g2net - 2nd Training School - Machine Learning and Signal processing for Time Series Analysis

Teaching material for 'g2net - 2nd Training School - Machine Learning and Signal processing for Time Series Analysis' held at the University of Malta, Valletta Campus, Malta, 9-13 March 2020.

---

TimeTable

Module Code	Topic
SP	Signal Processing
ML	Machine Learning
DL	Deep Learning
GRAV	Gravitational Data Talk
GEO	Geophysical Data Talk
HACK	ML Hackaton

Lectures

• SP - Signal Processing - Eftim Zdravevski

  • SP1 - Preparing Environment
  • SP2 - Introduction and Visualization
  • SP3 - Preparing GW Data

• ML - Machine Learning - Christopher Zerafa, Luigia Petre

  • CZ - ML1 - Introduction to Python & Machine Learning
  • CZ - ML2_3 - Classification vs Regression and Metrics
  • CZ - ML4 - Model Selection and Feature Engineering
  • LP - ML5_6 - Trees And Ensembles
  • LP - ML7 - Naïve Bayes
  • LP - ML8 - Support Vector Machines

• DL - Deep Learning - Luca Antiga

  • Lectures
    • DL1_2 - Neural networks and back-propagation & PyTorch basics
    • DL3_4 - Deep learning in PyTorch & Deep learning for g2net data
    • Lecture Recording - DL1
  • Notebooks
    • 1.1-PyTorch-Basics
    • 1.2-Linear-Regression
    • 1.3-Multi-Class-Logisitc-Regression
    • 2.1-Convolutional-Neural-Networks
    • 2.2-Pretrained-ResNet-Imagenet
    • 2.3-TransferLearning-MNIST
    • 2.4-Finetuning-Hymenoptera
    • 2.5-CharRNN

• GRAV - Gravitational Data Talk - Michal Bejger, Jade Powell

  • MB - GRAV1_2 - Introduction to Gravitational Waves
  • JP - GRAV3_4 - Gravitational Wave Data Analysis with Machine Learning

• GEO - Geophysical Data Talk - Tomek Bulik, Gregory Beroza

  • TB - GEO1_2
    • Slides
    • Recording
  • GB - GEO3 - Deep Learning Applied to Earthquake Signals
    • Slides
    • Recording

• HACK - Machine Learning Hackaton - Christopher Zerafa, Filip Morawski, Alberto Iess

  • Platform: g2net - Hack
  • Notebooks: g2net_malta_hackaton
  • Data: datasets
  • Deadline: Friday 13/03/2020 13:00 GMT+1
  • Baseline Model: notebook
  • Recording
  • Leaderboard:

  Rank	Participant Name	Notebook	Score (balanced_accuracy)
  1	Ondřej Zelenka	script_1	81%
  2	Filip Morawski	notebook_2	78%
  3	Christopher Zerafa	notebook_3	76%

Resources

• Deep Learning with PyTorch

Photos

You can find photo album with pictures here.

---