Parameterized Action Representation (PAR)

Natural Language often describes actions at a high level, leaving out many of the details that have to be specified for animation.  The PAR bridges the gap between natural language and animations. PAR gives a description of an action.  The PAR has to specify the agent of the action as well as any relevant objects and information about path, location, manner, and purpose for a particular action. There are linguistic constraints on how this information can be conveyed by the language; agents and objects tend to be verb arguments, path is often a prepositional phrase, and manner and purpose might be in additional clauses.  A parser and translator map the components of an instruction into the parameters or variables of the PAR, which is then linked directly to PaT-Nets (finite state machines) executing the specified movement generators.

We use the example “Walk to the door and turn the handle slowly,” to illustrate the function of the PAR.  Whether or not the PAR system processes this instruction, there is nothing explicit in the linguistic representation about grasping the handle or which direction it will have to be turned, yet this information is necessary to the action's actual visible performance.  The PAR has to include information about applicability, preparatory, and termination conditions in order to fill in these gaps. It also has to be parameterized, because other details of the action depend on the PAR's participants, including agents, objects, and other attributes.  The representation of the “handle” object lists the actions that the object can perform and what state changes they cause. The number of steps it will take to get to the door depends on the agent's size and starting location.

Another resource for information on verbs is the Unified Verb Index, which merges links and webpages from a few different natural language processing projects. WordNet contains valuable information on both verbs and nouns. WordsEye is an interesting system that uses natural language to construct virtual scenes.

Software download coming soon-ish.

Recommended Reference:

R. Bindiganavale, W. Schuler, J. Allbeck, N. Badler, A. Joshi, and M. Palmer. "Dynamically Altering Agent Behaviors Using Natural Language Instructions." Proceedings of the 4th International Conference on Autonomous Agents 2000 (AGENTS 2000) , pages 293-300.


C++ PaT-Nets Manual

PAR Database Documentation (2008)

PAR C++ API (2007)

Example working out database representation of PAR Constraint Graph (2005)


PAR Implementation

Image Gallery