Microsoft is looking to bolster its position in the race for artificial-intelligence breakthroughs with the announcement Wednesday of a deal to acquire Bonsai, a startup focused on creating methods for training autonomous systems.
Berkeley, Calif.-based Bonsai provides a faster way to build the "brains" necessary for robotics and other autonomous systems. The startup brings a "novel approach using machine teaching that abstracts the low-level mechanics of machine learning," wrote Gurdeep Pall, Microsoft corporate vice president for business AI, in a blog post.
Ultimately, Bonsai enables subject matter experts to "specify and train autonomous systems to accomplish tasks," regardless of their aptitude in AI, Pall wrote.
Bonsai's platform includes "unique machine-teaching innovations, automated model generation and management, a host of APIs and SDKs for simulator integration, as well as pre-built support for leading simulations all packaged in one end-to-end platform," Pall wrote.
Redmond, Wash.-based Microsoft did not disclose terms of the acquisition agreement.
Microsoft's venture capital arm, M12, already held a stake in Bonsai in connection with an investment announced in May 2017. Bonsai was founded in 2014.
Recent achievements from Bonsai have included a new benchmark for programming industrial control systems—which "performed 45 times faster than a comparable approach from Google’s DeepMind" at training a simulated robotic arm—as well as teaching Siemens subject matter experts to auto-calibrate a computer numerical control machine "30 times faster than the traditional approach," Pall said in the blog post.
"What I find exciting is that Bonsai has achieved some remarkable breakthroughs with their approach that will have a profound impact on AI development," Pall wrote.
Reed Wiedower, CTO of Washington, D.C.-based New Signature, said in an email that there are two big reasons for Microsoft partners to applaud the planned acquisition of Bonsai.
The first is that Bonsai's value proposition to partners is to make it quicker to build machine-learning systems, where "the principal is that you break a complex task into a set of easier tasks for training purposes," Wiedower said.
This enables partners with limited resources to "more quickly come up with a way to solve a problem than previously allowed," Wiedower said. "Right now, the entire industry is suffering from a severe shortage of experts in machine learning, and this will make that problem much less relevant."
The other big benefit to partners, he said, is that Bonsai has industry-specific solutions for verticals including energy, manufacturing and supply chain.
"As Microsoft itself starts to take on a more industry-specific leaning in the field sales organization, the ability to provide value to these specific verticals is a key value proposition of working within the Bonsai environment," Wiedower said. "In contrast to the value proposition for organizations who rapidly adopt Microsoft 365 and Dynamics 365, based on the strength of their broad capabilities, most of the value within Azure is specific to the industries in which it is deployed."
Having specific industry solutions "pre-baked" to solve certain types of problems, Wiedower said, "allows partners to enter an organization and provide value much more quickly."
The acquisition agreement for Bonsai follows notable comments about AI by Microsoft CEO Satya Nadella in April. Nadella contended at the time that the software giant is pulling away from competitors in artificial intelligence, with "advanced" AI for detecting security threats and a greater number of AI services "than any other cloud provider."
The move is also the latest in a recent string of acquisition deals for Microsoft that have ranged from the blockbuster agreement to acquire open source code repository GitHub, in a deal valued at $7.5 billion, in early June, to the acquisition of social learning startup Flipgrid this week.