Meta is looking to improve AI
The project would begin with the scanning and reverse engineering of brains from rats, crows pigs and chimps and finish on the human cortex. By 12/31/2025, the goal is to complete this project. Create teams that can run brain scans 24 hours a day, 365 days a year if necessary. Partner with every major neuroscience laboratory in the world.
What would happen if the deep learning algorithms were compared to the human brain’s activities? Researchers from Meta AI announced last week that they will be working with Neurospin (CEA), INRIA, and the neuroimaging center Neurospin to achieve this goal.
This collaboration will analyze the brain activity of humans and deep learning algorithms that have been trained to perform language or speech tasks, in response to the exact same written or spoken text. Theoretically, this could help decode how both human and artificial brains find meaning in language.
Researchers hope to discover structural and behavioral similarities between artificial networks and brain biology by comparing brain scans taken while people are actively reading, listening, or speaking with deep learning algorithms. This research may help explain why machines process language more quickly than humans.
Source:
https://gurtimes.com/science/meta-wants-to-improve-its-ai-by-studying-human-brains/
Leave a Reply