The Analytics track at IT OUG is very interesting this year. Along with some of the typical analytics talks, there are a large number of presentations on machine learning and related topics.

Machine Learning area has matured a lot over the past 12 months, and it seems everyone is talking about it. But the conversations now are very different. Gone are the days presentations on what is Machine Learning and the basics of it. Lots of people are working with it in their companies solving a wide range of problems, or have already learned the basics.

I’ll be giving two presentations in Milan and Rome, and the focus of these will be on how to use machine learning in your development languages/tools, in your day-to-day operational systems, etc., and I’ll be discussing some of the typical pitfalls and successful approaches for winning in these areas.

In Milan, my first presentation is called “Putting Machine Learning into Production: the Good, the Bad and the Ugly”. I’ve worked on many machine learning projects over the past 27 years, I’ve seen a lot of things happen, from having a pure machine learning or statistics focus, to it being used for Business Intelligence, to ML Ops and production deployment. This presentation will illustrate some of the good, the bad and the ugly scenarios I’ve encountered and give some recommendations on how to make it easy to productionalize your machine learning.

The second presentation in Milan is co-presented with Neil Chandler. Everyone is talking about Neural Networks. But there is lots of confusion about using Neural Networks, what they do, what you can do with them and how much processing power is needed. It isn’t just for identifying pictures of your Cat or Dog. In this presentation I will answer all of these questions before handing over to Neil who will bring you through how to use Oracle REST and how to call the neural network model I built in the first half of the presentation.

The conference then moves to Rome on the Friday, where I’ll be giving the “Putting Machine Learning into Production: the Good, the Bad and the Ugly” again (see above for a description). After lunch there will be four, ten minute presentations. This is great way to get a taste of different topics and helps the presenter to focus in on a specific topic and the solution to it. My topic for this ten minute presentation is “REST enabling in-Database Machine Learning models”. I’m going to attempt a slightly different approach to giving this presentation. It will be a mixture of slides and demo, but I won’t be using any presentation software or touching the keyboard for the demo! Fingers crossed it will work the way I want and if everything goes to plan I will be finished in 9 minutes 29 seconds. It will be a slightly Autonomous presentation. Will it work?