A little goes a long way: Improving toxic language classification despite data scarcity

Mika Juuti, Tommi Gröndahl, Adrian Flanagan, N. Asokan


Abstract
Detection of some types of toxic language is hampered by extreme scarcity of labeled training data. Data augmentation – generating new synthetic data from a labeled seed dataset – can help. The efficacy of data augmentation on toxic language classification has not been fully explored. We present the first systematic study on how data augmentation techniques impact performance across toxic language classifiers, ranging from shallow logistic regression architectures to BERT – a state-of-the-art pretrained Transformer network. We compare the performance of eight techniques on very scarce seed datasets. We show that while BERT performed the best, shallow classifiers performed comparably when trained on data augmented with a combination of three techniques, including GPT-2-generated sentences. We discuss the interplay of performance and computational overhead, which can inform the choice of techniques under different constraints.
Anthology ID:
2020.findings-emnlp.269
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2020
Month:
November
Year:
2020
Address:
Online
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
2991–3009
Language:
URL:
https://aclanthology.org/2020.findings-emnlp.269
DOI:
10.18653/v1/2020.findings-emnlp.269
Bibkey:
Cite (ACL):
Mika Juuti, Tommi Gröndahl, Adrian Flanagan, and N. Asokan. 2020. A little goes a long way: Improving toxic language classification despite data scarcity. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 2991–3009, Online. Association for Computational Linguistics.
Cite (Informal):
A little goes a long way: Improving toxic language classification despite data scarcity (Juuti et al., Findings 2020)
Copy Citation:
PDF:
https://preview.aclanthology.org/ingestion-script-update/2020.findings-emnlp.269.pdf
Video:
 https://slideslive.com/38940137
Code
 ssg-research/language-data-augmentation