PROJECT OYSTER
Our Technology
Project Oyster uses physics-informed neural networks, which combines real life physics constraints into neural networks. There is a breakthrough in studying the way in which particles are carried through substances that could be used for our project.
More InfoAccurate Results
Through modelling the whole ocean, maximum error bound is 2%
Cost Efficient
Training can occur with smaller datasets, and without extensive sensor infrastructure
Computationally Efficient
Current simulations require supercomputers, but our neural network can be runned on a household computer
