01.03.2021

Artificial Intelligence Neuron Discovery

A neuron previously only found in the human brain has been discovered in artificial intelligence for the first time.

Researchers working at the OpenAI startup, which was founded by none other than billionaire Elon Musk, have discovered the presence of a multinodal neuron in an artificially intelligent network. The multinodal neuron, which was discovered in 2005 and was thought to only be present in the human brain is solely responsible for identifying a common theme from a variety of sensory information.

In the case of the human brain, a single neuron is devoted to each and every person you meet so that when you see them, an image of them or see their name, the multinodal neuron will make a connection in the brain.

It was discovered by engineers at the startup inside a neural network called CLIP. CLIP trains itself on complex datasets in order to recognise distortions and changes in the lighting and pose of objects or people. It can go even further to recognise those objects or people in abstractions – such as sketches, cartoons or statues.

“We’ve discovered neurons in CLIP that respond to the same concept whether presented literally, symbolically, or conceptually” OpenAI explained in a blog post. “Our discovery of multinodal neurons in CLIP gives us a clue as to what may be a common mechanism of both synthetic and natural vision systems: abstraction.”  

“Using the tools of interpretability, we give an unprecedented look into the rich visual concepts that exist within the weights of CLIP,” the OpenAI researchers wrote. “Within CLIP, we discover high-level concepts that span a large subset of the human visual lexicon—geographical regions, facial expressions, religious iconography, famous people and more. By probing what each neuron affects downstream, we can get a glimpse into how CLIP performs its classification.”

Neural networks, like CLIP, are being increasingly used in pushing the boundaries of technology with the invention of systems such as facial recognition, self-driving cars and digital assistants like Amazon’s Alexa.

Posted by: Rock Consulting