Samsung’s team designed to seek out innovative, forward-thinking technology today announced the launch of the Q Fund, an investment initiative meant to advance early-stage AI startups.
Formerly named Embodied Intelligence, Covariant is teaching robots to learn how to complete tasks without human control and was founded by UC Berkeley professor Pieter Abbeel, together with other OpenAI researchers.
The Q Fund will focus on backing startups using transformational techniques for robot control, human-machine interaction, simulated learning, and AI that automatically builds custom AI models, like Google’s AutoML.
The fund will also look for ways state-of-the-art AI or novel approaches can help reshape computer science.
Samsung Next principal investor Vincent Tang believes machine learning can be used to improve algorithms, computer systems, and data structures and help solve decades-old computer science problems.
“Everything we take for granted today or everything the field of computer science has tackled in the past 70 years, we think it’s ripe for replacement or augmentation with AI,” Tang said. “Historically, these have been solved with hand-crafted algorithms, so you had a bunch of really smart computer scientists figuring out the right algorithms to figure out how to index or sort something. The thesis here is you can replace all that with learned systems. You don’t need fine-grain, fine-tuned, hand-crafted algorithms to design these systems for you.”
Investments will focus on startups in the United States initially but may expand to Europe and other parts of the world where Samsung has offices, like Berlin or Tel Aviv.
The Q Fund will draw its capital from the original Samsung Next Fund. Previously known as the Samsung Global Innovation Center and established in 2012, Samsung Next is a $150 fund for investment and acquisition of startups working in areas like artificial intelligence, VR/AR, and internet of things (IoT). Initially focused on the U.S., the fund grew to incorporate European startups nearly a year ago.
No commitment was made by Samsung to invest a specific amount of money or make a specific number of investments with the Q Fund, but investments will run in the neighborhood of $1 million, Samsung Next director and principal investor Ajay Singh told VentureBeat in a phone interview.
Samsung Next has invested in 78 companies thus far, a company spokesperson told VentureBeat in an email.
The Q Fund will work with researchers from Carnegie Mellon University, Columbia University, MIT, and other schools.
In its bid to compete with Facebook, Google, and other companies with global AI ambitions, Samsung last month opened AI research centers in Cambridge, Moscow, and Toronto.
Samsung Next has been in touch with the AI Center but does not have any specific plans to work with Samsung’s AI centers, Singh said.
Samsung Research seeks to employ more than 1,000 advanced AI researchers by 2020 to explore ways to bring machine intelligence to Samsung products.