Machine Learning is an expanding field with enormous potential for Physics applications. In particular, Deep-Learning methods can be used for improved data processing, signal analysis, and modeling. Some of these methods have already been successfully applied in fields ranging from particle physics and astrophysics to medical physics, but much of their scope and possible applications in physics are yet to be explored.
A proper understanding and a fair use of the machine-learning toolkit will be more and more necessary in research. In this context, the Institute of Particle and Cosmos Physics (IPARCOS) seeks to promote the knowledge of this type of techniques among its researchers, and among the university community in general.
This first "Machine Learning and Applications to Physics" workshop has a twofold objective:
- To carry out an introductory course on Machine Learning techniques, including theory and hands-on sessions, for all those interested in starting to use them.
- To learn from both national and international researchers, the use they are making of Machine Learning techniques in various areas of Physics.
It is going to be carried out in three days. In the morning there will be a theory session and talks from different experts working on this field and the afternoon will be devoted to a hands-on session.
The introductory course is structured as follows:
- Introduction to machine learning: basic concepts and methodology overview, by Prof. Pedro J. Zufiria Zatarain (UPM, Madrid)
- Deep learning and applications to Physics, by Prof. Juan Rojo (UV, Amsterdam)
- Bayesian methods and approximate bayesian inference, by Ekaterina Lobacheva (HSE University, Moscow)
- Bayesian neural networks, by Prof. Nadezhda Chirkova (HSE University, Moscow)
It is mainly focused on researchers from the Faculty of Physical Sciences of the UCM, but it is also open to researchers from other universities and research centers.