned in both of the two tasks. In addition, a pairwise attention model based on article definitions is incorporated into the classification model to help alleviate the label imbalance problem. Experimental results on two real-world datasets show that our proposed approach significantly outperforms state-of-the-art multi-label classification methods.

Original languageEnglish (US)
Title of host publication41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018
PublisherAssociation for Computing Machinery, Inc
Pages485-494
Number of pages10
ISBN (Electronic)9781450356572
DOIs
StatePublished - Jun 27 2018
Event41st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018 - Ann Arbor, United States
Duration: Jul 8 2018Jul 12 2018

Publication series

Name41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018

Other

Other41st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018
CountryUnited States
CityAnn Arbor
Period7/8/187/12/18

Fingerprint

Crime
Labels
Classifiers

All Science Journal Classification (ASJC) codes

  • Software
  • Computer Graphics and Computer-Aided Design
  • Information Systems

Keywords

  • Dynamic threshold predictor
  • Multi-label classification
  • Pairwise attention model

Cite this

Wang, P., Yang, Z., Niu, S., Zhang, Y., Zhang, L., & Niu, S. (2018). Modeling dynamic pairwise attention for crime classification over legal articles. In 41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018 (pp. 485-494). (41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018). Association for Computing Machinery, Inc. https://doi.org/10.1145/3209978.3210057
Wang, Pengfei ; Yang, Ze ; Niu, Shuzi ; Zhang, Yongfeng ; Zhang, Lei ; Niu, Shaozhang. / Modeling dynamic pairwise attention for crime classification over legal articles. 41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018. Association for Computing Machinery, Inc, 2018. pp. 485-494 (41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018).
@inproceedings{67314468dc7640b081e88bbb88b02e0a,
title = "Modeling dynamic pairwise attention for crime classification over legal articles",
abstract = "In juridical field, judges usually need to consult several relevant cases to determine the specific articles that the evidence violated, which is a task that is time consuming and needs extensive professional knowledge. In this paper, we focus on how to save the manual efforts and make the conviction process more efficient. Specifically, we treat the evidences as documents, and articles as labels, thus the conviction process can be cast as a multi-label classification problem. However, the challenge in this specific scenario lies in two aspects. One is that the number of articles that evidences violated is dynamic, which we denote as the label dynamic problem. The other is that most articles are violated by only a few of the evidences, which we denote as the label imbalance problem. Previous methods usually learn the multi-label classification model and the label thresholds independently, and may ignore the label imbalance problem. To tackle with both challenges, we propose a unified D ynamic P airwise A ttention M odel (DPAM for short) in this paper. Specifically, DPAM adopts the multi-task learning paradigm to learn the multi-label classifier and the threshold predictor jointly, and thus DPAM can improve the generalization performance by leveraging the information learned in both of the two tasks. In addition, a pairwise attention model based on article definitions is incorporated into the classification model to help alleviate the label imbalance problem. Experimental results on two real-world datasets show that our proposed approach significantly outperforms state-of-the-art multi-label classification methods.",
keywords = "Dynamic threshold predictor, Multi-label classification, Pairwise attention model",
author = "Pengfei Wang and Ze Yang and Shuzi Niu and Yongfeng Zhang and Lei Zhang and Shaozhang Niu",
year = "2018",
month = "6",
day = "27",
doi = "10.1145/3209978.3210057",
language = "English (US)",
series = "41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018",
publisher = "Association for Computing Machinery, Inc",
pages = "485--494",
booktitle = "41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018",

}

Wang, P, Yang, Z, Niu, S, Zhang, Y, Zhang, L & Niu, S 2018, Modeling dynamic pairwise attention for crime classification over legal articles. in 41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018. 41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018, Association for Computing Machinery, Inc, pp. 485-494, 41st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018, Ann Arbor, United States, 7/8/18. https://doi.org/10.1145/3209978.3210057

Modeling dynamic pairwise attention for crime classification over legal articles. / Wang, Pengfei; Yang, Ze; Niu, Shuzi; Zhang, Yongfeng; Zhang, Lei; Niu, Shaozhang.

41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018. Association for Computing Machinery, Inc, 2018. p. 485-494 (41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Modeling dynamic pairwise attention for crime classification over legal articles

AU - Wang, Pengfei

AU - Yang, Ze

AU - Niu, Shuzi

AU - Zhang, Yongfeng

AU - Zhang, Lei

AU - Niu, Shaozhang

PY - 2018/6/27

Y1 - 2018/6/27

N2 - In juridical field, judges usually need to consult several relevant cases to determine the specific articles that the evidence violated, which is a task that is time consuming and needs extensive professional knowledge. In this paper, we focus on how to save the manual efforts and make the conviction process more efficient. Specifically, we treat the evidences as documents, and articles as labels, thus the conviction process can be cast as a multi-label classification problem. However, the challenge in this specific scenario lies in two aspects. One is that the number of articles that evidences violated is dynamic, which we denote as the label dynamic problem. The other is that most articles are violated by only a few of the evidences, which we denote as the label imbalance problem. Previous methods usually learn the multi-label classification model and the label thresholds independently, and may ignore the label imbalance problem. To tackle with both challenges, we propose a unified D ynamic P airwise A ttention M odel (DPAM for short) in this paper. Specifically, DPAM adopts the multi-task learning paradigm to learn the multi-label classifier and the threshold predictor jointly, and thus DPAM can improve the generalization performance by leveraging the information learned in both of the two tasks. In addition, a pairwise attention model based on article definitions is incorporated into the classification model to help alleviate the label imbalance problem. Experimental results on two real-world datasets show that our proposed approach significantly outperforms state-of-the-art multi-label classification methods.

AB - In juridical field, judges usually need to consult several relevant cases to determine the specific articles that the evidence violated, which is a task that is time consuming and needs extensive professional knowledge. In this paper, we focus on how to save the manual efforts and make the conviction process more efficient. Specifically, we treat the evidences as documents, and articles as labels, thus the conviction process can be cast as a multi-label classification problem. However, the challenge in this specific scenario lies in two aspects. One is that the number of articles that evidences violated is dynamic, which we denote as the label dynamic problem. The other is that most articles are violated by only a few of the evidences, which we denote as the label imbalance problem. Previous methods usually learn the multi-label classification model and the label thresholds independently, and may ignore the label imbalance problem. To tackle with both challenges, we propose a unified D ynamic P airwise A ttention M odel (DPAM for short) in this paper. Specifically, DPAM adopts the multi-task learning paradigm to learn the multi-label classifier and the threshold predictor jointly, and thus DPAM can improve the generalization performance by leveraging the information learned in both of the two tasks. In addition, a pairwise attention model based on article definitions is incorporated into the classification model to help alleviate the label imbalance problem. Experimental results on two real-world datasets show that our proposed approach significantly outperforms state-of-the-art multi-label classification methods.

KW - Dynamic threshold predictor

KW - Multi-label classification

KW - Pairwise attention model

UR - http://www.scopus.com/inward/record.url?scp=85051497336&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85051497336&partnerID=8YFLogxK

U2 - 10.1145/3209978.3210057

DO - 10.1145/3209978.3210057

M3 - Conference contribution

AN - SCOPUS:85051497336

T3 - 41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018

SP - 485

EP - 494

BT - 41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018

PB - Association for Computing Machinery, Inc

ER -

Wang P, Yang Z, Niu S, Zhang Y, Zhang L, Niu S. Modeling dynamic pairwise attention for crime classification over legal articles. In 41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018. Association for Computing Machinery, Inc. 2018. p. 485-494. (41st International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2018). https://doi.org/10.1145/3209978.3210057