Commonsense causal reasoning between short texts

Zhiyi Luo, Yuchen Sha, Kenny Q. Zhu, Seungwon Hwang, Zhongyuan Wang

Research output: Contribution to conferencePaper

3 Citations (Scopus)

Abstract

Commonsense causal reasoning is the process of capturing and understanding the causal dependencies amongst events and actions. Such events and actions can be expressed in terms, phrases or sentences in natural language text. Therefore, one possible way of obtaining causal knowledge is by extracting causal relations between terms or phrases from a large text corpus. However, causal relations in text are sparse, ambiguous, and sometimes implicit, and thus difficult to obtain. This paper attacks the problem of commonsense causality reasoning between short texts (phrases and sentences) using a data driven approach. We propose a framework that automatically harvests a network of causal-effect terms from a large web corpus. Backed by this network, we propose a novel and effective metric to properly model the causality strength between terms. We show these signals can be aggregated for causality reasonings between short texts, including sentences and phrases. In particular, our approach outperforms all previously reported results in the standard SEMEVAL COPA task by substantial margins.

Original languageEnglish
Pages421-430
Number of pages10
Publication statusPublished - 2016 Jan 1
Event15th International Conference on Principles of Knowledge Representation and Reasoning, KR 2016 - Cape Town, South Africa
Duration: 2016 Apr 252016 Apr 29

Other

Other15th International Conference on Principles of Knowledge Representation and Reasoning, KR 2016
CountrySouth Africa
CityCape Town
Period16/4/2516/4/29

Fingerprint

Reasoning
Causality
Term
Causal Effect
Ambiguous
Data-driven
Margin
Natural Language
Attack
Text
Metric
Corpus
Model

All Science Journal Classification (ASJC) codes

  • Logic
  • Software

Cite this

Luo, Z., Sha, Y., Zhu, K. Q., Hwang, S., & Wang, Z. (2016). Commonsense causal reasoning between short texts. 421-430. Paper presented at 15th International Conference on Principles of Knowledge Representation and Reasoning, KR 2016, Cape Town, South Africa.
Luo, Zhiyi ; Sha, Yuchen ; Zhu, Kenny Q. ; Hwang, Seungwon ; Wang, Zhongyuan. / Commonsense causal reasoning between short texts. Paper presented at 15th International Conference on Principles of Knowledge Representation and Reasoning, KR 2016, Cape Town, South Africa.10 p.
@conference{628f6b65b8f2444ebf2fead36eeda42b,
title = "Commonsense causal reasoning between short texts",
abstract = "Commonsense causal reasoning is the process of capturing and understanding the causal dependencies amongst events and actions. Such events and actions can be expressed in terms, phrases or sentences in natural language text. Therefore, one possible way of obtaining causal knowledge is by extracting causal relations between terms or phrases from a large text corpus. However, causal relations in text are sparse, ambiguous, and sometimes implicit, and thus difficult to obtain. This paper attacks the problem of commonsense causality reasoning between short texts (phrases and sentences) using a data driven approach. We propose a framework that automatically harvests a network of causal-effect terms from a large web corpus. Backed by this network, we propose a novel and effective metric to properly model the causality strength between terms. We show these signals can be aggregated for causality reasonings between short texts, including sentences and phrases. In particular, our approach outperforms all previously reported results in the standard SEMEVAL COPA task by substantial margins.",
author = "Zhiyi Luo and Yuchen Sha and Zhu, {Kenny Q.} and Seungwon Hwang and Zhongyuan Wang",
year = "2016",
month = "1",
day = "1",
language = "English",
pages = "421--430",
note = "15th International Conference on Principles of Knowledge Representation and Reasoning, KR 2016 ; Conference date: 25-04-2016 Through 29-04-2016",

}

Luo, Z, Sha, Y, Zhu, KQ, Hwang, S & Wang, Z 2016, 'Commonsense causal reasoning between short texts', Paper presented at 15th International Conference on Principles of Knowledge Representation and Reasoning, KR 2016, Cape Town, South Africa, 16/4/25 - 16/4/29 pp. 421-430.

Commonsense causal reasoning between short texts. / Luo, Zhiyi; Sha, Yuchen; Zhu, Kenny Q.; Hwang, Seungwon; Wang, Zhongyuan.

2016. 421-430 Paper presented at 15th International Conference on Principles of Knowledge Representation and Reasoning, KR 2016, Cape Town, South Africa.

Research output: Contribution to conferencePaper

TY - CONF

T1 - Commonsense causal reasoning between short texts

AU - Luo, Zhiyi

AU - Sha, Yuchen

AU - Zhu, Kenny Q.

AU - Hwang, Seungwon

AU - Wang, Zhongyuan

PY - 2016/1/1

Y1 - 2016/1/1

N2 - Commonsense causal reasoning is the process of capturing and understanding the causal dependencies amongst events and actions. Such events and actions can be expressed in terms, phrases or sentences in natural language text. Therefore, one possible way of obtaining causal knowledge is by extracting causal relations between terms or phrases from a large text corpus. However, causal relations in text are sparse, ambiguous, and sometimes implicit, and thus difficult to obtain. This paper attacks the problem of commonsense causality reasoning between short texts (phrases and sentences) using a data driven approach. We propose a framework that automatically harvests a network of causal-effect terms from a large web corpus. Backed by this network, we propose a novel and effective metric to properly model the causality strength between terms. We show these signals can be aggregated for causality reasonings between short texts, including sentences and phrases. In particular, our approach outperforms all previously reported results in the standard SEMEVAL COPA task by substantial margins.

AB - Commonsense causal reasoning is the process of capturing and understanding the causal dependencies amongst events and actions. Such events and actions can be expressed in terms, phrases or sentences in natural language text. Therefore, one possible way of obtaining causal knowledge is by extracting causal relations between terms or phrases from a large text corpus. However, causal relations in text are sparse, ambiguous, and sometimes implicit, and thus difficult to obtain. This paper attacks the problem of commonsense causality reasoning between short texts (phrases and sentences) using a data driven approach. We propose a framework that automatically harvests a network of causal-effect terms from a large web corpus. Backed by this network, we propose a novel and effective metric to properly model the causality strength between terms. We show these signals can be aggregated for causality reasonings between short texts, including sentences and phrases. In particular, our approach outperforms all previously reported results in the standard SEMEVAL COPA task by substantial margins.

UR - http://www.scopus.com/inward/record.url?scp=85011110659&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85011110659&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:85011110659

SP - 421

EP - 430

ER -

Luo Z, Sha Y, Zhu KQ, Hwang S, Wang Z. Commonsense causal reasoning between short texts. 2016. Paper presented at 15th International Conference on Principles of Knowledge Representation and Reasoning, KR 2016, Cape Town, South Africa.