Discovering and exploiting abstract symmetry elements obtained from low-level data within the CMS experiment

Description

One of the important aspects of searches for new physics at the Large Hadron Collider (LHC) involves the identification and reconstruction of single particles, jets and event topologies of interest in collision events. The End-to-End Deep Learning (E2E) project in the CMS experiment focuses on the development of these reconstruction and identification tasks with innovative deep learning approaches.

This project explores the development of Physics-Aware Neural Networks (PANN) which require sophisticated data-efficient neural networks that can learn hidden underlying symmetries and generalise from a small dataset.

Duration

Total project length: 175/350 hours.

Task ideas

Develop a deep learning model capable of uncovering the hidden symmetries present in the dataset.
Utilise this hidden symmetry to build a physics aware neural network.

Expected results

Discover hidden symmetries present in the CMS dataset
Benchmark the models with other previous works in terms of and data efficiency and invariance with respect to symmetry operations.

Difficulty level

Advanced

Requirements

C++, Python, PyTorch, Tensorflow and some previous experience in Deep Learning.

Mentors

Diptarko Choudhury (NISER)
Sergei Gleyzer (University of Alabama)
Ruchi Chudasama (University of Alabama)
Eric Reinhardt (University of Alabama)
Samuel Campbell (University of Alabama)
Emanuele Usai (University of Alabama)

Please DO NOT contact mentors directly by email. Questions should instead be directed to ml4-sci@cern.ch which is forwarded to mentors. To submit your proposal, CV, and test task solutions, please use this Google form.