Please log in to watch this conference skillscast.
Artificial neural networks are not made only to play Go or distinguish tanks from cats on photos. Nowadays, in the world of BigData and TensorFlow, it's easy to forget what inspired them: neurons. In this talk you will take a closer look at them, both those made of jelly and their digital counterparts. You will discover how they differ and explore some of their non-trivial features. Maciej will try to prove that learning is not so crucial to intelligence as it is usually thought, whereas the ability to operate in real time is. Then you will go through a short introduction of the actor model and Akka, a Scala's toolkit emphasizing actor-based concurrency, which Maciej will subsequently use to demonstrate how to build a very small asynchronous neural network based on the features discussed before.
YOU MAY ALSO LIKE:
- LDNUG September 2017 - #ProgNET Special with Richard Campbell (in London on 12th September 2017)
- Progressive .NET 2017 (in London on 13th - 15th September 2017)
- London Unreal Engine Meetup (in London on 20th September 2017)
- Fast Track to F# with Tomas Petricek & Phil Trelford (in London on 16th - 17th October 2017)
Artificial Neural Networks in Akka - Beginner
Maciek is a Scala and Android developer at Wire in Berlin and an AI enthusiast on weekends. He is interested mostly in research on the border between artificial neural networks and neurobiology, and in using AI to control drones and non-player characters in computer games. He started coding in Scala in 2013 and never looked back. An advocate of healthy minimalism, clean code, and bicycles.