# QCD-Aware Neural Networks for Jet Physics

- Institution: CERN
- Department: Theory Colloquium
- Speaker: Kyle Stuart Cranmer
- Year: 2017
- Language: English
- Subjects: Physics
- Source: [CERN](https://videos.cern.ch/record/3002475)
- 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)

Recent progress in applying machine learning for jet physics has been built upon an analogy between calorimeters and images. In this work, we present a novel class of recursive neural networks built instead upon an analogy between QCD and natural languages. In the analogy, four-momenta are like words and the clustering history of sequential recombination jet algorithms is like the parsing of a sentence. Our approach works directly with the four-momenta of a variable-length set of particles, and the jet-based neural network topology varies on an event-by-event basis. Our experiments highlight the flexibility of our method for building task-specific jet embeddings and show that recursive architectures are significantly more accurate and data efficient than previous image-based networks. We extend the analogy from individual jets (sentences) to full events (paragraphs), and show for the first time an event-level classifier operating on all the stable particles produced in an LHC event. I will discuss future directions for this style of hybrid physics-aware machine learning algorithms. .

0:00:00 Slide 1

0:00:49 Slide 2

0:01:11 Slide 3

0:03:31 Slide 4

0:04:27 Slide 5

0:05:44 Slide 6

0:05:57 Slide 7

0:07:36 Slide 8

0:08:19 Slide 9

0:09:17 Slide 10

0:11:47 Slide 11

0:22:02 Slide 12

0:23:02 Slide 13

0:24:34 Slide 14

0:25:55 Slide 15

0:26:31 Slide 16

0:27:38 Slide 17

0:28:24 Slide 18

0:29:46 Slide 19

0:30:11 Slide 20

0:31:56 Slide 21

0:32:02 Slide 22

0:32:51 Slide 23

0:33:07 Slide 24

0:33:41 Slide 25

0:34:08 Slide 26

0:34:28 Slide 27

0:34:33 Slide 28

0:34:49 Slide 29

0:35:27 Slide 30

0:35:37 Slide 31

0:36:08 Slide 32

0:37:23 Slide 33

0:37:44 Slide 34

0:37:55 Slide 35

0:38:07 Slide 36

0:39:12 Slide 37

0:40:48 Slide 38

0:40:59 Slide 39

0:41:17 Slide 40

0:41:28 Slide 41

0:42:32 Slide 42

0:43:31 Slide 43

0:44:40 Slide 44

0:45:02 Slide 45

0:46:22 Slide 46

0:46:49 Slide 47

0:46:53 Slide 48

0:47:00 Slide 49

0:48:09 Slide 50…

## Lectures

1. [QCD-Aware Neural Networks for Jet Physics](https://invisiblenetwork.ai/learn/cern/3002475/65e62357be48), 1 h 5 min (CERN's player)