Hermann von Helmholtz

tags
People
source
https://plato.stanford.edu/entries/hermann-helmholtz/

Perception

Sign Theory

The mind makes a series of “unconscious inferences”, or mental adjustments, to construct a “picture” of the experience. Reasonably, we can lay this as a significant step towards representationalism. A key aspect of sign theory is that the rising learned signals are not exact copies of their initial sensations, but instead a sign of these sensations (or representation of these sensations).

Reinforcement Learning: An Introduction

We learn by interacting with the world around us. This has been thought about since the time of Hermann von Helmholtz, with his unconscious inferences.

Hierarchical Predictive Coding

Unconscious Inference: Hermann von Helmholtz

Sign Theory

Sign theory is a theory of perception developed by Hermann von Helmholtz of how the brain processes and propagates information (i.e. a theory of perception). The main contribution of the theory is that stimuli present themselves as signals and those signals are modified as they propagate forward. The mind makes a series of “unconscious inferences”, or mental adjustments, to construct a “picture” of the experience. Reasonably, we can lay this as a significant step towards representationalism. A key aspect of sign theory is that the rising learned signals are not exact copies of their initial sensations, but instead a sign of these sensations (or representation of these sensations).

clark2013: Whatever next? Predictive brains, situated agents, and the future of cognitive science

Work in connectionist networks was also inspired by Hermann von Helmholtz (specifically Hinton, McClelland, Rumelhart…). Specifically, the Helmholtz machin sought to learn new representations in a multilevel system without requiring the provision of copious pre-classified samples (i.e. a generative model).

These hierarchical models follow from Hermann von Helmholtz (1860) in depicting perception as a process of probabilistic, knowledge-driven inference.