Anticipatory learning classifier systems
Anticipations are an important part of human behaviour. Anticipatory Learning Classifier Systems intend to apply this insight by simulation and utilization of anticipations in an artificial behavioural learning system. The book introduces such a system, namely ACS2. The system autonomously evolves an internal environmental model represented by a population of condition-action-effect classifiers. ACS2 is acting on an environment and is forming a model of the encountered environment. In seven chapters (Background, ACS2, Experiments with ACS2, Limits, model exploitation, related systems, concluding summary) it is shown how ACS2 evolves a complete, accurate, and compact environmental model, i.e. a model that anticipates the perceptual effects of an action in each situation accurately with a least number of maximally general classifiers. The model is built by two distinct processes: an anticipatory learning process and a genetic generalization. The two processes achieve the evolution of the intended environmental model. Despite some small over-generalization cases caused by the genetic generalization process, the extensive experimental study confirmed that the model evolves very reliably and genetic generalization results in great generalization capabilities. The final part discusses future work with ACS2. The appendix contains a detailed algorithmic description of ACS2 and a C++ program code documentation to investigate own experiments with anticipatory learning classifier systems.
- The anticipatory classifier system and genetic generalization
- scientific article; zbMATH DE number 1154419
- Introduction to learning classifier systems
- Combining latent learning with dynamic programming in the modular anticipatory classifier system
- Rule-based evolutionary online learning systems. A principle approach to LCS analysis and design.
- Ten years of genetic fuzzy systems: Current framework and new trends.
- YACS: A new learning classifier system using anticipation
- The anticipatory classifier system and genetic generalization
- Combining latent learning with dynamic programming in the modular anticipatory classifier system
- An approach to active learning for classifier systems
- The anticipatory profile. An attempt to describe anticipation as process
- Anticipation in biological and cognitive systems: the need for a physical theory of biological organization
- scientific article; zbMATH DE number 1154419 (Why is no real title available?)
- scientific article; zbMATH DE number 2040732 (Why is no real title available?)
- scientific article; zbMATH DE number 2087893 (Why is no real title available?)
- scientific article; zbMATH DE number 2087906 (Why is no real title available?)
- scientific article; zbMATH DE number 1912669 (Why is no real title available?)
- Categorical system theory and the phenomenological calculus
This page was built for publication: Anticipatory learning classifier systems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4328477)
