At some point over the last decade you may have heard something about Numenta, a company dedicated to creating Artificial Intelligence software that works more like the human brain. Jeff Hawkins, founder of mobile computing company Palm, has dedicated his time and his fortune to developing a cohesive theory meant to explain the workings of the human brain and therefore provide a blueprint for a powerful and innovative new kind of AI software. While his work has been revolutionary in nature to say the least, it has been widely ignored by the tech industry, even as major companies like Google become increasingly involved in machine learning processes. Now thanks to tech giant IBM, this amazing software may soon see the light of day.

IBM has taken a special interest in Numenta because it functions more similarly to the human brain than other AI software. Such an interest, in fact, that they have established a research group of around 100 people to work specifically on Numenta’s learning algorithms. The special project group, known internally as the Cortical Learning Center, is led by veteran researcher Winfried Wilcke who has nothing but high praise for the project. At a recent technology conference he described Numenta as “being closer to biological reality than other machine learning software” and added that it can learn how to make sense of raw data more efficiently. In other words, it performs in a more human brain-like way than its rivals currently do. While most machine learning software requires expert training with sample data in order to work, the hope is that the algorithms being developed by Numenta will eschew the “deep learning” method, which trains multi-layered networks of artificial neurons to find patterns in data, and mimic biological learning methods more closely.

“Our goal is not to be biologically inspired; I want to re-create it exactly.” This is the driving force behind founder Jeff Hawkins’s last decade of research into the functions of the human brain’s neocortex. This isn’t to say that his algorithms don’t also depend on the same multi-layered networks as other AI technology, but that their aim is to recreate the behavior of repeating circuits in the roughly 100 neurons found in the neocortex. The key to true machine learning software, according to Hawkins, is in these repeating circuits. If he is able to successfully mimic them in software, it would be a real game-changer. “This is how you would really build a machine intelligence,” he says. As we speak, Numenta is hard at work teaching the software to control physical equipment to be used in future robots.

As with any emerging technology, Numenta’s work is not without critics. However, most of the criticism so far is from other AI companies that point out what the technology is unable to do as of yet. But considering that the whole goal of Numenta is to function completely differently than any other AI currently does, and more like the human brain does, I’d say they’re on the right track.