GRADnet Machine learning and AI workshop

15 Jan 2020, 09:00 → 16 Jan 2020, 17:15 Europe/London

LT (Arts II)

A Bevan (Queen Mary), A Pourtsidou (Queen Mary), Marcella Bona (QMUL), U Blumenschein (Queen Mary)

This is a 2-day workshop that consists of a series of presentations and tutorials, to ensure delegates get the chance to learn about modern methods, and have the opportunity to get some hands on experience. This workshop will be of interests to PhD students keen to learn about, and start using modern data science and AI in the context of physics and astronomy.

Wednesday 15 January

10:30 → 11:00 Break

11:00 → 12:30 Introduction and Orientation Session

This session will lay down the housekeeping rules and provide some brief data wrangling issues.

11:00 Welcome and housekeeping

Speaker: Dr A Bevan (Queen Mary)

Slides Housekeeping.pdf

11:10 Why is data wrangling important?

Speaker: Dr Nick Barlow (Alan Turing Institute)

Slides DataWrangling_QMUL_150120.pdf

11:30 Data wrangling exercises

Speaker: Dr Nick Barlow (Alan Turing Institute)

Slides https://github.com/nbarlowATI/DataWranglingWorkshop https://github.com/wrattler/wrattler-examples

12:30 → 13:30 Lunch

13:30 → 15:00 Tutorial

13:30 Introduction to Python and TensorFlow (linear regression and neural network examples)

Speakers: Dr A Bevan (Queen Mary), JMM Davies (Queen Mary), Dr Marcella Bona (QMUL), TP Charman (Queen Mary)

Slides

• DecisionTrees.pdf
• https://github.com/adrianbevan/GRADnet-ML/tree/master/NN
• NNs.pdf
• NNs_Tutorial.pdf

15:00 → 15:30 Break

15:30 → 17:00 Plenary

15:30 Reinforcement learning

Speaker: Dr A Pourtsidou (Queen Mary)

Slides RL-lecture-AP.pdf

16:15 Sci ML team @ RAL

Speaker: Dr Jeyan Thiyagalingam (STFC)

Slides 2020_01_15_SciML_QMUL_GradNet_Jeyan.pdf

18:00 → 20:00 Dinner

Thursday 16 January

09:00 → 10:30 Tutorial

09:00 Linear regression, neural network model building, and decision trees (tutorial using SciKit Learn and Keras)

Speakers: Dr A Bevan (Queen Mary), JMM Davies (Queen Mary), Dr Marcella Bona (QMUL), TP Charman (Queen Mary)

Slides

• DecisionTrees.pdf
• gradnet_ml.pdf
• https://github.com/adrianbevan/GRADnet-ML/tree/master/Pulsars
• NNs.pdf

10:30 → 11:00 Break

11:00 → 12:30 Plenary

11:00 NNPDF - using neural networks to compute parton density functions

Speaker: Dr Maria Ubiali (DAMPT)

Slides NNPDF.pdf

11:45 ABBA - Automatically Building Better Algorithms.

Speaker: Dr John Woodward (QMUL)

Slides john_20200116.pdf john_20200116.pptx

12:30 → 13:30 Lunch

13:30 → 15:00 Tutorial

13:30 SQL tutorial

Speaker: Dr A Pourtsidou (Queen Mary)

Slides https://github.com/Alkistis/DSI/blob/master/SQL-tutorial-DSI.ipynb SQL.pdf

14:15 Data in finance

Speaker: Dr Thomas Babbedge (Gresham Investment Management LLC)

15:00 → 15:30 Break

15:30 → 17:15 Plenary

15:30 Challenges of Machine Learning in Physics

Speaker: Dr A Bevan (Queen Mary)

Slides 2020-GRADnet-ML-PhysicsChallenges.pdf

16:15 Machine Learning applied to neutrino physics

Speaker: Dr Sophie King (Kings College London)

Slides 20_01_16_ml_nu.pdf