Affordance Discovery using Simulated Exploration. Allevato, A., Thomaz, A., & Pryor, M. In Proceedings of the 17th International Conference on Autonomous Agents and MultiAgent Systems, of AAMAS '18, pages 2174–2176, Stockholm, Sweden, July, 2018. International Foundation for Autonomous Agents and Multiagent Systems.
abstract   bibtex   
Allowing robots to understand their world in terms of affordances allows for generalization, learning, and complex planning, while also being intuitive for humans to understand. In recent work, affordances are often learned with hand-coded robot actions, which can limit or bias the model. Real-world training has also been used to learn affordances and manipulation models, but is timeconsuming and unsafe for the robot and its environment.
@inproceedings{allevato_affordance_2018,
	address = {Stockholm, Sweden},
	series = {{AAMAS} '18},
	title = {Affordance {Discovery} using {Simulated} {Exploration}},
	abstract = {Allowing robots to understand their world in terms of affordances allows for generalization, learning, and complex planning, while also being intuitive for humans to understand. In recent work, affordances are often learned with hand-coded robot actions, which can limit or bias the model. Real-world training has also been used to learn affordances and manipulation models, but is timeconsuming and unsafe for the robot and its environment.},
	urldate = {2020-05-09},
	booktitle = {Proceedings of the 17th {International} {Conference} on {Autonomous} {Agents} and {MultiAgent} {Systems}},
	publisher = {International Foundation for Autonomous Agents and Multiagent Systems},
	author = {Allevato, Adam and Thomaz, Andrea and Pryor, Mitch},
	month = jul,
	year = {2018},
	keywords = {affordances, feature selection, human-guided learning, mental simulation, robot simulation, transfer learning},
	pages = {2174--2176},
}

Downloads: 0