A Model of Perceptual Change by Domain Integration
Year: 2001
Type: conference-paper
Abstract: A neural network model is presented that shows how the perception of stimuli can be changed due to their integration between different domains. The model develops representations based on the correlational structure between the stimuli of the domains. It shows how prototypicality, categorical perception, and interference effects between the domains can arise without the explicit labeling of categories. The model is extended to learn the sensori-motor integration between articulatory parameters and speech sounds and it is shown how it can, in accordance with the ideomotor principle, imitate sounds based on the developed mappings in a “babbling phase”.
Author: Gert Westermann
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2621593743