Summary of Applied AI Workshop 23.10.2019

Workshop team

AI Hub Tampere's October workshop explored the many possibilities and frameworks around Applied AI. The presentations covered simulations, demos, implementation ideas, comparisons, hardware suggestions and other useful theory and practical tips.

AI Hub Tampere workshop team (from left to right): Mikko Pohja and Lauri Anttila (Futurice), Ritva Savonsaari, Andrei Ahonen, Reza Ghabcheloo, Timo Hämäläinen, Henri Lunnikivi and Kai Jylkkä. The workshop contents in a nutshell include:

  • Professor Timo Hämäläinen: opening the morning session with an introduction to the upcoming themes
  • Kai Jylkkä: some hardware recommendations for different machine learning use cases, their price range, and other potential. Jylkkä’s demo featured vehicle detection.
  • Henri Lunnikivi: an introduction to AirSim, what it is, and how or why it could be used. This free open source software is available on GitHub. It is possible to run autonomous vehicle tests on this Unreal Engine simulation environment. His demos visualized manually collected data and imagery applied to a neural network that drove a car in a simulated Hawaii environment.
  • Lunnikivi also explained how RUST can accelerate Python, which can sometimes run a bit slowly
  • Mikko Pohja (Futurice): a presentation on data science, data augmentation and neural networks project. The example case involved markku.ai, an AI-driven remote control car.  The operating code is given out as open source resource. The car itself can be bought or 3d printed.
  • Andrei Ahonen: robotics, ROS (Robot Operating Systems), DDS (Distributed Data Service), IoT, and how to solve connectivity issues.
  • Associate Professor Reza Ghabcheloo: closing the afternoon session with robot learning, trajectory optimization, (robot/computer) vision based control and Autonomous Mobile Machines

AI Hub Tampere would like to thank all our presenters and ~30 workshop participants. Please find more updates and photos about the workshop on Twitter @ai_hub_tampere