We demonstrate how a robot can learn new knowledge about objects, objects parts, and actions from a single spoken instruction and use it immediately for performing tasks. We detail the modifications to several architectural components required to enable such fast learning.
@inproceedings{scheutzetal17aamas, title={Spoken Instruction-Based One-Shot Object and Action Learning in a Cognitive Robotic Architecture}, author={Matthias Scheutz and Evan Krause and Brad Oosterveld and Tyler Frasca and Robert Platt}, year={2017}, booktitle={Proceedings of the 16th International Conference on Autoomous Agents and Multiagent Systems}, url={https://hrilab.tufts.edu/publications/scheutzetal17aamas.pdf} }