Artificial intelligence and machine learnings are topics that can provoke mixed reactions, with a good dose of fear that machines may firstly take over your jobs and, ultimately, the world. The world of machine learning is advancing rapidly and an announcement today in the field of unsupervised machine learning from Cortica, a company from Israel founded in 2007, is exciting, while perhaps stirring up a few feelings of disquiet.
Cortica has spent the past 10 years studying the behavior of the mammalian cortex and reverse engineering the process exercised by the human brain to think and to learn. The result is next-generation AI that mimics human cognitive function; enabling machines to think and learn on their own…just like you and I do.
The process, known as ‘Autonomous AI’, is a significant departure from the current, widely used ‘deep learning’ approach that now supports AI capability in state-of-the-art technology. It is the long-anticipated missing piece of the puzzle that brings true artificial intelligence to life; imparting machines with the capability to evolve beyond the parameters and the limits prescribed by human dictate through data input. Autonomous AI enables machines to think and learn like people.
Deep Learning, the current widely accepted paradigm for AI enablement, relies on the input of an enormous amount of annotated data to “teach” a machine to recognize a given set of possibilities and parameters. The machine carries out functions and completes tasks subject to those parameters. The more data the machine ingests, the greater its capabilities.
However, the same machine cannot reason or plan or understand on any level beyond the parameters of its data program. It is incapable of reading context or interpreting any variation of an object/idea/factor that it has not previously encountered. It cannot observe, infer, and comprehend simultaneously as people do. Neither can it learn in real-time, all at once, in the manner of a sentient being. AI based on Deep Learning will always be limited to the parameters prescribed by its creator. It is also dependent upon an enormous stockpile of data in an attempt to cover every detail of every possibility of any given subject.
Cortica takes a fundamentally different approach to AI, an approach that is unrivaled in the AI industry. By tapping into the live brain tissue of rats and observing and replicating the structures and functions of the cortical network, Cortica has uncovered a process called ‘Unsupervised Learning;’ a re-creation of the way in which people receive, process and react to information or stimuli. Autonomous AI is inspired by this biological understanding, and its re-creation enables machines to mimic human cognitive function outside the confines of the human brain.
Unsupervised Learning allows the system to comb through heavy amounts of data, organize it automatically based on commonalities and, in doing so, the machine begins to learn and comprehend on its own.
“The real-world applications that are enabled by Cortica technology are astounding,” said Igal Raichelgauz, Cortica CEO. “In the short term, Cortica will be the bridge that brings driverless (autonomous) cars to life. Our approach enables the car to identify the moving parts in its environment, make informed decisions, and even predict the changes in path of the elements in its surroundings. We enhance the machine’s environmental understanding beyond a simple data based description of past experiences. The technology is applicable across the board in not just the automotive setting, but in security, environmental and industrial settings as well. It will truly reinvent our world.”