Incorporating Sub-Word Level Information in Language Invariant Neural Event Detection

Suhan Prabhu, Pranav Goel, Alok Debnath, Manish Shrivastava

Abstract
Detection of TimeML events in text have traditionally been done on corpora such as TimeBanks. However, deep learning methods have not been applied to these corpora, because these datasets seldom contain more than 10,000 event mentions. Traditional architectures revolve around highly feature engineered, language specific statistical models. In this paper, we present a Language Invariant Neural Event Detection (ALINED) architecture. ALINED uses an aggregation of both sub-word level features as well as lexical and structural information. This is achieved by combining convolution over character embeddings, with recurrent layers over contextual word embeddings. We find that our model extracts relevant features for event span identification without relying on language specific features. We compare the performance of our language invariant model to the current state-of-the-art in English, Spanish, Italian and French. We outperform the F1-score of the state of the art in English by 1.65 points. We achieve F1-scores of 84.96, 80.87 and 74.81 on Spanish, Italian and French respectively which is comparable to the current states of the art for these languages. We also introduce the automatic annotation of events in Hindi, a low resource language, with an F1-Score of 77.13.
Anthology ID:
2019.icon-1.5
Volume:
Proceedings of the 16th International Conference on Natural Language Processing
Month:
December
Year:
2019
Address:
International Institute of Information Technology, Hyderabad, India
Editors:
Dipti Misra Sharma, Pushpak Bhattacharya
Venue:
ICON
SIG:
Publisher:
NLP Association of India
Note:
Pages:
36–44
Language:
URL:
https://aclanthology.org/2019.icon-1.5
DOI:
Bibkey:
Cite (ACL):
Suhan Prabhu, Pranav Goel, Alok Debnath, and Manish Shrivastava. 2019. Incorporating Sub-Word Level Information in Language Invariant Neural Event Detection. In Proceedings of the 16th International Conference on Natural Language Processing, pages 36–44, International Institute of Information Technology, Hyderabad, India. NLP Association of India.
Cite (Informal):
Incorporating Sub-Word Level Information in Language Invariant Neural Event Detection (Prabhu et al., ICON 2019)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-4/2019.icon-1.5.pdf
PDF Search