Scientists of the Center for Systems Biology Dresden (CSBD), the Max Planck Institute of Molecular Cell Biology and Genetics (MPI-CBG), and the Helmholtz-Zentrum Dresden-Rossendorf (HZDR) – all part of the DRESDEN-concept research alliance – are organizing an exceptional workshop from 21st to 25th August 2017: With methods that are used by Google, Facebook, and other internet giants, you can also analyse scientific data. Even Google engineers will visit the workshop to share their knowledge.
When it comes to products offered online, internet groups count more and more on the forecasting power of deep neuronal networks. Those networks are mathematical abstract models trying to interpret events around them, much like the neuronal networks of our brains. Even though these methods have been known already for decades, the real breakthrough emerged only in recent years because of modern graphic cards. With the incredible computing power of these cards, nowadays you can train networks of unthinkable size in only a few hours. Such trained networks are used, for example, in driverless cars or to create product proposals on Amazon.
The goal of the one-week „Deep Learning Bootcamp“ at the CSBD is to introduce the use of deep neuronal networks for science. In that way, it is possible to substantially improve the quality of microscopic pictures, create a more realistic virtual reality, or detect cells or cell structures in pictures – to name a few possible applications.
One highlight of the bootcamp is a mini workshop by engineers and researchers of Google on 24th August 2017. Google will demonstrate how to use the tools of „Google Cloud“ to train neuronal networks. This mini workshop is open to all interested people from all disciplines or professions.
The idea and the realization of this workshop was a result of a cooperation within the interdisciplinary GPU Center of Excellence under the patronage of the Technische Universität Dresden. The organizers hope to create a multiplier effect within the Saxon science community and to further develop Dresden as a knowledge platform for neuronal networks, graphic card based calculations, and artificial intelligence.
Sign up for the mini-workshop on August 24th with Google: https://indico.mpi-cbg.de/event/68/overview