# Learning Invariant Representations using Mutual Information Regularization

- Institution: CERN
- Department: LPCC Workshops
- Speaker: Justin Tan
- Year: 2019
- Language: English
- Subjects: Computer science
- Source: [CERN](https://videos.cern.ch/record/3016933)
- 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)

Invariance of learned representations of neural networks against certain sensitive attributes of the input data is a desirable trait in many modern-day applications of machine learning, such as precision measurements in experimental high-energy physics and enforcing algorithmic fairness in the social and financial domain. We present a method for enforcing this invariance through regularization of the mutual information between the target variable and the classifier output. Applications of the proposed technique to rare decay searches in experimental high-energy physics are presented, and demonstrate improvement in statistical significance over conventionally trained neural networks and classical machine learning techniques..

0:00:00 Slide 1

0:00:15 Slide 2

0:02:01 Slide 3

0:03:00 Slide 4

0:04:51 Slide 5

0:06:25 Slide 6

0:07:56 Slide 7

0:08:20 Slide 8

0:09:46 Slide 9

0:11:30 Slide 10

0:13:26 Slide 11

0:14:04 Slide 12

0:14:40 Slide 13

0:16:09 Slide 14

0:16:57 Slide 15

0:17:22 Slide 16

0:18:02 Slide 17

0:18:30 Slide 18

0:18:49 Slide 19

0:19:05 Slide 20

0:19:10 Slide 21

0:19:14 Slide 22

0:19:48 Slide 23

0:20:06 Slide 24

0:20:20 Slide 25

0:21:03 Slide 26

## Lectures

1. [Learning Invariant Representations using Mutual Information Regularization](https://invisiblenetwork.ai/learn/cern/3016933/860f2b50ba8c), 23 min (CERN's player)