Scientists make a pocket-sized AI brain with help from monkey neurons : NPR

Source Domain: www.npr.org

Researchers successfully developed an efficient AI model that mimics a part of the human visual system for better understanding of brain functions.
Originating from training data from macaque monkeys, an AI model initially composed of 60 million variables was compressed into one with just 10,000 variables.
The compact AI mimics functionalities of human V4 neurons, which encode colors, textures, and curves, providing insights into the capabilities of human brains.
The study highlights how human brains operate efficiently despite using less power, raising questions on the complexity of current AI systems.
Potential implications include the development of simpler yet more powerful AI systems that could perform tasks like self-driving cars with less computational power.
Understanding the brain’s functions might be essential to developing AI models that work as flexibly as a human brain, despite smaller sizes.

You may have missed