Bonsai is working on a reinforcement learning platform and has been funded in part by Microsoft’s startup venture fund, M12. This platform is targeted for companies utilizing industrial control systems within the framework of HVAC systems, energy allocation, robotics, and manufacturing.
The platform will allow experts of various subject matters to train autonomous systems, regardless of their level of Artificial Intelligence (AI) aptitude or expertise, to carry out specific tasks. This development dovetails with Microsoft’s ambitions to increase AI accessibility and user-friendliness.
Science fiction has incorporated intelligent autonomous systems in stories and movies for many years, but in reality they have remained an elusive, much sought after prize. As Microsoft explained:
“The reality today is that training autonomous systems that function amidst the many unforeseen situations in the real world is very hard and requires deep expertise in AI — essentially making it unscalable.”
As an example of the machine teaching system and AI platform Bonsai has developed, Siemens personal with expertise in AI were able to teach auto-calibration to an AI model. This training was for a computer numerical control machine and, compared to the traditional method, the AI model was trained 30 times faster.