TALENT School 2021 | ONLINE
ECT* - Villa Tambosi
Strada delle Tabarelle, 286
Trento - Italy
Machine Learning (ML) is one of the most exciting and dynamic areas of modern research and application. The purpose of this Nuclear Talent course is to provide an introduction to the core concepts and tools of machine learning in a manner easily understood and intuitive to physicists and nuclear physicists in particular. We will start with some of the basic methods from supervised learning, such as various regression methods before we move into deep learning methods for both supervised and unsupervised learning.
Organizers
-
Daniel Bazin (Michigan State University)
-
Morten Hjorth-Jensen (Michigan State University & University of Oslo)
-
Michelle Kuchera (Davidson College)
-
Sean Liddick (Michigan State University)
- Raghuram Ramanujan (Davidson College)
Student Coordination
-
Morten Hjorth-Jensen (Michigan State University & University of Oslo)
Registration
Registration no longer available.
Secretariat
Barbara Gazzoli (ECT*)
Schedule
The course will be taught as an online intensive course of duration of two weeks, with a total time of 20 h of lectures and 10 h of exercises, questions and answers. Videos and digital learning material will be made available one week before the course begins. It is possible to work on a final assignment of 2 weeks of work. The total load will be approximately 80 hours, corresponding to 5 ECTS in Europe.
Time and Activity
- 2pm-4pm (Central European time=CET) Lectures, project relevant information and directed exercises
- 5pm-6pm (CET) Questions and answers, Computational projects, exercises and hands-on sessions
The tentative (prone to revisions) lecture plan is as follows
Week 1
- Monday Linear Regression and intro to statistical data analysis (Morten Hjorth-Jensen MHJ). Learning slides at https://nucleartalent.github.io/MachineLearningECT/doc/pub/Introduction/html/Introduction.html and https://nucleartalent.github.io/MachineLearningECT/doc/pub/Day1/html/Day1.html
- Tuesday Logistic Regression and classification problems, intro to gradient methods (MHJ). Learning slides at https://nucleartalent.github.io/MachineLearningECT/doc/pub/Day2/html/Day2.html
- Wednesday Decision Trees, Random Forests and Boosting methods (MHJ). Learning slides at https://nucleartalent.github.io/MachineLearningECT/doc/pub/Day3/html/Day3.html
- Thursday Basics of Neural Networks and writing your own Neural Network code (MHJ). Learning slides at https://nucleartalent.github.io/MachineLearningECT/doc/pub/Day4/html/Day4.html
- Friday Beta-decay experiments, how to analyze various events, with hands-on examples . (Sean Liddick) Videos and teaching material https://nucleartalent.github.io/MachineLearningECT/doc/pub/Day5/html/Day5.html.
Week 2
- Monday Neural Networks and Deep Learning (Raghu Ramanujan, RR). PDF file of the presented slides at https://nucleartalent.github.io/MachineLearningECT/doc/pub/Day6/pdf/Day6.pdf. Jupter-Notebook at https://nucleartalent.github.io/MachineLearningECT/doc/pub/Day6/ipynb/Day6.ipynb.
- Tuesday From Neural Networks to Convolutional Neural Networks and how to analyze experiment (classification of events and real data) (Michelle Kuchera, MK). Jupyter-notebook of lecture at https://nucleartalent.github.io/MachineLearningECT/doc/pub/Day7/ipynb/Day7.ipynb.
- Wednesday Discussion of nuclear experiments and how to analyze data, presentation of simulated data from Active-Target Time-Projection Chamber (AT-TPC) (Daniel Bazin). Slides of lectures (PDF) at https://nucleartalent.github.io/MachineLearningECT/doc/pub/Day8/pdf/Day8.pdf.
- Thursday Generative models (MK). Slides of lectures (PDF) at https://nucleartalent.github.io/MachineLearningECT/doc/pub/Day9/pdf/Day9.pdf. Jupyter-notebook at https://nucleartalent.github.io/MachineLearningECT/doc/pub/Day9/ipynb/Day9.ipynb.
- Friday Reinforcement Learning (RR). Slides of lectures (PDF) at https://nucleartalent.github.io/MachineLearningECT/doc/pub/Day10/pdf/Day10.pdf. Future directions in machine learning and summary of course.