Differentiable programming tools and techniques for modelling with small data in Python

Poster

See poster

This poster will first present the concepts of differential programming and how it can be used for modelling when available data is limited, but a priori knowledge is avaiable and high explainability is required. It then review the state of the python ecosystem (with libraries lire JAX or taichi) and how it compares to Julia’s SciML ecosystem.

Speaker

Fabian Dubois

Tokyo-based, doing consulting, research and development for sustainable projects on Earth and in Space. Current projects include robotic control for lunar surface exploration and simulation for water treatment plants. I have been working with python and data for 10 years, with a focus on robotics and IoT.