@inproceedings{toraman-etal-2022-large,
    title = "Large-Scale Hate Speech Detection with Cross-Domain Transfer",
    author = "Toraman, Cagri  and
      {\c{S}}ahinu{\c{c}}, Furkan  and
      Yilmaz, Eyup",
    editor = "Calzolari, Nicoletta  and
      B{\'e}chet, Fr{\'e}d{\'e}ric  and
      Blache, Philippe  and
      Choukri, Khalid  and
      Cieri, Christopher  and
      Declerck, Thierry  and
      Goggi, Sara  and
      Isahara, Hitoshi  and
      Maegaard, Bente  and
      Mariani, Joseph  and
      Mazo, H{\'e}l{\`e}ne  and
      Odijk, Jan  and
      Piperidis, Stelios",
    booktitle = "Proceedings of the Thirteenth Language Resources and Evaluation Conference",
    month = jun,
    year = "2022",
    address = "Marseille, France",
    publisher = "European Language Resources Association",
    url = "https://preview.aclanthology.org/add-emnlp-2024-awards/2022.lrec-1.238/",
    pages = "2215--2225",
    abstract = "The performance of hate speech detection models relies on the datasets on which the models are trained. Existing datasets are mostly prepared with a limited number of instances or hate domains that define hate topics. This hinders large-scale analysis and transfer learning with respect to hate domains. In this study, we construct large-scale tweet datasets for hate speech detection in English and a low-resource language, Turkish, consisting of human-labeled 100k tweets per each. Our datasets are designed to have equal number of tweets distributed over five domains. The experimental results supported by statistical tests show that Transformer-based language models outperform conventional bag-of-words and neural models by at least 5{\%} in English and 10{\%} in Turkish for large-scale hate speech detection. The performance is also scalable to different training sizes, such that 98{\%} of performance in English, and 97{\%} in Turkish, are recovered when 20{\%} of training instances are used. We further examine the generalization ability of cross-domain transfer among hate domains. We show that 96{\%} of the performance of a target domain in average is recovered by other domains for English, and 92{\%} for Turkish. Gender and religion are more successful to generalize to other domains, while sports fail most."
}