The 'information bottleneck' theory has been suggested as a possible explanation for how human brains learn.
Notes on verification
Multiple sources confirm the information bottleneck theory is discussed as a framework for understanding learning in both deep neural networks and, by extension, human brains, but it remains a hypothesis rather than settled consensus in neuroscience. [tier=gold indep_score=0.867 clusters=3 claim_tier=notable]
Sources
- New Theory Cracks Open the Black Box of Deep Neural Networks (seed:technology_and_ai)
- https://www.quantamagazine.org/new-theory-cracks-open-the-black-box-of-deep-learning-20170921/ (corroboration)
- https://elifesciences.org/reviewed-preprints/89369 (corroboration)
- https://www.techopedia.com/why-is-the-information-bottleneck-an-important-theory-in-deep-learning/7/33056 (corroboration)