[Next] [Previous] [Up] [Top] [Contents]
7. A Bottom-up Investigation of Internal Context in an Artificial Agent
7.1. Overview
The idea is to place an artificial agent in a environment that is well beyond its capacity to model but one which exhibits some regularities. The agent is designed to learn about its environment using feedback from its attempts to predict outcomes in that environment. It learns using a method that allows the development of pattern-recognition, inference and context-like constructs. I then study the structure of the knowledge that the agent learns in order to identify whether context-like constructs emerged.
The Pragmatic Roots of Context - Bruce Edmonds - 31 MAR 99
[Next] [Previous] [Up] [Top] [Contents]
Generated with CERN WebMaker