IRIS-HEP Fellow: Andrii Len



Fellowship dates: Jun – Sep, 2022
Jul – Sep, 2023
Jun – Aug, 2024
Jun – Aug, 2025

Home Institution: Taras Shevchenko National University of Kyiv


Project: The usage of Deep Learning for QCD background estimation

The focus of the present project is to find optimal deep learning models to be used for the separation of signal and background events.



More information: My project proposal

Mentors:
  • Ece Asilar (CERN)


Project: Predict CMS data popularity to improve its availability for physics analysis

The focus of the project is to aggregate and extract data usage information, find data’s features and optimal Machine Learning models to predict the probability that a dataset will be accessed in the next month.



More information: My project proposal

Mentors:
  • Dmytro Kovalskyi

  • Rahul Chauhan

  • Hasan Ozturk


Project: Topological Rare Hadron Decay Tagging with DNN: Deep neural net topological tagger for rare hadron decay identification

One of the main challenges of this project will be to identify and build an effective DNN architecture to train a new model that will not only match BDT in performance but gives a significant improvement to the analysis sensitivity.



More information: My project proposal

Mentors:
  • Dmytro Kovalskyi (MIT)


Project: Mitigating the Impact of Simulation Mis-Modeling on DNN Training: Building Robust DNNs in the Presence of Detector Mis-Modeling

In this project, we will compare two methods to mitigate Simulation Mis-modeling impact on training: First one is to exclude simulated data completely (use only real data for training) and the second one is to modify loss function to include penalty terms for mis-modeling. We will assess relative performance and identify common trends of these approaches to find an optimal solution.



More information: My project proposal

Mentors:
  • Dmytro Kovalskyi (MIT)

Presentations and Publications
Current Status


Contact me: