https://github.com/yandexdataschool/pyretina
https://github.com/yandexdataschool/pyretina
Last synced: about 1 month ago
JSON representation
- Host: GitHub
- URL: https://github.com/yandexdataschool/pyretina
- Owner: yandexdataschool
- License: apache-2.0
- Created: 2016-05-27T16:28:57.000Z (about 9 years ago)
- Default Branch: master
- Last Pushed: 2016-10-01T16:16:28.000Z (over 8 years ago)
- Last Synced: 2024-12-26T12:10:30.255Z (6 months ago)
- Language: Python
- Size: 10.3 MB
- Stars: 1
- Watchers: 5
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Artificial Retina algorithm via optimization
=======
This repo contains an implementation of the Artificial Retina algorithm for LHCb VELO tracking.
Currently contains:
- numerical optimization methods as main subroutine for local maximum search;
- multi-start as a global search method;
- 2 initial seeding algorithms;
- Retina response function as well as its gradient and Hessian matrix implemented via [theano](https://github.com/Theano/Theano) (thus supposed to be computed on a GPU or a multi-core CPU);
- simplified LHCb VELO simulation with parameters inspired by the upgrade TDR;
- utils for efficiency measuments.
Future work:
- primary vertex fitting;
- fine tuning of the meta-parameters (sigma cooling, optimizer's parameters);
- advanced initial seeding (based on hits and VELO's geometry);
- hit's timing;
- multi-stage helix curve fitting.