7th #MLDD - Learning on the fast lane

Hi Everyone,

we are happy to announce that the next MLDD “Learning on the fast lane” will be on May, 28th. This time, Oliver Lange will present a method which speeds up the learning phase (specifically, deep q-learning) of reinforcement learning. For those who are not familiar with the term reinforcement learning: This is a special technique in machine learning which lets artificial agents (could be robots, for example) learn in a similar fashion as humans do: By exploring the space of possible actions (like moving forward or standing still) and observing the effect this has on the agent itself (like falling) and its environment. The “AI“ then tries to learn to predict the outcome of an action before the action is taken. If you want to know how that looks in practice, go and watch an AI learn how to play PacMan: https://youtu.be/LbWXGw1sWD0. Notice how it makes silly moves and gets eaten by the ghosts? Yes, it has to learn the hard way…

We’ll meet at LOVOO’s fancy kitchen with free food and beverages, sponsored by LOVOO. Big thank you again :)

There will also be time for your 2 minute pitches after the talk: If you have an idea for a new project, are looking for a master’s thesis, or have the best startup idea ever but not the team or the customers (yet) – this is your opportunity to present it to an expert audience and get feedback right away, for free.

If you don’t feel like talking in front of a big group of people – just put your ideas on a poster and let us know in advance (via email) so we can organize a proper poster session.

Also, if you’d like to present yourself at one of our following meetups, just write a message on the group’s board on meetup.com or contact us directly.