Second Machine Learning in High Energy Physics Summer School 2016

Europe/Copenhagen
Lundmarkssalen (Lund University)

Lundmarkssalen

Lund University

Lund University, Lund, Sweden
Andrey Ustyuzhanin (Yandex School of Data Analysis (RU)), Caterina Doglioni (Lund University (SE))
Description

The Second Machine Learning summer school organized by Yandex School of Data Analysis and Laboratory of Methods for Big Data Analysis of National Research University Higher School of Economics will be held in Lund, Sweden from 20 to 26 June 2016. It is hosted by Lund University.

The school is intended to cover the relatively young area of data analysis and computational research that has started to emerge in High Energy Physics (HEP). It is known by several names including “Multivariate Analysis”, “Neural Networks”, “Classification/Clusterization techniques”. In more generic terms, these techniques belong to the field of “Machine Learning”, which is an area that is based on research performed in Statistics and has received a lot of attention from the Data Science community.

There are plenty of essential problems in High energy Physics that can be solved using Machine Learning methods. These vary from online data filtering and reconstruction to offline data analysis.

Students of the school will receive a theoretical and practical introduction to this new field and will be able to apply acquired knowledge to solve their own problems. Topics ranging from decision trees to deep larning and hyperparameter optimization will be covered with concrete examples and hands-on tutorials. A special data-science competition will be organized within the school to allow participants to get better feeling of real-life ML applications scenarios.

The MLHEP school is a satellite event to the LHCP2016 conference, so its dates and venue (Lund University) are well-aligned with the conference.

Expected number of students for the school is 40-50 people.

Pre-requisites for participation

Upon completion of the school participants would be able to

  • formulate a HEP-related problem in ML-friendly terms
  • select quality criteria for a given problem
  • understand and apply principles of widely-used classification models (e.g. boosting, bagging, BDT, neural networks, etc) to practical cases
  • optimize features and parameters of a given model in efficient way under given restrictions
  • select the best classifier implementation amongst a variety of ML libraries (scikit-learn, xgboost, deep learning libraries, etc)
  • define & conduct reproducible data-driven experiments

Github repository, with material and slides from the school

https://github.com/yandexdataschool/mlhep2016

Organizers

Partners

Local information
Registration
Accommodation & dinners
    • Organisational: Welcome Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Andrey Ustyuzhanin (Yandex School of Data Analysis (RU))
      • 1
        Registration
      • 2
        Welcome to MLHEP
        Speaker: Caterina Doglioni (Lund University (SE))
      • 3
        Competition introduction
      • 10:05
        Break
    • Lectures: Day 1 lectures Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Aleksei Rogozhnikov (Yandex School of Data Analysis (RU))
      • 4
        Intro: General pipeline, ML at a glance, model evaluation

        Featuring cross-validation and ROC AUC

      • 11:40
        Break
      • 5
        Metric ML algorithms

        SVM, KNN, Linear regression

    • Organisational: Lunch Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
    • Seminars: Day 1 seminars Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Nikita Kazeev (Yandex School of Data Analysis (RU))
      • 6
        Course technicalities, working environment
      • 15:55
        Break
      • 7
        Python for data analysis; Probability density estimation

        numpy, root_numpy, pandas, matplotlib

    • Organisational: Break Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
    • Invited lectures: Jet parton and particle identification and applications Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: J Michael Williams (Massachusetts Inst. of Technology (US))
    • Organisational: Welcome dinner http://www.stadsparkscafeet.se/

      http://www.stadsparkscafeet.se/

      Stadsparken i Lund Stadsparksgången 222 29
    • Lectures: Day 2 lectures Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Aleksei Rogozhnikov (Yandex School of Data Analysis (RU))
      • 8
        Decision trees
      • 10:50
        Break
      • 9
        Ensembles

        bagging & boosting

    • Organisational: Lunch Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
    • Seminars: Day 2 seminrs Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Nikita Kazeev (Yandex School of Data Analysis (RU))
      • 10
        Model evaluation

