Abstract

The goal of this workshop is to provide an insight into deep learning on the use case of detecting handwritten numbers. After an introduction to the basics of neural networks, a workflow for machine learning tasks is solved in groups. Therefore, Jupyter notebooks are provided to guide the course through ANN development with Tensorflow/Keras. Finally, the trained networks will be applied for the detection and recognition of handwritten text on images.

Prerequisite: Basic Python knowledge and, ideally a Laptop with an Anaconda 3 installation (Instructions can be found here in Ger+Engl). As a minimal requirement, a Laptop or Tablet (with keyboard) can be used to work in Google-Colab.
IMPROTATNT: For the Colab-Backup, a Google account is mandatory and the last slide of the Instruction should be completed before the event.