Third Machine Learning in High Energy Physics Summer School 2017

Europe/London
Reading University (Reading)

Reading University

Reading

Reading University, Reading, United Kingdom
Andrey Ustyuzhanin (Yandex School of Data Analysis (RU)), Ulrik Egede (Imperial College (GB))
Description

The Third Machine Learning summer school organized by Yandex School of Data AnalysisLaboratory of Methods for Big Data Analysis of National Research University Higher School of Economics and Imperial College London will be held in Reading, UK from 17 to 23 July 2017.

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 learning 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.

Expected number of students for the school is 50-60 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

School competition

Details on the school Machine Learning competition are provided on the dedicated page.

Organizers

Imperial College London

Partners

Sponsors

The school is sponsored by Winton which allows us to provide a subsidy for students who are not able to afford the registration fee out of their own funds. See registration fee page for details. 

GPU resources in the cloud are provided by Microsoft Azure.

    • 08:50 09:00
      Organisational: Opening words Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      Conveners: Andrey Ustyuzhanin (Yandex School of Data Analysis (RU)), Ulrik Egede (Imperial College (GB))
      • 08:50
        'Welcome to MLHEP' Opening Words 10m
        Speakers: Andrey Ustyuzhanin (Yandex School of Data Analysis (RU)), Ulrik Egede (Imperial College (GB))
    • 09:00 10:30
      Lectures: Day 1 Lectures Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      Convener: Alexey Artemov (Yandex)
      • 09:00
        Lecture 1. Introduction. Why ML works. Overfitting. Model Selection. Figures of Merits. Linear Models. Regularization. Logistic Regression. 1h 30m
        Speaker: Alexey Artemov (Yandex)
    • 10:30 10:50
      Coffee break 20m Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 10:50 11:10
      Practice: challenge & industry: Competition Introduction Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      Convener: Andrey Ustyuzhanin (Yandex School of Data Analysis (RU))
      • 10:50
        Competition Introduction 20m
        Speaker: Andrey Ustyuzhanin (Yandex School of Data Analysis (RU))
    • 11:10 12:40
      Seminars: Day 1 Seminars Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      Convener: Alexander Panin (Yandex School of Data Analysis (RU))
      • 11:10
        Seminar 1. Python data crunching: numpy, root_numpy, pandas (very short), matplotlib. 1h 30m
        Speaker: Alexander Panin (Yandex School of Data Analysis (RU))
    • 12:40 14:00
      Organisational: Lunch Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 14:00 15:30
      Lectures: Day 1 Lectures Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      Convener: Alexey Artemov (Yandex)
      • 14:00
        Lecture 2. Decision Trees. Bagging. Ensembles. RandomForest. AdaBoost. GB-Reweighting. Gradient Boosting. Decorrelation of features with predictions. 1h 30m
        Speaker: Alexey Artemov (Yandex)
    • 15:30 15:50
      Coffee break 20m Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 15:50 17:20
      Seminars: Day 1 Seminars Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      Convener: Alexander Panin (Yandex School of Data Analysis (RU))
      • 15:50
        Seminar 2. Matplotlib, Trees & Linear Models in SciKit Learn. Overfitting checks & prevention. 1h 30m
        Speaker: Alexander Panin (Yandex School of Data Analysis (RU))
    • 19:00 20:30
      Organisational: Welcome Reception Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 09:00 10:30
      Lectures: Day 2 Lectures Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      • 09:00
        Lecture 3. Ensembling: Adaboost, Stacking. Gradient Boosting. 1h 30m
    • 10:30 11:00
      Coffee break 30m Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 11:00 12:30
      Seminars: Day 2 Seminars Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      • 11:00
        Seminar 3. Ensembling: Adaboost, Stacking. Gradient Boosting. 1h 30m
    • 12:30 13:30
      Organisational: Lunch Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 13:30 15:00
      Lectures: Day 2 Lectures Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      • 13:30
        Lecture 4. Hyperparameters optimization techniques and methods. Feature selection. TPE. Gaussian Processes 1h 30m
    • 15:00 15:30
      Coffee break 30m Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 15:30 17:00
      Seminars: Day 2 seminrs Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      • 15:30
        Seminar 4. Optimization of hyper-parameters, applications to challenge. 1h 30m
    • 17:00 17:15
      Organisational: Break Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 17:15 18:45
      Invited lectures: Mike Williams Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 09:00 21:00
      Free day Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 09:00 10:30
      Lectures: Day 4 Lectures Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      • 09:00
        Lecture 5. Introduction into Neural Network models. Multi-layered Perceptron. Backpropagation. 1h 30m
    • 10:30 11:00
      Coffee break 30m Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 11:00 12:30
      Seminars: Day 4 Seminars Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      • 11:00
        Seminar 5. Computing gradient by hand. Tensorflow. Keras. Remote machine setup. Getting familiar with GPU. 1h 30m
    • 12:30 13:30
      Organisational: Lunch Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 13:30 15:00
      Lectures: Day 4 Lectures Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      • 13:30
        Lecture 6. Intro into Deep learning. Convolutional Neural Network. Model zoo. Augmentation. 1h 30m
    • 15:00 15:30
      Coffee break 30m Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 15:30 17:00
      Seminars: Day 4 Seminars Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      • 15:30
        Seminar 6. Image recognition using convolutional neural networks. 1h 30m
    • 17:00 17:15
      Organisational: Break Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 17:15 18:45
      Invited lectures: Guest Lecture Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      • 17:15
        Guest Lecture 1h 30m
        Speaker: Noel Dawe (University of Melbourne (AU))
    • 09:00 10:30
      Practice: challenge & industry: ML Projects & Problems @ HEP Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 10:30 11:00
      Coffee break 30m Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 11:00 12:30
      Practice: challenge & industry: ML at industry Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 12:30 13:30
      Organisational: Lunch Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 13:30 15:00
      Lectures: Day 5 Lectures Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      • 13:30
        Lecture 7. Dimensionality Reduction. PCA. LDA. LLE. TSNE. Autoencoders. 1h 30m
    • 15:00 15:30
      Coffee break 30m Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 15:30 17:00
      Seminars: Day 5 Seminars Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      • 15:30
        Seminar 7. PCA, Autoencoders (faces) 1h 30m
    • 17:00 17:15
      Organisational: Break Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 17:15 18:45
      Invited lectures: Tim Head Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 19:00 21:00
      Dinner 2h

      School dinner at the Bel & Dragon placed at Blake's lock on the river Kennet. https://goo.gl/maps/UFhMauYXAs12

    • 09:00 10:30
      Lectures: Day 6 Lectures Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      • 09:00
        Lecture 8 Generative Adversarial Networks, Metric Learning. 1h 30m
    • 10:30 11:00
      Coffee break 30m Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 11:00 12:30
      Seminars: Day 6 Seminars Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      • 11:00
        Seminar 8 Generative Adversarial Networks (physics) 1h 30m
    • 12:30 13:30
      Organisational: Lunch Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 13:30 16:00
      Lectures: Day 6 Lectures Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
      • 13:30
        Lecture 9. RNNs. 1h 30m
      • 15:00
        Coffee break 30m
      • 15:30
        Certificates, Free Discussion 30m
    • 09:00 10:30
      Lecture 10. Why Deep Learning works. Tips & tweaks. 1h 30m Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 10:30 11:00
      Break 30m Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom
    • 11:00 12:00
      Practice: challenge & industry: Awards & Winning solution presentation Reading University

      Reading University

      Reading

      Reading University, Reading, United Kingdom