Our hands-on workshops are provided by our sponsors. Each workshop is open for a limited number of participants and you can register in advance.

Just send an email to and tell us which workshops you would like to attend.

Workshop 1:

Tuesday, June 13th, 9:45

Cloudera: powering possibilities in machine learning

Steffen Märkl, Systems Engineer, Cloudera

Machine learning is all about the data, but it's often out of reach for analytics teams working at scale. Cloudera customers such as can store, process and analyse 550 million events each day to help them improve gamers’ experiences and increase customer lifetime value.

Whether you are new to machine learning and advanced analytics, or you already take advantage of the possibilities, this interactive workshop will explore practical examples and give you some new ideas to take away. Discover how enterprise organisations can accelerate machine learning from exploration to production by empowering their data scientists with R, Python, Apache Spark and more in one secure, unified and collaborative platform.

Workshop 2:

Tuesday, June 13th, 10:45

F&F GmbH: TensorFlow step-by-step

Dr. Federica Fusco, Data Scientist, F&F GmbH

Nowadays deep learning is everywhere and along with it TensorFlow, Google’s deep learning library, which has widely spread since its first release at the end of 2015. Learning TensorFlow can be quite challenging mainly for two reasons. On the one hand the low-level of this tool introduces some obvious difficulties, on the other hand the available code examples are sometimes hard to understand for beginners, especially if they are related to a previous version of the software. Hence, in this presentation we will introduce TensorFlow from scratch. We will start from a couple of simple lines of code, and will arrive at the implementation of a full neural network. Every line of code will be described and live executed. Our purpose is to give a deep insight into TensorFlow, in order to make it accessible to everybody who is interested in learning this tool. A standard TensorFlow execution starts with the definition of the computational graph. This graph defines the placeholders of the input parameters, which have to be fed with data on execution, and the operations that one wants to execute. Once the definition is complete, a session, i.e. a runtime context for executing the graph, is launched. We will start the presentation from a very simple graph definition, in order to solve a multinomial logistic regression. This example will give the opportunity to understand the basics of graphs, sessions, and more generally TensorFlow itself. The way we implement the logistic regression, it can be seen as a very simple neural network (NN), consisting of one layer of neurons, a softmax activation function and a categorical crossentropy cost function. Therefore, to conclude our presentation we will present a NN, which will be composed of a few convolutional layers.
In particular, we will focus on the architecture of the network, the input and the output of every layer, and we will use it to perform a simple classification task.