Founded in 2014 by former Microsoft engineers, Bonsai helps enterprises add machine learning and AI capabilities to their existing operations. Its AI platform is designed to enable subject matter experts – even those with no AI background – to program their expertise directly into an AI model and teach it how to solve real-world business problems.
Microsoft says the Bonsai acquisition is a move to “build ‘brains’ for autonomous systems.” Training truly autonomous systems that can function amidst the many unforeseen situations in the real world, says the company, is difficult and requires deep expertise in AI — essentially making it unscalable.
“To achieve this inflection point in AI’s growth, traditional machine learning methodologies aren’t enough,” says Gurdeep Pall, Microsoft Corporate Vice President, Business AI. “Bringing intelligence to autonomous systems at scale will require a unique combination of the new practice of machine teaching, advances in deep reinforcement learning, and leveraging simulation for training.”
The Bonsai acquisition, says the company, is another major step forward in the tech giant’s vision to make it easier for developers and subject matter experts to build the “brains” – machine learning models – for autonomous systems of all kinds. Bonsai’s general-purpose, deep reinforcement learning platform – which includes unique machine-teaching innovations, automated model generation and management, APIs and SDKs for simulator integration, as well as pre-built support for leading simulations – is especially suited for enterprises leveraging industrial control systems such as robotics, energy, HVAC, manufacturing, and autonomous systems in general.
Bonsai’s approach has achieved several breakthroughs, says Microsoft, including successfully training a simulated robotic arm to grasp and stack blocks on top of one another by breaking down the task into simpler sub-concepts – a technique that performed 45 times faster than a comparable approach from Google’s DeepMind. In another example, subject matter experts from Siemens, with no AI expertise, used Bonsai’s AI Platform and machine teaching to train an AI model to autocalibrate a Computer Numerical Control (CNC) machine 30 times faster than the traditional approach – representing a huge milestone in industrial AI, and suggesting “staggering” implications across the broader sector, says Microsoft.
“Bonsai’s platform combined with rich simulation tools and reinforcement learning work in Microsoft Research becomes the simplest and richest AI toolchain for building any kind of autonomous system for control and calibration tasks,” says Pall. “This toolchain will compose with Azure Machine Learning running on the Azure Cloud with GPUs and Brainwave, and models built with it will be deployed and managed in Azure IoT, giving Microsoft an end-to-end solution for building, operating and enhancing “brains” for autonomous systems.”
Microsoft had previously invested in Bonsai last year via its corporate venture capital unit, M12. Terms of the acquisition deal weren’t disclosed.
AI wars heat up with new Microsoft research lab
Microsoft acquires deep learning startup
Self-taught Google AI achieves ‘superhuman’ chess proficiency in hours
Microsoft previews real-time AI platform for Azure
Microsoft buys conversational AI startup