Projective Visualization: Acting from Experience
Abstract
This paper describes Projective Visualization which uses previous observation of a process or activity to project the results of an agent's actions into the future. Actions which seem likely to sueceed are selected and applied. Actions which seem likely to fail are rejected, and other actions can be generated and evaluated. This paper presents a description of the architecture for Projective Visualization, preliminary results on learning to act from observations of a reactive system, and a comparison of two types of Case Projection (how situations are projected into the future).
Cite
Text
Goodman. "Projective Visualization: Acting from Experience." AAAI Conference on Artificial Intelligence, 1993.Markdown
[Goodman. "Projective Visualization: Acting from Experience." AAAI Conference on Artificial Intelligence, 1993.](https://mlanthology.org/aaai/1993/goodman1993aaai-projective/)BibTeX
@inproceedings{goodman1993aaai-projective,
title = {{Projective Visualization: Acting from Experience}},
author = {Goodman, Marc},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {1993},
pages = {54-59},
url = {https://mlanthology.org/aaai/1993/goodman1993aaai-projective/}
}