Microsoft has announced the acquisition of machine learning specialist, Bonsai.
The Berkeley, California-based startup is developing a deep reinforcement learning platform that abstracts the low-level mechanics of the machine learning process, so that subject matter experts, regardless of their AI expertise or aptitude, can train autonomous systems to carry out different tasks.
In short, machine teaching.
Bonsai is partly funded by Microsoft’s M12 startup venture fund. The value of the acquisition has not been disclosed.
Simulating for success
Bonsai’s machine training regime takes place inside a simulated environment, and the system is targeted at organisations that are using industrial control systems in contexts such as robotics, energy, HVAC, and manufacturing, along with other autonomous applications.
For example, using Bonsai’s AI platform and machine teaching system, subject matter experts from Siemens, with no AI expertise, trained an AI model to auto-calibrate a computer numerical control machine 30 times faster than the traditional approach: a significant milestone in industrial AI.
Bonsai’s platform includes automated model generation and management, and what Microsoft calls “a host of APIs and SDKs for simulator integration, as well as pre-built support for leading simulations”.
While envisioned in science fiction for a long time, truly intelligent autonomous systems are still elusive and remain “a holy grail”, said Microsoft of the acquisition.
“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.
“To achieve this inflection point in AI’s growth, traditional machine learning methodologies aren’t enough. 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 Microsoft toolchain
Microsoft said that the new “toolchain” brought about by the acquisition 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”.
Gurdeep Pall, Microsoft’s corporate VP of Business AI said of the deal, “What I find exciting is that Bonsai has achieved some remarkable breakthroughs with their approach that will have a profound impact on AI development.”
According to Pall, Bonsai established a new reinforcement learning benchmark for programming industrial control systems last autumn. The platform successfully trained a simulated robotic arm to grasp and stack blocks on top of one another by breaking down the task into simpler sub-concepts.
“Their novel technique performed 45 times faster than a comparable approach from Google’s DeepMind,” he said. “Then, earlier this year, they extended deep reinforcement learning’s capabilities beyond traditional game play, where it’s often demonstrated, to real-world applications.”
Internet of Business says
At the heart of this latest acquisition is Microsoft’s oft-stated desire to make AI more accessible and useful for everyone. As Pall said, “we have to remove the barriers to development, empowering every developer, regardless of machine learning expertise, to be an AI developer”.
The deal comes in the wake of Microsoft’s ongoing acquisition of developer community and platform GitHub, and its purchase of deep learning specialist Maluuba.