: An Architecture for Knowledge Representation and Reasoning in Robotics" by Shiqi Zhang and colleagues.
Using probabilistic models to handle the messy, quantitative "noise" of the real world, like unreliable sensors or imperfect physical actions. Key Findings Kr (3) mp4
In experiments with physical robots, this combined architecture reduced the time it took to finish tasks by 39% compared to traditional methods. : An Architecture for Knowledge Representation and Reasoning
Using a declarative "action language" to help the robot understand qualitative knowledge and prioritized rules. quantitative "noise" of the real world