Research
Fernando Peruani in Nature Physics
on August 10, 2022
Research
Published on September 19, 2022– Updated on September 19, 2022
Digital computing through randomness and order in neural networks
Alexandre Pitti, researcher at ETIS laboratory, explains results published in the American journal "Proceedings of the National Academy of Sciences" (PNAS) on August 10, 2022 in the article "Digital computing through randomness and order in neural networks".
We hypothesize that the brain performs some kind of digital computing to overcome the unreliability of its neurons by exploiting random synaptic connections. Theoretical equations and computer experiments show that surprisingly, only a small number of neurons is enough for coding sequences formed from items taken within a repertoire of a billion different values.
Despite the learning limitation of neurons, which perform large errors, random connections help to generate high entropy, conveying maximum information simultaneously. This result led us to derive a general equation of the learning capabilities in neural networks, which has deep implications in neuroscience modeling and artificial intelligence toward developing a neural information theory.
Read the article on PNAS website
Despite the learning limitation of neurons, which perform large errors, random connections help to generate high entropy, conveying maximum information simultaneously. This result led us to derive a general equation of the learning capabilities in neural networks, which has deep implications in neuroscience modeling and artificial intelligence toward developing a neural information theory.
Read the article on PNAS website
