Deep learning for hate speech detection: a comparative study (Aithor)

A thorough empirical comparison of deep and shallow hate-speech detection techniques using three widely used datasets is presented in the study. The research intends to highlight advancements in the field and pinpoint the advantages and disadvantages of the state-of-the-art at the moment. Measures of practical performance, such as domain generalization, computational efficiency, detection efficacy, and the capacity to employ pre-trained models, are the major emphasis of the examination. The objective is to quantify the state-of-the-art, suggest future research areas, and offer advise on the practical use of hate speech identification. Using three sizable and openly accessible hate speech detection benchmarks, the paper assesses 14 shallow/deep classification-based hate speech detectors that are enabled by various word representation techniques.

https://aithor.com/paper-summary/deep-learning-for-hate-speech-detection-a-comparative-study