Quorum Robotics Releases the Engram Artificial Intelligence Engine
System capable of learning and making decisions in realtime
February 10, 2015
CHICAGO – Quorum Robotics today announced the availability of their Engram AI Engine, a next-generation Artificial Intelligence system that is capable of learning and making decisions in realtime, using a fraction of the resources required by traditional AI methods.
Engram was designed as a learning engine for distributed robotics systems – robots that have independent controllers but that must operate in concert with one another. In these situations, the number of possible interactions between robots is immeasurable, so Quorum Robotics focused on adaptability – enabling the AI to learn and make decisions in real-time, to adapt to new or unknown situations.
“We quickly learned that the Engram AI engine was able to learn much more than robotics control, and we tested it in a variety of environments to see how it stood up compared to other methods,” said Quorum Robotics CEO Noah Schwartz.
One such test involved the classic video game, Pong. While existing Deep Learning methods required several hundred thousand games to achieve super-human performance, the Engram AI Engine was able to master the game after only playing a few points. “The Pong results really opened our eyes to the power and versatility of the engine,” said Schwartz.
Like other AI methods such as Deep Learning and reinforcement learning, the algorithms that drive Engram are inspired by the brain. Engram is unique, however, because it is derived from Schwartz’s own 12-year research program examining how the brain learns and organizes in response to experience.
“Much of the brain-inspired AI that we see is based on high-level concepts about the brain,” added Schwartz. “Our algorithms are different because they mimic what is actually happening on a molecular level when the brain is learning and adapting. The rest of the AI structure organizes automatically in response to whatever input we feed into it.”
The efficiency of the Engram AI Engine allows it to run on stand-alone devices, such as smartphones or a household IoT devices like smart thermostats. “We’re just beginning to explore the ways that the engine can be used, and we’re testing the market to see how the Engine can provide value, not only for automated hardware systems, but for software-based AI applications as well,” said Schwartz.