Churchland & Sejnowski - Neural computation

Traditional reliance on intuition and logic for understanding for representation. Philosophers: Emphasis on a priori reasoning, dismissing empirical investigation (experiences) The naturalistic approach: incorporating psychological and neurobiological data.

Dualist perspective: Mind is separate from physical brain. “The ghost in the machine” is a materialistic critique of the concept of the mind existing alongside and separate from the body. The mind is not some spooky-unknown matter, but material and a part of the brain.

Rapid completion of cognitive tasks within neural limitations.

It functions as a highly interconnected parallel system.

It is modeled through complex nonlinear dynamical systems.

It challenges David Marr’s Dream by proposing diverse task descriptions and algorithms. Recognition of the brain’s structural complexity across multiple levels is emphasized. (An interdisciplinary approach is needed to understand the brain)

Structure of neuronal processes likely differs from overt language structures. Modeling information processing through complex nonlinear dynamical systems.

Learning tasks without symbols or rules, but with a self-learning neural network. The organization of the nodes emerges through experience and learning.

Shape-from-shading network mimics the visual cortex’s receptive fields. Inference of projective fields in real neural networks.