FlorUniTo@TRAC-2: Retrofitting Word Embeddings on an Abusive Lexicon for Aggressive Language Detection

摘要

This paper describes our participation to the TRAC-2 Shared Tasks on Aggression Identification. Our team, FlorUniTo, investigated the applicability of using an abusive lexicon to enhance word embeddings towards improving detection of aggressive language. The embeddings used in our paper are word-aligned pre-trained vectors for English, Hindi, and Bengali, to reflect the languages represented in the shared task datasets. The embeddings are retrofitted to a multilingual abusive lexicon, HurtLex. We experimented with an LSTM model using the original as well as the transformed embeddings and different language and setting variations. Overall, our systems placed toward the middle of the official rankings based on weighted Fl score. Furthermore, the results on the development and test sets show promise for this novel avenue of research.