EFI Data Analytics for Physics

30 Oct 2017, 09:00 → 1 Nov 2017, 17:00 America/Chicago

https://efi.uchicago.edu/content/efi-data-analytics-physics

For hands-on sessions participants will have access to personal JupyterLab instances. To get address of your instance please go to this address, select a free instance and add your name to it.

https://goo.gl/TTrbDt

You have full freedom on it, if you manage to destroy it ask for it to be recreated.

 

Monday, 30 October

09:30 → 10:00 Coffee and tea

10:00 → 12:00 Intro session

10:00 Introduction and round table

Speaker: David Miller (University of Chicago (US))

Miller_Intro_30Oct2017.pdf

10:20 Discussion of projects for the workshop

Link to GitHub with various exercises

10:40 Integrative Cyberinfrastructure for the EFI

Speaker: Robert William Gardner Jr (University of Chicago (US))

Slides

10:50 Brief introduction to modern data analysis environments

Speaker: Dr Benedikt Riedel (University of Chicago)

Brief Introduction to Modern Analysis Environments

11:10 Hands-on session with Python, Jupyter Notebooks, and ROOT

Make sure to sign up for a particular JupyterLab instance.

Then navigate to that instance by typing the URL into your browser. The password for JupyterLab is written on the board in the room.

Speaker: Giordon Holtsberg Stark (University of Chicago (US))

Tutorial

• 0 - Overview
• 1 - JuPyter Tutorial
• 2 - Python Tutorial
• 3 - NumPy Tutorial
• 4 - Plotting Basics

12:00 → 13:00 Lunch break

13:00 → 17:30 Afternoon session: Doing stuff

13:00 Introduction to pattern recognition algorithms and Machine Learning

Speaker: Ben Nachman (Lawrence Berkeley National Lab. (US))

Nachman_IntroMachineLearning.pdf

13:45 Hands-on: 1D optimal classifier, e.g. a Gaussian process classifier, compared with a simple Neural Network

Speakers: Ben Nachman (Lawrence Berkeley National Lab. (US)), Mats Joakim Robert Olsson (University of Chicago (US))

14:30 Coffee break

15:00 Advanced machine learning application in high energy physics

Speaker: Ben Nachman (Lawrence Berkeley National Lab. (US))

Nachman_AdvancedMachineLearning.pdf

15:45 Keras deep learning images

Speaker: Mats Joakim Robert Olsson (University of Chicago (US))

16:35 Demo: Balancing a pole using "Reinforcement Learning"

Speaker: Mats Joakim Robert Olsson (University of Chicago (US))

Tuesday, 31 October

09:30 → 12:00 Morning session: Statistical Analysis Tools

09:30 Introduction to statistical analysis tools

Counting experiments: RooStats / HistFitter
Spectrum experiments: BumpHunter

Speaker: Maximilian J Swiatlowski (University of Chicago (US))

swiatlow.pdf

09:55 Hands-on with HistFitter for counting experiments

Speaker: Maximilian J Swiatlowski (University of Chicago (US))

10:20 Coffee Break

10:40 Data visualization with ROOT, D3, Bokeh, and HoloViews

Speaker: Teodora Szasz (University of Chicago Research Computing Center)

Presentation

12:00 → 13:00 Lunch break

13:00 → 18:00 Break for 50th anniversary UChicago-Fermilab Event

Wednesday, 1 November

09:30 → 12:00 Morning session: Distributed computing

09:30 Introduction to distributed computing

Speaker: Lincoln Bryant (University of Chicago (US))

Slides

09:50 Hands-on session with OSG Connect

Speaker: Benedikt Riedel (University of Chicago)

10:20 Coffee break

10:45 Introduction to Elasticsearch

Speaker: Ilija Vukotic (University of Chicago (US))

Elasticsearch - what, why, how

11:05 Hands on: using Elasticsearch for ML

Speaker: Ilija Vukotic (University of Chicago (US))

12:00 → 13:00 Lunch break

13:00 → 15:30 Afternoon session: Synthesis of previous work

13:00 Run jobs on distributed computing

13:20 MVA and/or machine learning examples

13:50 Visualizations

14:10 Wrap up