The first open-source large model in the marine field, OceanGPT (Cangyuan), was officially launched in Hangzhou, Zhejiang. This innovative achievement was developed by the National Key Laboratory of Marine Precision Sensing Technology at Zhejiang University, marking an important step forward for China in the field of marine technology.
OceanGPT has the ability to answer basic questions about marine knowledge and can interpret multi-modal data such as sonar images and marine observation maps through natural language. This capability enables OceanGPT to perform well when processing complex marine data. The model also employs a "slow thinking" reasoning mechanism enhanced with domain-specific knowledge, which means it can effectively reduce the risk of generating incorrect information and solve the common "hallucination errors" in existing general large models.
Image source note: The image is AI-generated, and the image licensing service provider is Midjourney.
In practical applications, OceanGPT is not only capable of answering marine-related questions but also suitable for key scenarios such as controlling marine robots and underwater embodied intelligence. Professor Chen Huajun from the School of Computer Science at Zhejiang University stated that with OceanGPT's code generation capability, users only need to input a natural language instruction, and the model can generate corresponding robot control code, enabling rapid deployment and task execution. This technology aims to allow even non-professionals to drive underwater robots to complete complex tasks through simple voice commands.
Additionally, OceanGPT integrates the MCP large model protocol, aiming to promote the realization of multi-robot collaboration. In the future, this model will be directly deployed on marine robots, leveraging its reasoning capabilities to further enhance the autonomous operation capabilities and operational efficiency of marine equipment.
Currently, OceanGPT has completed technical verification on the Haiying series of underwater robot platforms at Zhejiang University. Test results show that after using OceanGPT, the efficiency of writing robot code has improved from "hour-level" to "second-level," greatly improving work efficiency.
Through the release of OceanGPT, China's research and development capabilities in the field of marine technology have been further demonstrated, and it is expected to play an important role in more extensive application scenarios in the future.