Everything .NET developer wanted to find out about deep learning, but was afraid to ask

RU / Day 2 / 10:30 / Track 1

Machine learning and neural networks are quite popular topics now, which allow making a howling success in various fields: from image processing to music synthesis. For some reasons, Python is often used for such tasks, but it's hated by .NET developers. The bitter truth is that .NET platform hasn't got convenient tools that allow you to solve the problems of neural network learning. Because of this all tasks of machine learning are usually given to data scientist and further separated in the form of distinct services. .NET developers prefer not to think about it.

In this talk, Dmitry will tell about main types of modern neural network and how .NET developers can use finished trained networks to solve problems. Also, how to teach these networks to the end. The main emphasis of the demonstration will be on image processing — we will learn to recognize and find objects on the image. We will use ML.NET, the audience must stay calm during the demonstration of Python code.