Audio vs. Text: Identify a Powerful Modality for Effective Hate Speech Detection
Keywords: Hate Speech, Audio Hate, Text Hate
TL;DR: Multimodal hate speech detection
Abstract: The boom in social media platforms has witnessed a significant jump in offensive, hate, and toxic languages. These toxic contents leave a significant effect on one's personality which can even lead to depression and can be in various forms including audio and text. \textbf{However, the primary concern is that the benchmark datasets such as Toxigen for hate speech detection, are only text-based}. \textit{Therefore, in this research, for the first time, we have collected the audio-based hate speech dataset for unified security}. Utilizing both these modalities (text and audio), we have performed the benchmark study for audio and text hate detection and proposed a multimodal hate detection algorithm.
Submission Number: 31
Loading