User:Eugene M. Izhikevich/Proposed/Invariance learning
This article will briefly cover: Concept of invariance in the context of sensory processing. Selectivity versus invariance. Types of invariance (low level: shift, size, rotation, intermediate: e.g. cue-invariance, coordinate system, high-level: object-level, concept-level) Various solutions to building invariant representations (invariant features, pooling, routing). Supervised and "self-supervised" (e.g. cross-modal) approach. Unsupervised approach, using heuristics to learn invariance (anti-Hebbian, persistence, slow features, trace learning). Use in models of object recognition. The generative perspective.