Commonsense Reasoning to Guide Deep Learning for Scene Understanding (Extended Abstract)

Commonsense Reasoning to Guide Deep Learning for Scene Understanding (Extended Abstract)

Mohan Sridharan, Tiago Mota

Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence
Sister Conferences Best Papers. Pages 4760-4764. https://doi.org/10.24963/ijcai.2020/661

Our architecture uses non-monotonic logical reasoning with incomplete commonsense domain knowledge, and incremental inductive learning, to guide the construction of deep network models from a small number of training examples. Experimental results in the context of a robot reasoning about the partial occlusion of objects and the stability of object configurations in simulated images indicate an improvement in reliability and a reduction in computational effort in comparison with an architecture based just on deep networks.
Keywords:
Knowledge Representation and Reasoning: Non-monotonic Reasoning, Common-Sense Reasoning
Machine Learning: Deep Learning
Machine Learning: Online Learning
Robotics: Robotics and Vision