In 2006, two of Neurala’s co-founders were sitting in a coffee shop when one of them said something that made the other spit out his coffee: “Someday we’ll be able to run these AI models on a cell phone.” At the time, that sounded absurd. AI required server racks, not handsets.
“Someday we’ll be able to run these AI models on a cell phone.” – Boston, 2006
The founding team had spent years studying how biological brains learn — specifically how the complementary systems of the neocortex and hippocampus allow humans to acquire new knowledge quickly, continuously, and without forgetting. Their research, conducted for NASA, DARPA, and the Air Force Research Labs, pointed to a fundamental flaw in conventional deep learning.
The solution they developed — the Lifelong Deep Neural Network (L-DNN) — mimics the brain’s dual-system architecture in software. It learns from a handful of examples, in seconds, on modest hardware, and continues learning throughout its entire deployed life without catastrophic forgetting. That technology is now protected by 19+ patents and has been embedded in enterprise products used by customers around the world.
In 2013, Neurala emerged from stealth and joined the Techstars program to begin commercial deployment. Today, Neurala’s L-DNN technology is embedded inside products built by FLIR, Sony Semiconductor Solutions, IHI, and more — and we are actively working with integrators and OEM hardware companies to put L-DNN inside the next generation of intelligent products.
