Deep Learning in Practice


Details
Dear Meetup members,
I'm exited to announce we gonna have a practice-oriented Meetup next Wednesday.
Agenda:
18.00 Gathering and networking
18.30 A Pragmatic Introduction To Restricted Boltzmann Machines
Dr. Christian Leibig, Zeiss Vision Science Lab in Tübingen
In this talk, we will explore the fundamental ideas and mathematical intuitions behind Restricted Boltzmann Machines and how they can be used as effective machine learning units in both supervised and unsupervised learning problems. This talk will also cover examples of successful applications of Restricted Boltzmann machines and provide references to existing machine learning libraries that can speed up development and prototyping processes.
Christian Leibig received a PhD in Neural Information Processing from the IMPRS Tübingen and is currently working as a post doc at the Zeiss Vision Science Lab in Tübingen.
19.15 Break and networking
19.45 Implementation of Restricted Boltzmann Machines in Python
Peter Schneider and David Haber, Soma Analytics
This interactive coding session will take participants through the implementation of a Restricted Boltzmann Machine in Python. We will use standard Python packages such as numpy and scipy. Participants will have a chance to benchmark their implementations on a variety of datasets to compare performance. We will also discuss ways to integrate GPU acceleration (e.g., using Theano or PyCuda) and other tweaks to improve speed and analysis capabilities."It is highly recommended that participants bring their own laptops to this session and install the Anaconda Python distribution prior to the session: https://www.continuum.io/downloads (https://www.continuum.io/downloads.)
Peter is Co-founder at Soma Analytics. He studied quantitative economics at the University of Kiel and computer science at the University of Hagen.David is a Lead Software Engineer at Soma. He graduated with a degree in computer science from Imperial College London.
The meetup is going to be limited to 75 participants this time, so be quick!
Looking forward to see you there!
Thanks a lot in advance to our sponsor and host TNG, and to Peter and David for putting this together to a large extend.
Uwe

Deep Learning in Practice