QGNN for HEP at the LHC
Google Summer of Code 23 with ML4SCI
21 Jun
Created on June 21|Last edited on June 21
Comment
Datasets
Quark Gluon
Original simulated top quark initiated jet from the paper [1]
Graph Construction
KNN Algorithm
Sample image for image to graph processing. Segmenting calorimeter cells [2]
Model Architectures
Hybrid Quantum GCNN
Uses Data Re-uploading Quantum Circuit
Hybrid neural network architecture from [3]. Extended using data reuploading circuits.
Ego QGCNN
Ego QGCNN architecture from [4]
Status
- Currently reading the paper: Hyperparameter-free and Explainable Whole Graph Embedding.
- Coding the pipeline for preprocessing for Quark gluon to graphs
- Implementing Ego QGCNN.
References
- Hariri, Ali & Dyachkova, Darya & Gleyzer, Sergei. (2021). Graph Variational Autoencoder for Detector Reconstruction and Fast Simulation in High-Energy Physics. EPJ Web of Conferences. 251. 03051. 10.1051/epjconf/202125103051.
- Graph Neural Networks in Particle Physics Jonathan Shlomi (Weizmann Inst.), Peter Battaglia, Jean-Roch Vlimant (Caltech) e-Print: 2007.13681 [hep-ex] DOI: 10.1088/2632-2153/abbf9a (publication)
- Tüysüz, C., Rieger, C., Novotny, K. et al. Hybrid quantum classical graph neural networks for particle track reconstruction. Quantum Mach. Intell. 3, 29 (2021). https://doi.org/10.1007/s42484-021-00055-9
- Ai, Xing & Zhang, Zhihong & Sun, Luzhe & Yan, Junchi & Hancock, Edwin. (2022). Decompositional Quantum Graph Neural Network.
Add a comment