Enhancing Image Caption Generation Using Reinforcement Learning with Human Feedback

arXiv:2403.06735v1 Announce Type: new
Abstract: Research on generative models to produce human-aligned / human-preferred outputs has seen significant recent contributions. Between text and image-generative models, we narrowed our focus to text-based generative models, particularly to produce captions for images that align with human preferences. In this research, we explored a potential method to amplify the performance of the Deep Neural Network Model to generate captions that are preferred by humans. This was achieved by integrating Supervised Learning and Reinforcement Learning …

abstract arxiv captions cs.ai cs.cv feedback focus generative generative models human human feedback image images reinforcement reinforcement learning research text type