Course Learning Machine Learning
Learning Machine Learning 2019 (LML19) is a 9-day summer course organized by the Departments of Electrical and Electronic Engineering and Biomedical Engineering at Universidad de los Andes with the support of Universidad del Rosario and the Center for Astrophysics at Harvard University. LML19 will bring together beginners and experts in Machine Learning (ML) in a multilevel school that will cover basic concepts in this area with the opportunity to develop real-world projects.
Note: There will no class on Saturday, June 1
Anyone who is interested in learning current concepts of Machine Learning, preferably with basic skills in mathematics and programming.
Basic skills in mathematics and programming.
This course will start by equipping the students with basic tools in python, statistics, and optimization. Then, lectures and laboratories on current theory and techniques in ML will be offered, followed by a day where students will have the opportunity to test their learned abilities in ML and to win a prize in a competition. During the last two days of the course, students will take part in multidisciplinary workshops along with top researchers in ML.
Lectures, workshops, challenges, computer laboratories.
This course will be taught in English. Students are expected to have a good level of communication in English (reading, listening, and speaking).
May 30th and 31st
Basics in Optimization, Probability, and Python. These sessions will be recorded, and the videos will be available for those participants that will not be able to attend them.
Introductions & Course Goals
-(Lecture 1) Computational and Inferential Thinking (June 4th )
- Machine learning & statistics introduction
- Problem definition & Data
- Exploration: visualization & data preparation
- Featurization and Pipelining
-(Lecture 2) Supervised Learning (I) (June 4th )
- Regression: logistic regression, kNN, Gaussian Processes
- Classification: Random Forest & LightGBM
-(Lecture 3) Unsupervised Learning
- Clustering approaches
- Anomaly detections with random forests
-(Lecture 4) Neural Networks
- Introductory algorithms and frameworks
- Fully connected networks for regression
-(Lecture 5) Deep Convolutional Neural Networks
- Imaging classification
- Time-series classification
- Temporal convolution NN (TCNs)
(Lecture 6) Generative and Compressive Modelling
- Auto-encoders for (semi- or unsupervised) learning
- Gurrogate emulation
-(Lecture 7) Semi-supervised & Reinforcement Learning
-(Lecture 8) ML In the Real World
- Business considerations
- Deployability, Scaling, and Maintainability
- Bias, Reproducibility, GDPR, and Ethics in ML
June 10th and 11th