Machine Learning

Name: Machine Learning
Start: 2019-05-20T11:00:00+02:00
End: 2019-05-22T16:00:00+02:00
Location: Camogli

20 May 2019, 11:00 → 22 May 2019, 16:00 Europe/Rome

Hotel Cenobio dei Dogi (Camogli)

Hotel Cenobio dei Dogi

Camogli

Via N. Cuneo, 34

Description

Supporto

battaglieri@ge.infn.it

Monday, 20 May
- 11:00 → 11:30
  
  Introduction 30m
- 11:30 → 13:00
  
  Lecture 1 - Overview on ML/DL in HEP 1h 30m
  
  This lecture aims at giving a broad overview of main concepts in ML in the context of HEP applications.
  
  Speaker: Amir Farbin (University of Texas at Arlington)
  
  INFN-ML-Lecture-1.pdf
- 13:00 → 14:30
  
  Lunch break 1h 30m
  
  On your own
- 14:30 → 16:00
  
  Lecture 2 - Introduction to the ML landscape and methods 1h 30m
  
  An introduction to the (general, not HEP specific) ML landscape is given.
  
  Speaker: Daniele Bonacorsi (BO)
  
  Camogli2019__DBonacorsi_MLlandscape.pdf
- 16:00 → 16:30
  
  Coffee break 30m
- 16:30 → 17:30
  
  Hands-on session 1h
  
  github for the hands-on part
- 17:30 → 18:15
  
  Compute vision and applications 45m
  
  Speaker: Dr Ferrari A. (Argo Vision)
  
  Dataset per training e test
  
  MiniGuida alla configurazione di OpenCV
  
  Slides - full video
  
  Slides - no video
- 18:15 → 18:40
  
  ML in the online data acquisition 25m
  
  Speaker: Cristiano Fanelli (ISS)
  
  ML_online_data_acquisition-C.Fanelli.pdf
- 18:40 → 19:05
  
  ML for the EIC 25m
  
  Speaker: Dmitry Romanov (Jefferson Lab)
  
  EIC ML.pptx
Tuesday, 21 May
- 09:00 → 10:30
  
  Lecture 3 - End-2-end projects discussion: regression and classification 1h 30m
  
  The attendees will be guided through an example of regression techniques.
  
  Speaker: Daniele Bonacorsi (BO)
  
  Camogli2019__DBonacorsi_e2eProjectClassification.pdf
  
  Camogli2019__DBonacorsi_e2eProjectRegression.pdf
- 10:30 → 11:00
  
  Coffee break 30m
- 11:00 → 12:30
  
  Lecture 4 - Rapid walk-through of Deep Learning 1h 30m
  
  The attendees will be guided through concepts and methods in Deep Learning.
  
  Speaker: Amir Farbin (University of Texas at Arlington)
  
  INFL-ML-Lecture-2.pdf
- 12:30 → 14:30
  
  Lunch break 2h
  
  On your own
- 14:30 → 15:30
  
  Hands-on session (cont'd) 1h
- 15:30 → 16:00
  
  Coffee break 30m
- 16:00 → 17:00
  
  Hands-on session (cont'd) 1h
- 17:00 → 17:45
  
  Graph Network for High Energy Physics 45m
  
  Speaker: Dr Jean-Roch Vlimant (Caltech)
  
  vlimant_CamoglieINFN_May19.pdf
- 17:45 → 18:30
  
  Improving Reaction Identification with ROOT TMVA 45m
  
  Speaker: Derek Glazier (University of Edinburgh)
  
  ReactionID.pdf
Wednesday, 22 May
- 09:00 → 10:30
  
  Lecture 5 - Training models and Gradient Descent (bonus: Decision Trees) 1h 30m
  
  The attendees will be guided through what's under the hood in terms of training models. An additional lecture on Decision Trees is posted.
  
  Speaker: Daniele Bonacorsi (BO)
  
  Camogli2019__DBonacorsi_DecisionTrees.pdf
  
  Camogli2019__DBonacorsi_TrainingModels.pdf
- 10:30 → 11:00
  
  Coffee break 30m
- 11:00 → 12:30
  
  Lecture 6 - Keras and/or Deep Learning in HEP 1h 30m
  
  The attendes opted to be given a guided walk-through of the basics of Keras as from keras.io. Alternative choice was a seminar on Deep Learning in HEP, whose slides are anyway posted.
  
  Speaker: Amir Farbin (University of Texas at Arlington)
  
  INFN-ML-Lecture-3.pdf
- 12:30 → 14:30
  
  Lunch break 2h
  
  On your own
- 14:30 → 16:00
  
  Closeout 1h 30m
  
  The closeout includes speakers presenting references and post-course self-training suggestions, as well as a discussion of attendees' open projects.
  
  Post-course references