|
The inherent differentiable nature of optics provides opportunities to harness the powerful tools of machine learning. Automatic differentiation underpins the rise of machine learning in the modern world, enabling the optimisation of million and billion parameter models, shedding the curse of dimensionality. By constructing optical models within automatic differentiation frameworks, simple approaches to challenging problems arise, and access to even more powerful optimisation and inference algorithms is granted. dLux is a recently developed fully differentiable open-source optical modelling framework that enables users to take full advantage of these benefits. |
|