# Recursive Neural Networks in Quark/Gluon Tagging

- Institution: CERN
- Department: Machine Learning
- Speaker: Taoli Cheng
- Year: 2018
- Language: English
- Subjects: Computer science
- Source: [CERN](https://videos.cern.ch/record/3009889)
- Licence: [CERN](http://copyright.web.cern.ch)
- Who can watch: Tweens profiles and older (sources give no ratings; placed by Invisible Network from the source's own words)

Vidyo contribution

Based on the natural tree-like structure of jet sequential clustering, the recursive neural networks (RecNNs) embed jet clustering history recursively as in natural language processing. We explore the performance of RecNN in quark/gluon discrimination. The results show that RecNNs work better than the baseline BDT by a few percent in gluon rejection at the working point of 50\% quark acceptance. We also experimented on some relevant aspects which might influence the performance of networks. It shows that even only particle flow identification as input feature without any extra information on momentum or angular position is already giving a fairly good result, which indicates that most of the information for q/g discrimination is already included in the tree-structure itself..

0:00:00 Slide 1

0:00:14 Slide 2

0:01:18 Slide 3

0:01:29 Slide 4

0:02:05 Slide 5

0:03:37 Slide 6

0:06:25 Slide 7

0:07:12 Slide 8

0:08:39 Slide 9

0:10:29 Slide 10

0:11:33 Slide 11

0:12:20 Slide 12

0:12:53 Slide 13

0:14:07 Slide 14

0:16:41 Slide 15

0:17:28 Slide 16

0:17:56 Slide 17

0:18:36 Slide 18

0:20:38 Slide 19

0:21:32 Slide 20

0:22:03 Slide 21

0:22:44 Slide 22

0:23:32 Slide 23

0:25:01 Slide 24

0:26:34 Slide 25

0:29:35 Slide 26

0:29:40 Slide 27

## Lectures

1. [Recursive Neural Networks in Quark/Gluon Tagging](https://invisiblenetwork.ai/learn/cern/3009889/7c5eb35835a9), 33 min (CERN's player)