The accessibility of Machine Learning techniques and the increasing amount of available data enable effective data-driven solutions for the water domain. Join this course for a practical understanding of Machine Learning and take the first step towards developing your own data-driven model.

WHO SHOULD ATTEND? 

If you are curious about Machine Learning (ML) and have a basic grasp of common Python libraries, this course is designed for you. It will provide a structured overview of the concepts and a guided hands-on experience.

This is a blended course, so you have access to all the content at once and you can decide how much time to invest in the course. Each module contains a deck of slides accompanied by a pre-recorded presentation from the instructors. You will also find guided Jupiter notebooks for putting in practice the newly learnt concepts. Most modules end with a quiz for self-assessment of the learning progress. You will learn online, through video lectures, handouts, Jupyter notebooks and quizzes. Additionally, you get 1 online, one-to-one meeting with your own DHI ML tutor. In this way, you have maximum flexibility in managing the course workload, but you are still able to interact directly with a specialist. Approximately two hour work per module is expected to get the maximum out of the course. You are expected to complete the course within 3 months to earn your course certificate and be allocated a tutoring session.

CONTENT

In the first part of the course (module 1 to 3) you will be introduced to the foundation blocks of ML, while the second part of the course (module 4 to 6) covers more advanced topics closer to practical applications.

1. Machine Learning mindset and Data Analysis
   • Definitions and examples of machine learning models
   • Common types of ML models
   • Problem framing
2. Classification models
   • Training, cross validation and evaluation
   • Decision trees and random forests
   • Underfitting and overfitting
3. Regression models
   • Linear and polynomial regression models
   • Regularization techniques
   • Bias/variance tradeoff
4. Neural networks
   • Building blocks and multi-layer perceptron
   • Training, gradient descent and learning curves
5. Working with time series
   • Time series pre-processing
   • Forecasting with Neural Networks
6. Working with images
   • Images as tables and clustering
   • Convolutional neural networks and transfer learning

BENEFITS

Upon completing the course you will be able to: 

• Navigate through the ML concepts and types of models
• Frame, build, test and tune your first ML model
• Use the core functions of scikit-learn and keras python library

REQUIREMENTS

• This is an online course, therefore, a reliable internet connection of sufficient bandwidth is critical.
• Familiarity with common Python libraries (numpy, pandas, matplotlib) is required.
• Expect to allocate between 4 and 12 hours to go through the course content

LANGUAGE

Lectures and training materials are in English. 

Don’t have an account? Follow these instructions to set one up today!