# How the machine learning conquers reconstruction in neutrino experiments

- Institution: CERN
- Department: EP-IT Data science seminars
- Speaker: Robert Sulej
- Year: 2017
- Language: English
- Subjects: Computer science
- Source: [CERN](https://videos.cern.ch/record/3008282)
- 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)

An evolution from the purely algorithmic approaches towards the machine learning solutions started a few years ago in the neutrino experiments. Now, this process turns into a true boom, especially in the experiments based on the imaging technologies, such as LArTPC's used in MicroBooNE and DUNE experiments or liquid scintillator detector implemented by the NOvA Collaboration. High resolution, image-like projections of events obtained with these detectors proved to be hard pattern recognition problems for the conventional reconstruction techniques. In the seminar, I will present why the neutrino events are so challenging and how the essential difficulties are now being attacked with the machine learning. .

0:00:00 Slide 1

0:02:23 Slide 2

0:03:54 Slide 3

0:04:30 Slide 4

0:06:31 Slide 5

0:07:14 Slide 6

0:07:51 Slide 7

0:08:22 Slide 8

0:09:20 Slide 9

0:09:40 Slide 10

0:10:30 Slide 11

0:11:54 Slide 12

0:12:51 Slide 13

0:13:07 Slide 14

0:14:24 Slide 15

0:14:31 Slide 16

0:16:00 Slide 17

0:17:04 Slide 18

0:17:47 Slide 19

0:18:05 Slide 20

0:18:48 Slide 21

0:20:05 Slide 22

0:20:57 Slide 23

0:22:27 Slide 24

0:26:22 Slide 25

0:27:22 Slide 26

0:28:24 Slide 27

0:30:13 Slide 28

0:30:46 Slide 29

0:32:02 Slide 30

0:33:00 Slide 31

0:34:19 Slide 32

0:35:50 Slide 33

0:37:04 Slide 34

0:37:44 Slide 35

0:37:56 Slide 36

0:38:02 Slide 37

0:38:07 Slide 38

0:38:18 Slide 39

0:39:46 Slide 40

0:42:48 Slide 41

0:47:41 Slide 42

0:52:10 Slide 43

## Lectures

1. [How the machine learning conquers reconstruction in neutrino experiments](https://invisiblenetwork.ai/learn/cern/3008282/175e40a5a92f), 56 min (CERN's player)