This research uses multimodal large language models (MLLMs) to enhance art appreciation education. The study introduces LLaVA-Docent, a model developed using MLLMs that aims to improve access and engagement in art education. The model was developed through extensive literature review and expert consultation, and trained using a virtual dialogue dataset on GPT-4. The effectiveness of LLaVA-Docent was assessed through quantitative and qualitative evaluations, revealing its strengths and weaknesses. The study concludes that LLaVA-Docent can significantly contribute to the field of art education.
Publication date: 12 Feb 2024
Project Page: N/A
Paper: https://arxiv.org/pdf/2402.06264
