BATON: Aligning Text-to-Audio Model with Human Preference Feedback
This addresses the challenge of aligning AI-generated audio with human preferences for users of text-to-audio models, representing an incremental improvement through fine-tuning with feedback.
The paper tackles the problem of text-to-audio models generating audio misaligned with human preferences due to language complexity and model limitations, and introduces BATON, a framework using human feedback to improve alignment, resulting in significant quality improvements in audio integrity, temporal relationships, and preference alignment.
With the development of AI-Generated Content (AIGC), text-to-audio models are gaining widespread attention. However, it is challenging for these models to generate audio aligned with human preference due to the inherent information density of natural language and limited model understanding ability. To alleviate this issue, we formulate the BATON, a framework designed to enhance the alignment between generated audio and text prompt using human preference feedback. Our BATON comprises three key stages: Firstly, we curated a dataset containing both prompts and the corresponding generated audio, which was then annotated based on human feedback. Secondly, we introduced a reward model using the constructed dataset, which can mimic human preference by assigning rewards to input text-audio pairs. Finally, we employed the reward model to fine-tune an off-the-shelf text-to-audio model. The experiment results demonstrate that our BATON can significantly improve the generation quality of the original text-to-audio models, concerning audio integrity, temporal relationship, and alignment with human preference.