This is a great idea.But what are the current limitations?Current models used in probabilistic perception are at the surface level.

This idea arises in many perception tasks. In vision, we can have high-level models of activities, and of how people interact with each other, giving us much better models for visual tracking.However, I donít have much time left, soIíll spend just a couple of slides talking
about it in the context of natural language processing.