2020/09/06 by Nickil Maveli, Maveli, Nickil
Computer Science · Social Sciences · #Computation and Language (cs.CL) #FOS: Computer and information sciences #Hate Speech and Cyberbullying Detection #Information Retrieval (cs.IR) #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Misinformation and Its Impacts #Social and Information Networks (cs.SI) #Topic Modeling
paper · pdf · doi:10.48550/arxiv.2009.06375
openalex publication_date 2020/09/06 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Twitter and, in general, social media has become an indispensable\ncommunication channel in times of emergency. The ubiquitousness of smartphone\ngadgets enables people to declare an emergency observed in real-time. As a\nresult, more agencies are interested in programmatically monitoring Twitter\n(disaster relief organizations and news agencies). Therefore, recognizing the\ninformativeness of a Tweet can help filter noise from the large volumes of\nTweets. In this paper, we present our submission for WNUT-2020 Task 2:\nIdentification of informative COVID-19 English Tweets. Our most successful\nmodel is an ensemble of transformers, including RoBERTa, XLNet, and BERTweet\ntrained in a Semi-Supervised Learning (SSL) setting. The proposed system\nachieves an F1 score of 0.9011 on the test set (ranking 7th on the leaderboard)\nand shows significant gains in performance compared to a baseline system using\nFastText embeddings.\n