Much more on IBM PowerAI - this time with live demos

This time it is much more about live demos and not so much PowerPoints. It is still all about IBM PowerAI and how to use the specific IBM tools to facilitate everything Deep Learning. Since we've got a lot of new Deep Learners since our last MeetUp, we will as well do a short recap on, what IBM PowerAI is, and what it can offer you. The seminar is targeted people with an interest in Deep Learning, whether that interest is fueled by business or technical reasons. All presentations will be in English.

Agenda

08:45-09:00 Registration

09:00-09:05 Welcome (Kim Gregers, Partner Technical Advocate, IBM)

09:05-09:30 Update on IBM PowerAI including recent use cases (Kim Gregers)

09:30-10:20 IBM PowerAI helps you training DL for images. How to move on from there (Chris Parsons, Machine Learning Engineer, IBM):
So you've built a model using PowerAI Vision. How do you go from there to integrating it into your application so your users can benefit? In this demo, we define, train and deploy an image classification model to an iOS app

10:20-10:35 Break with coffee

10:35-11:15 How to optimize your hyper parameters (Ari Juntunen, CTO, Elinar Oy):
How to use Deep Learning Impact to optimize your hyper parameters automatically. In this session, we cover a simple image detection use case and see how Deep Learning Impact can optimize the model automatically

11:15-11:40 PowerAI Vision - Eye Mouse (Sophie Mai Chau, Student Programmer, IBM)
Here we will use the Vision tool to build a convolutional neural network for eye-tracking. We will first detect face and eyes and then compute a gaze prediction. Then we can use this prediction to move a mouse pointer

11:40-11.55 Finishing off the day with sandwich'es