This paper proposes a novel framework that enables a robot to learn ordinal object relations. While most related work focuses on classifying objects into dis- crete categories, such approaches cannot learn ob- ject properties (e.g., weight, height, size, etc.) that are context-specific and relative to other objects. To address this problem, we propose that a robot should learn to order objects based on ordinal ob- ject relations. In our experiments, the robot ex- plored a set of 32 objects that can be ordered by three properties: height, weight, and width. Next, the robot used unsupervised learning to discover multiple ways that the objects can be ordered based on the haptic and proprioceptive perceptions de- tected while exploring the objects. Following, the robot’s model was presented with labeled object se- ries, allowing it to ground the three ordinal relations in terms of how similar they are to the orders dis- covered during the unsupervised stage. Finally, the grounded models were used to recognize whether new object series were ordered by any of the three properties as well as to correctly insert additional objects into an existing series.

PDF, HTML

In Proceedings of the 25th International Joint Conference on Artificial Intelligence (IJCAI), New York City, USA, Jult 2016.

@inproceedings{IJCAI16-sinapov, title={Learning to Order Objects Using Haptic and Proprioceptive Exploratory Behaviors}, author={Jivko Sinapov and Priyanka Khante and Maxwell Svetlik and Peter Stone}, booktitle={Proceedings of the 25th International Joint Conference on Artificial Intelligence (IJCAI)}, month={Jult}, address={New York City, USA}, url="http://www.cs.utexas.edu/users/ai-lab?sinapov:ijcai16", year={2016} }

Slides (PDF)