Computational Reproducibility, from theory to practice.
The Reproducibility Crisis has received much attention in the past decade and has even been proposed as the biggest challenge of our academic generation. While debate about the extent and implications of the crisis continues, few would argue that more transparent research, where the raw materials underlying findings are openly shared, would not improve reliability and is, therefore, a reasonable expectation of modern research.
We will look at how some of CRAN's principles and mechanisms have evolved over the past, and what they might turn into during the next decade. We provide both our personal views on challenges and opportunities, and try to answer some key questions.
Talk with your model! Towards the language for exploration and explanation of machine learning models.
It is getting easier to build complex predictive models. However, without proper verification, these models carry a hidden risk that they will eventually stop working and we won't even notice. To prevent this from happening we need good diagnostic tools to explore, maintain and debug models.
Deep Learning with Keras and TensorFlow
Dr. Shirin Elsinghorst (b. Glander)
This workshop begins with a short theoretical part to make everyone familiar with the basics of deep learning, including cross-entropy, loss, activation functions and how weights and biases are optimized with backpropagation and gradient descent. Then we will delve into how to build (deep) neural networks with Keras and TensorFlow in R.
Bernd Bischl & Michel Lang
The mlr3 ecosystem provides a one-stop solution for all machine learning (ML) needs, spanning preprocessing, model learning and evaluation, ensembles, visualization, and hyperparameter tuning.
Analyzing and visualising spatial and spatiotemporal data cubes
Edzer Pebesma & Martijn Tennekes
Data cubes are a modern way to denote array data, and we focus on array data where some of the dimensions refer to space and time. Examples are spatial raster data, multivariate time series for multiple locations, and time series of raster images such as satellite data or weather predictions. In this workshop we will show a variety of cases where such data arise, and demonstrate how they can be analysed and visualised with R.
Scan the barcode with Eventor to have the schedule always with you.
This website is licensed under a Creative Commons Attribution-ShareAlike 4.0 International LicenseImpressum
Template by Bootstrapious.
Ported to Hugo by DevCows