IBM Research - Ireland Internship Project: Machine learning to forecast dynamic physical systems - overview
Machine learning is hugely successful in learning a set of features that relate an input vector x to an output vector y. A more challenging case exists when the input vector is correlated in time – i.e. the output vector y depends on both x and previous values of x. This is a common problem in physics where the temporal relation is generally encoded via a set of partial differential equations.
This project will apply deep neural networks to the problem and develop a forecasting model for fast-changing physical systems.
- Knowledge of machine learning
- Programming in Python