Deep Learning with CNTK and F#

Explore deep learning using Microsoft's Cognitive Toolkit (CNTK) and F# in this 55-minute conference talk. Delve into the powerful world of neural networks and their applications in computer vision, speech recognition, and automated language translation. Learn the fundamental concepts behind deep learning, understand its inner workings, and discover how F# simplifies the process of training and implementing models within .NET. Through practical demonstrations, gain insights into learning models from data and integrating them into your .NET projects. Examine topics such as activation gates, reverse engineering, learning by example, robustness, circuit learning, model composition, tensors, and sequential models. Witness a hands-on demo using the MNIST dataset and compare CNTK with other frameworks like TensorFlow and PyTorch.

Intro
Context
Activation Gate
Can we reverse engineer this?
Learning by example
Robustness
Learning a Circuit / Board
Composing more complex circuits
Model: Composition of Layers
What is CNTK?
What does CNTK do?
Core components
What the hell is a Tensor?
Tensors
Functions
Sequential Models
Demo: MNIST dataset
MNIST input
Conclusion
What about TensorFlow? PyTorch?
Thank you :