Google powers up AI, machine learning accelerator for healthcare

Launchpad Studio’s first class includes startups working to automate scribing with Google Glass, apply machine learning to help people who can no longer walk, pinpoint sepsis earlier and build a platform for wearables.
By Tom Sullivan
12:10 PM

With the mandate of fostering an ecosystem of applied machine learning startups, Google on Wednesday revealed the first four companies to join its Launchpad Studio and said this initial track is aimed squarely at healthcare and biotech.

It’s no secret that Google and rivals Amazon, Apple, IBM and Microsoft are eyeing the $2.7 trillion healthcare market as fertile ground for technological disruption -- though it appears Google is the first of the titans to formally establish a program for working with startups specific to the industry.  

“Launchpad Studio is the accelerator engine of Google,” said Malika Cantor, Program Manager for Launchpad Studio. “We’re focused on machine learning startups, we look at technology problems and want to bridge the gap between healthcare and the frontier tech industry because there’s a lot of promise but general skepticism about the role AI and machine learning will play.” 

[Also: Convincing C-suite to invest in AI: A new mode of ROI]

The first four startups, Augmedix, BrainQ, Byteflies and Cytovale, get what Google deftly described as “equity-free support,” and access to Google mentors, community engagement as well as datasets and testing environments for prototyping, as examples.

Augmedix is working to minimize the time doctors spend on a computer during patient visits by leveraging Google Glass to automate scribing and collect audio, video and written notes then use natural language processing to help clinicians make sense of that information.

[Also: New partnerships are fast-tracking AI tech]

Cantor described BrainQ as a research project concentrating on taking advances in neural networks and applying machine learning to signal processing to develop customized treatment protocols for people who cannot walk anymore, whether because of a stroke, spinal or brain injuries.

Byteflies is building a plug-n-play platform for wearables that other software developers could use to harness machine learning to transform raw data into actionable insights.

And Cytovale’s computer vision technology looks at how a cell transforms when a patient gets infected with sepsis to apply technologies that can detect sepsis earlier. 

[Also: Watson exec lays out 3 principles of AI: Purpose, transparency, skill]

Byteflies CEO Hans Daneels said that taking part in the Launchpad Studio program will give his company access to experts in the field of artificial intelligence and machine learning, as well people inside Google and in the industry who know how to make platforms scalable and understand the business aspects of doing so.

“We see a huge amount of potential applications so we want to make this scalable,” Daneels said. “It’s our goal to allow other companies to build wearable solutions on our platform without the need to reinvent everything.”

Cantor said the selection process involved Google engineers, product people, experts from DeepMind and Verily and others to ensure the due diligence includes determining that they are, in fact, addressing a valid problem.

“We’re going to be working with them for about 6 months,” Cantor added.  “We’ll touch base once a month, with them coming to meet with Google experts and we want the companies to spend time with their products to iterate because re-training a model takes a lot of time.”

After the initial six months, the four companies will operate similarly to a portfolio company in a VC firm, Cantor said.

“We’re hoping to inform what specialized tools we need to build for the healthcare industry, which APIs they need to go and build amazing things,” Cantor explained. “A lot of what we’re trying to do is support companies.”

Whereas the healthcare and biotech track is Launchpad’s first, Google intends to establish other industry tracks going forward, Cantor said.

Twitter: SullyHIT
Email the writer:

Want to get more stories like this one? Get daily news updates from Healthcare IT News.
Your subscription has been saved.
Something went wrong. Please try again.