Using Contextual Representations for Suicide Risk Assessment from Internet Forums

Ashwin Karthik Ambalavanan, Pranjali Dileep Jagtap, Soumya Adhya, Murthy Devarakonda


Abstract
Social media posts may yield clues to the subject’s (usually, the writer’s) suicide risk and intent, which can be used for timely intervention. This research, motivated by the CLPsych 2019 shared task, developed neural network-based methods for analyzing posts in one or more Reddit forums to assess the subject’s suicide risk. One of the technical challenges this task poses is the large amount of text from multiple posts of a single user. Our neural network models use the advanced multi-headed Attention-based autoencoder architecture, called Bidirectional Encoder Representations from Transformers (BERT). Our system achieved the 2nd best performance of 0.477 macro averaged F measure on Task A of the challenge. Among the three different alternatives we developed for the challenge, the single BERT model that processed all of a user’s posts performed the best on all three Tasks.
Anthology ID:
W19-3022
Volume:
Proceedings of the Sixth Workshop on Computational Linguistics and Clinical Psychology
Month:
June
Year:
2019
Address:
Minneapolis, Minnesota
Editors:
Kate Niederhoffer, Kristy Hollingshead, Philip Resnik, Rebecca Resnik, Kate Loveys
Venue:
CLPsych
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
172–176
Language:
URL:
https://aclanthology.org/W19-3022
DOI:
10.18653/v1/W19-3022
Bibkey:
Cite (ACL):
Ashwin Karthik Ambalavanan, Pranjali Dileep Jagtap, Soumya Adhya, and Murthy Devarakonda. 2019. Using Contextual Representations for Suicide Risk Assessment from Internet Forums. In Proceedings of the Sixth Workshop on Computational Linguistics and Clinical Psychology, pages 172–176, Minneapolis, Minnesota. Association for Computational Linguistics.
Cite (Informal):
Using Contextual Representations for Suicide Risk Assessment from Internet Forums (Ambalavanan et al., CLPsych 2019)
Copy Citation:
PDF:
https://preview.aclanthology.org/naacl24-info/W19-3022.pdf