In this paper, we propose a novel deep-learning architecture for text classification, named cross segment-and-concatenate multi-task learning (CSC-MTL). We use CSC-MTL to improve the performance of cyber-aggression detection from text. Our approach provides a robust shared feature representation for multi-task learning by detecting contrasts and similarities among polarity and neutral classes. We participated in the cyber-aggression shared task under the team name uOttawa. We report 59.74% F1 performance for the Facebook test set and 56.9% for the Twitter test set, for detecting aggression from text.
Ahmed Husseini Orabi, Mahmoud Husseini Orabi, Qianjia Huang, Diana Inkpen, and David Van Bruwaene. 2018. Cyber-aggression Detection using Cross Segment-and-Concatenate Multi-Task Learning from Text. In Proceedings of the First Workshop on Trolling, Aggression and Cyberbullying (TRAC-2018), pages 159–165, Santa Fe, New Mexico, USA. Association for Computational Linguistics.
@inproceedings{husseini-orabi-etal-2018-cyber,
title = "Cyber-aggression Detection using Cross Segment-and-Concatenate Multi-Task Learning from Text",
author = "Husseini Orabi, Ahmed and
Husseini Orabi, Mahmoud and
Huang, Qianjia and
Inkpen, Diana and
Van Bruwaene, David",
editor = "Kumar, Ritesh and
Ojha, Atul Kr. and
Zampieri, Marcos and
Malmasi, Shervin",
booktitle = "Proceedings of the First Workshop on Trolling, Aggression and Cyberbullying ({TRAC}-2018)",
month = aug,
year = "2018",
address = "Santa Fe, New Mexico, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/W18-4419/",
pages = "159--165",
abstract = "In this paper, we propose a novel deep-learning architecture for text classification, named cross segment-and-concatenate multi-task learning (CSC-MTL). We use CSC-MTL to improve the performance of cyber-aggression detection from text. Our approach provides a robust shared feature representation for multi-task learning by detecting contrasts and similarities among polarity and neutral classes. We participated in the cyber-aggression shared task under the team name uOttawa. We report 59.74{\%} F1 performance for the Facebook test set and 56.9{\%} for the Twitter test set, for detecting aggression from text."
}
%0 Conference Proceedings
%T Cyber-aggression Detection using Cross Segment-and-Concatenate Multi-Task Learning from Text
%A Husseini Orabi, Ahmed
%A Husseini Orabi, Mahmoud
%A Huang, Qianjia
%A Inkpen, Diana
%A Van Bruwaene, David
%Y Kumar, Ritesh
%Y Ojha, Atul Kr.
%Y Zampieri, Marcos
%Y Malmasi, Shervin
%S Proceedings of the First Workshop on Trolling, Aggression and Cyberbullying (TRAC-2018)
%D 2018
%8 August
%I Association for Computational Linguistics
%C Santa Fe, New Mexico, USA
%F husseini-orabi-etal-2018-cyber
%X In this paper, we propose a novel deep-learning architecture for text classification, named cross segment-and-concatenate multi-task learning (CSC-MTL). We use CSC-MTL to improve the performance of cyber-aggression detection from text. Our approach provides a robust shared feature representation for multi-task learning by detecting contrasts and similarities among polarity and neutral classes. We participated in the cyber-aggression shared task under the team name uOttawa. We report 59.74% F1 performance for the Facebook test set and 56.9% for the Twitter test set, for detecting aggression from text.
%U https://aclanthology.org/W18-4419/
%P 159-165
[Cyber-aggression Detection using Cross Segment-and-Concatenate Multi-Task Learning from Text](https://aclanthology.org/W18-4419/) (Husseini Orabi et al., TRAC 2018)
Ahmed Husseini Orabi, Mahmoud Husseini Orabi, Qianjia Huang, Diana Inkpen, and David Van Bruwaene. 2018. Cyber-aggression Detection using Cross Segment-and-Concatenate Multi-Task Learning from Text. In Proceedings of the First Workshop on Trolling, Aggression and Cyberbullying (TRAC-2018), pages 159–165, Santa Fe, New Mexico, USA. Association for Computational Linguistics.