The neuronal empiricism hypothesis: Cerebral cortex is an unsupervised Knowledge-seeking neural network

摘要

Recent studies demonstrate clearly that larhe-, medium-, and small-scale characteristics of the primary visual cortex re-emerge in auditory cortex when visual afferent information is redirected to auditory cortex. The neuronal empiricism hypothesis offers a natural explanation of this remarkable phenomenon, proposing that all cortex utilizes a knowledge-seeking unsuper- vised learning algorithm and that the specific characteristics exhibited by cortical cells are determined by the interaction of that algorithm with the statistical structure of real-world sensory input.