Recent success in the domain of unsupervised and semi-supervised learning has been lately a pivotal factor for development of Physics Aware and Symmetry Aware Machine Learning techniques where a model learns the symmetry of a dataset as a meta task and ends up learning the physics through the same.
Although most of the symmetries that we work with for SM physics are well defined and formulated, they can be well interpreted in 4-vector or 4-momenta basis. With change of representation the symmetries become elusive and difficult to write and work with. This calls for machine learning techniques that can learn the representation of the given symmetry through the means of a conserved quantity for a given abstract representation space.
Learning these symmetries not only makes us more prepared to deal with the physics constraints in these abstract spaces and coordinates but also makes us able to build neural networks that are invariant to these symmetries. Such neural networks as seen from the existing literature are more robust, stable, interpretable and data efficient.
This project will focus on ways to learn hidden symmetries combining the works of
Total project length: 175/350 hours.
Advanced
Please use this link to access the test for this project.
Please DO NOT contact mentors directly by email. Instead, please email ml4-sci@cern.ch with Project Title and include your CV and test results. The mentors will then get in touch with you.