        Overfitting, Cross-validation

      • 14:50
        Break
      • 11
        sklearn, simple algorithms
    • Organisational: Break Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
    • Invited lectures: Online, Collaborative Machine Learning Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Dr Joaquin Vanschoren (Eindhoven University of Technology)
    • Lectures: Day 3 lectures Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Aleksei Rogozhnikov (Yandex School of Data Analysis (RU))
      • 12
        Feature engineering, Dimensionality reduction
      • 10:50
        Break
    • Organisational: Lunch Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
    • Seminars: Day 3 seminars Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Nikita Kazeev (Yandex School of Data Analysis (RU))
      • 14
        Ensemble algorithms, dimensionality reduction

        Random forest, gradient boosting, PCA

      • 14:50
        Break
    • Invited lectures: Data Doping solution to the "Flavour of Physics" Kaggle challenge Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Dr Vicens Gaitan (Grupo AIA R&D Director)
    • Organisational: Break Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
    • Invited lectures: Approximating Likelihood Ratios with Calibrated Classifiers (TBC) Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Gilles Louppe (New York University (US))
    • Lectures: Day 4 lectures Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Tatiana Likhomanenko (National Research Centre Kurchatov Institute (RU))
      • 15
        Boosting reweighting and flatness
      • 10:50
        Break
    • Seminars: Boosting reweighting and flatness - seminar Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Tatiana Likhomanenko (National Research Centre Kurchatov Institute (RU))
    • Organisational: Lunch Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
    • Seminars: Day 4 seminars H418, RYDBERGSALEN (Lund Univeristy)

      H418, RYDBERGSALEN

      Lund Univeristy

      Sölvegatan 14, 223 62 Lund
      • 16
        Feature engineering and selection

        Flatness boosting and reweighting

        Speaker: Nikita Kazeev (Yandex School of Data Analysis (RU))
      • 15:20
        Break
      • 17
        Shallow neural networks
        Speaker: Alexander Panin (Yandex School of Data Analysis (RU))
      • 17:00
        Break
      • 18
        Deep neural networks
        Speaker: Alexander Panin (Yandex School of Data Analysis (RU))
    • Midsommar Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden

      From https://sweden.se/culture-traditions/midsummer/ In mid-June, school is out and nature has burst into life. It seems like the sun never sets. In fact, in the north of Sweden it doesn’t, and in the south only for an hour or two. This calls for celebration! Friends and family gather for the most typically Swedish tradition of all: Midsummer.

    • Lectures: Deep learning Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Alexander Panin (Yandex School of Data Analysis (RU))
      • 19
        Deep learning I
        Speaker: Mr Alexander Panin (Yandex School of Data Analysis (RU))
      • 10:50
        Break
      • 20
        Deep learning II
        Speaker: Mr Alexander Panin (Yandex School of Data Analysis (RU))
    • Organisational: Lunch Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
    • Seminars: Deep learning Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Alexander Panin (Yandex School of Data Analysis (RU))
      • 21
        Deep learning III
        Speaker: Mr Alexander Panin (Yandex School of Data Analysis (RU))
      • 14:50
        Break
      • 22
        Deep learning IV
        Speaker: Mr Alexander Panin (Yandex School of Data Analysis (RU))
    • Organisational: Break Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
    • Invited lectures: Ultra High Energy Cosmic Rays and the CRAYFIS Experiment Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      Convener: Chase Owen Shimmin (University of California Irvine (US))
    • Challenge hacking & conclusion Lundmarkssalen

      Lundmarkssalen

      Lund University

      Lund University, Lund, Sweden
      • 23
        Challenge hacking
        Time to apply all the knowledge and crack the challenge. Lecturers are available for questions. Feel free to grab coffee and lunch when it suits you.
      • 10:50
        Break
      • 24
        Challenge hacking
      • 14:30
        Break
      • 25
        Presentations preparation
        Time to prepare the solution presentation.
      • 26
        Awards and best solutions presentations