You will learn how deep neural networks work and how they are optimised. During our hands-on sessions, you will get the chance to work on our high-performance systems and train neural networks to solve an image classification problem. We cover different neural network architectures: from a basic fully connected network, to a convolutional neural network and variational auto-encoders (time permitting).

The language of instruction is English.

Sign up

What will you learn in this course?

• Understand how a neuron and neural network work
• Understand how a neural network is trained
• Investigate the effect of hyperparameters on the performance of neural networks
• Working with a high-level machine learning API (Keras)

For whom.

Anyone who is interested in deep learning, but has no (or little) current experience & knowledge.

Prerequisites

• Python
• Basic linear algebra
• Basic statistics

Topics

• Neural network: basic principles
  • How does a neuron work?
  • Fully connected networks
  • Training/optimising a neural network
• Convolutional neural networks (CNNs)
• (Variational) Auto-encoders

This training is a collaboration between SURF and the UvA, sponsored by EuroCC.

Sign in