This academic article discusses the potential of GPT-4V (multi-modal large language models) in marine analysis. Despite the success of GPT-4V in various fields, its use in domain-specific analysis like marine analysis has not been extensively studied. The researchers conducted a systematic evaluation of GPT-4V’s performance in marine research, setting a new benchmark for future developments. However, the experimental results showed that GPT-4V’s responses are still not meeting the specific requirements of marine professionals.
Publication date: 4 Jan 2024
Project Page: https://github.com/hkust-vgd/Marine GPT-4V Eval
Paper: https://arxiv.org/pdf/2401.02147
