We demonstrate a novel computational architecture for affordance perception. Our architecture allows robotic agents to reason that dirty knives are typically not used for cutting vegetables, even though they can functionally accomplish the task.
@article{sarathyetal16ieeej,
title={A Logic-based Computational Framework for Inferring Cognitive Affordances},
author={Vasanth Sarathy and Matthias Scheutz},
year={2016},
journal={IEEE Transactions on Cognitive and Developmental Systems},
volume={8},
url={https://hrilab.tufts.edu/publications/sarathyetal16ieeej.pdf}
}