HeadlineCause: A Dataset of News Headlines for Detecting Causalities

Ilya Gusev, Alexey Tikhonov


Abstract
Detecting implicit causal relations in texts is a task that requires both common sense and world knowledge. Existing datasets are focused either on commonsense causal reasoning or explicit causal relations. In this work, we present HeadlineCause, a dataset for detecting implicit causal relations between pairs of news headlines. The dataset includes over 5000 headline pairs from English news and over 9000 headline pairs from Russian news labeled through crowdsourcing. The pairs vary from totally unrelated or belonging to the same general topic to the ones including causation and refutation relations. We also present a set of models and experiments that demonstrates the dataset validity, including a multilingual XLM-RoBERTa based model for causality detection and a GPT-2 based model for possible effects prediction.
Anthology ID:
2022.lrec-1.662
Volume:
Proceedings of the Thirteenth Language Resources and Evaluation Conference
Month:
June
Year:
2022
Address:
Marseille, France
Venue:
LREC
SIG:
Publisher:
European Language Resources Association
Note:
Pages:
6153–6161
Language:
URL:
https://aclanthology.org/2022.lrec-1.662
DOI:
Bibkey:
Cite (ACL):
Ilya Gusev and Alexey Tikhonov. 2022. HeadlineCause: A Dataset of News Headlines for Detecting Causalities. In Proceedings of the Thirteenth Language Resources and Evaluation Conference, pages 6153–6161, Marseille, France. European Language Resources Association.
Cite (Informal):
HeadlineCause: A Dataset of News Headlines for Detecting Causalities (Gusev & Tikhonov, LREC 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.lrec-1.662.pdf
Code
 ilyagusev/headlinecause
Data
HeadlineCauseCOPASuperGLUE