Machine Learning at Lebedev Physical Institute

This course has been designed to introduce the topic of Machine Learning (ML) to a High Energy Physics (HEP) audience, following a classical Data Science approach with a gradual addition of HEP specifics on the way. Although enriched with HEP (with a strong bias towards its experimental part) examples in particular, the course aims to provide a general and basic description of ML fundamentals regardless of its application.

The common platform initially was the Lebedev Physical Institute (LPI), where its predecessor took place in the spring semester of 2019, but the course's coverage is expanding. Currently, during the autumn semester of 2020 it is being taught to HEP students of MIPT and MEPhI, so the materials will be uploaded accordingly. For a more comprehensive course overview and introduction to the topic, please have a look at the kick-off lecture slides.

NB: materials of the course are provided in English, recordings and some of the references throughout the course - in Russian.

Syllabus

0. Kick-off

• materials
• recording

1. Python

• materials
• recordings: seminar (part A), seminar (part B)

2. Introduction to ML

• materials
• recordings: lecture 1, lecture 2, seminar

3. Trees

• materials
• recordings: (lecture not available), seminar

4. Neural Networks

• materials
• recordings: lecture, seminar

5. Computer Vision & Generative Models

• materials
• recordings: lecture, (seminar not available)

6. ML examples in HEP

• materials
• recordings: lecture, seminar

Prerequisites

There is not much needed to start the journey: basics of calculus, linear algebra, statistics and some understanding of programming (not necessarily Python). However, you still need to setup a few things to go ahead with coding, so please refer to these instructions where we guide you through the essential process of preparing the coding environment.

Announcements

For Russian-speaking audience: all the announcements and news related to the course are being posted in this Telegram channel. Also, there is a dedicated chat for discussions.

References

While building this course we were inspired by many other great courses, tutorials and articles. Here we tried to collect them so that one would have a chance to look up for more additional materials. Note that since it wasn't our primarily goal, this is by a large extent incomplete, unstructured and in a mixture of RU and EN compilation -- but you might find it helpful as we did.

Developers

• Oleg Filatov [DESY] ([email protected])
• Stepan Zakharov [NSU] ([email protected])
• Vladimir Bocharnikov [DESY, MEPhI] ([email protected])
• Olga Razuvaeva [MEPhI, ITEP] ([email protected])
• Sergey Korpachev [MIPT, LPI RAS] ([email protected])
• Andrey Filatov [MIPT] ([email protected])
• Daniil Yakovlev [HSE, YSDA] ([email protected])
• Kirill Bukin [MIPT] ([email protected])