LLMs-Integrated Automatic Hate Speech Recognition Using Controllable Text Generation Models

TL;DR

This paper introduces a novel LLM-integrated ASR system that combines transcription with hate speech censorship, utilizing curriculum learning and text generation techniques to improve hate word masking accuracy.

Contribution

It presents a new method for hate speech recognition and censorship in speech using LLMs, with a curriculum learning approach to enhance performance.

Findings

Achieved 58.6% masking accuracy for hate words, surpassing baselines.

Demonstrated that curriculum training improves transcription and censorship efficiency.

Generated hate speech datasets with controlled hate content levels.

Abstract

This paper proposes an automatic speech recognition (ASR) model for hate speech using large language models (LLMs). The proposed method integrates the encoder of the ASR model with the decoder of the LLMs, enabling simultaneous transcription and censorship tasks to prevent the exposure of harmful content. Instruction tuning of the LLM to mask hate-related words with specific tokens requires an annotated hate speech dataset, which is limited. We generate text samples using an LLM with the Chain-of-Thought (CoT) prompting technique guided by cultural context and examples and then convert them into speech samples using a text-to-speech (TTS) system. However, some of them contain non-hate speech samples with hate-related words, which degrades the censorship performance. This paper filters the samples which text classification models correctly label as hate content. By adjusting the threshold for the number of correct answer models, we can control the level of hate in the generated dataset, allowing us to train the LLMs through curriculum learning in a gradual manner. Experimental results show that the proposed method achieves a masking accuracy of 58.6\% for hate-related words, surpassing previous baselines. We also confirm that the curriculum training contributes to the efficiency of both transcription and censorship tasks.