Deep Racer

Deep racer is a way for everybody to get started with machine learning, sponsored by AWS. You do not need to have any prior machine learning knowledge and will playful get introduced in the methods. This is done by building a machine learning model that will autonomously drive a toy car. In the first two step the car will drive in a simulation (but you can get a real video and see how it drives). In the third and final step the top teams will hopefully meet in Stockholm and see real live cars compete on a real physical track.

More about AWS Deep racer here. Voidmark is hosting this deep racer event in three phases:

  • Phase 1: A classic warm up race.
    For your organization (incubator, company aso) there will be a 2-3 month period where you can train your models. The 5 top teams/models in the end of this period will continue to the race event in September, You can only spend 30 minutes to make a single model and just hope for the best, or you can improve it. In any case, this is a fun way to get started with machine learning.
    You have probably received an invitation link to this first part from Sascha.
    Resources:
  • Phase 2: A simulated synchronous race (September 2022).
    Top candidates are invited to submit their models (or improved ones) to this race event. Participants have the option of using interactive speed controls to give their model a competitive edge on their turn. This event will be LIVE publicly streamed. More information to come
  • Phase 3: In-person DeepRacer race on a physical track.
    This event will probably be in Stockholm. We will have two leagues, on open league for anybody that has participated in any race in phase 1 and 2 and a competition league for the top 10 models of phase 2, the synchronous race in September