NEU at WNUT-2020 Task 2: Data Augmentation To Tell BERT That Death Is Not Necessarily Informative
This work addresses the challenge of filtering useful information from social media during the COVID-19 pandemic, but it is incremental as it builds on existing BERT methods with a specific data augmentation tweak.
The authors tackled the problem of identifying informative COVID-19 tweets using a BERT classifier, showing that BERT exploits easy signals like mentions of deaths, with performance dropping from 92.63 to 7.28 F1-score when such patterns are added, and they proposed a data augmentation technique to improve robustness.
Millions of people around the world are sharing COVID-19 related information on social media platforms. Since not all the information shared on the social media is useful, a machine learning system to identify informative posts can help users in finding relevant information. In this paper, we present a BERT classifier system for W-NUT2020 Shared Task 2: Identification of Informative COVID-19 English Tweets. Further, we show that BERT exploits some easy signals to identify informative tweets, and adding simple patterns to uninformative tweets drastically degrades BERT performance. In particular, simply adding 10 deaths to tweets in dev set, reduces BERT F1- score from 92.63 to 7.28. We also propose a simple data augmentation technique that helps in improving the robustness and generalization ability of the BERT classifier.