Name: MLHEP 2015 summer school
Start: 2015-08-27T08:00:00+03:00
End: 2015-08-30T23:00:00+03:00
Location: Academic University

MLHEP 2015 summer school

from Thursday, 27 August 2015 (08:00) to Sunday, 30 August 2015 (23:00)

Monday, 24 August 2015
Tuesday, 25 August 2015
Wednesday, 26 August 2015
Thursday, 27 August 2015

08:00 Registration

08:00 - 08:15

08:15 breakfast
breakfast
08:15 - 09:00
09:00 Welcome
09:00 - 09:35
Room: 204
Contributions

09:00 Welcome to MLHEP

09:20 MLHEP Kaggle competition - Nikita Kazeev (Yandex School of Data Analysis (RU))
09:35 break
break
09:35 - 09:40
09:40 Advanced track: lectures
lectures
09:40 - 12:40
Room: 206
Contributions

09:40 Reminder about major algorithms. Advanced aspects of their use. Model evaluation.

11:20 Feature selection. Regularization.
Introductory track: lectures
lectures
09:40 - 12:40
Room: 204
Contributions

09:40 Intro: General pipeline, ML at a glance

11:20 Intro: kNN, basic overfitting, roc curve, logistic regression
12:40 lunch
lunch
12:40 - 13:45
13:45 Advanced track: seminars
seminars
13:45 - 17:10
Room: 206
Contributions

13:45 MLHEP introduction. ML vs MLHEP: correlation and agreement restrictions - Tatiana Likhomanenko (National Research Centre Kurchatov Institute (RU))

15:40 sPlot technique. ML on sPlot data.
Introductory track: seminars
seminars
13:45 - 17:10
Room: 204
Contributions

13:45 Introduction, course technicalities

15:40 NumPy, Pandas, Matplotlib
17:10 break
break
17:10 - 17:20
17:20 Applying Machine Learning to LHC triggers optimization - Mika Anton Vesterinen (Ruprecht-Karls-Universitaet Heidelberg (DE))
Applying Machine Learning to LHC triggers optimization
- Mika Anton Vesterinen (Ruprecht-Karls-Universitaet Heidelberg (DE))
17:20 - 19:00
Room: 204
19:00 Welcome dinner
Welcome dinner
19:00 - 20:30
Friday, 28 August 2015

08:15 breakfast
breakfast
08:15 - 09:30
09:30 Advanced track: lectures #2
lectures #2
09:30 - 12:30
Room: 206
Contributions

09:30 model ensembling #1

11:10 model ensembling #2
Introductory track: lectures #2
lectures #2
09:30 - 12:30
Room: 204
Contributions

09:30 neural networks

11:10 decision tree, regression tree
12:30 lunch
lunch
12:30 - 13:45
13:45 Advanced track: seminars #2
seminars #2
13:45 - 17:10
Room: 206
Contributions

13:45 Brave new boosting world: flatness boosting

15:40 Brave new boosting world: reweighing boosting. Iterative learning
Introductory track: seminars #2
seminars #2
13:45 - 17:10
Room: 204
Contributions

13:45 Overfitting, Cross-validation

15:40 Sklearn: KNN, decision tree, logistic regression
17:10 break
break
17:10 - 17:30
17:30 Challenges of searching for physics signatures with a large amount of background - Patrick Haworth Owen (Imperial College Sci., Tech. & Med. (GB))
Challenges of searching for physics signatures with a large amount of background
- Patrick Haworth Owen (Imperial College Sci., Tech. & Med. (GB))
17:30 - 19:10
Room: 204
Saturday, 29 August 2015

08:15 breakfast
breakfast
08:15 - 09:30
09:30 Advanced track: lectures #3
lectures #3
09:30 - 12:30
Room: 206
Contributions

09:30 Linear dimensionality reduction.

11:10 Non-linear dimensionality reduction. Kernel trick. Common kernels.
Introductory track: lectures #3
lectures #3
09:30 - 12:30
Room: 204
Contributions

09:30 Ensembles

11:10 Gradient boosting; boosting to uniformity.
12:30 lunch
lunch
12:30 - 13:45
13:45 Advanced track: seminars #3
seminars #3
13:45 - 17:10
Room: 206
Contributions

13:45 Meta Algorithms Applications in HEP.

15:40 Hypotheses testing.
Introductory track: seminars #3
seminars #3
13:45 - 17:10
Room: 204
Contributions

13:45 Ensemble algorithms: random forest, gradient boosting

15:40 Hyperparam Optimization
17:10 break
break
17:10 - 17:30
17:30 MVA vs cut-and-count techniques in the SM model processes searches and measurements - Lesya Shchutska (University of Florida (US))
MVA vs cut-and-count techniques in the SM model processes searches and measurements
- Lesya Shchutska (University of Florida (US))
17:30 - 19:10
Room: 204
Sunday, 30 August 2015

08:15 breakfast
breakfast
08:15 - 09:00
09:00 Advanced track: lectures #4
lectures #4
09:00 - 12:00
Room: 206
Contributions

09:00 Kernelized algorithms: SVM, regression, K-NN, PCA and their properties.

10:40 Deep learning. Different kinds of learning.
Introductory track: lectures #4
lectures #4
09:00 - 12:00
Room: 204
Contributions

09:00 GBDT tuning, reweighting, testing hypotheses, gaussian processes for optimization

10:40 Unsupervised learning, Latent Dirichlet Allocation, PCA. Fast predictions and pruning
12:00 lunch
lunch
12:00 - 13:00
13:00 Advanced track: seminars #4
seminars #4
13:00 - 16:15
Room: 206
Contributions

13:00 Summary: HEP ML analysis sketch, from data to discovery.

14:45 My own first NN using theano.
Introductory track: seminars #4
seminars #4
13:00 - 16:15
Room: 204
Contributions

13:00 PCA, RBM and alike on MNIST

14:45 Clustering
16:15 break
break
16:15 - 16:20
16:20 Multivariate techniques in the Higgs search - Josh Bendavid (CERN)
Multivariate techniques in the Higgs search
- Josh Bendavid (CERN)
16:20 - 18:00
Room: 204
18:00 transfer to the boat
transfer to the boat
18:00 - 19:00

19:00 channel cruise + closing dinner + awards
channel cruise + closing dinner + awards
19:00 - 22:00

22:00 transfer to the hotel
transfer to the hotel
22:00 - 23:00