Anomaly Detection Using Contrastive Learning (Electron-Photon)

Overview:

This project aims to detect Electrons (Anomaly) by training the model only on the Photons using Contrastive Learning.

Dataset:

For the top quark dataset:

https://drive.google.com/drive/folders/1WXc1-wetvaiufNzAcVg23DEBK2QBYhp9?usp=sharing

And the Electron-Photon dataset can be downloaded from these links:

https://cernbox.cern.ch/index.php/s/sHjzCNFTFxutYCj/download

https://cernbox.cern.ch/index.php/s/69nGEZjOy3xGxBq/download

Requirements:

Strong knowledge of Python; Keras.

No need for a strong math background.

Note:

The code is still under verification.

This project was mentored by Prof. Sergei V. Gleyzer and Ali Hariri.