Advanced artificial intelligence agent for modeling physical interactions

Abstract

Described herein are advanced artificial intelligence agents for modeling physical interactions. In one embodiment, an apparatus to provide an active artificial intelligence (AI) agent includes at least one database to store physical interaction data and compute cluster coupled to the at least one database. The compute cluster automatically obtains physical interaction data from a data collection module without manual interaction, stores the physical interaction data in the at least one database, and automatically trains diverse sets of machine learning program units to simulate physical interactions with each individual program unit having a different model based on the applied physical interaction data.

1 . (canceled) 
     
     
         2 . A method comprising:
 applying physical interaction data to multiple machine learning program units to train the machine learning program units to simulate physical interactions, wherein each of the machine learning program units includes a different model based on the applied physical interaction data;   jointly approximating and modeling behaviors of each of the machine learning program units to train a master program unit; and   applying input to a master program unit to generate predicted physical interactions based on the training of the machine learning program units and the master program unit.   
     
     
         3 . The method of  claim 2 , wherein:
 the physical interaction data represents interactions between a human and a physical object.   
     
     
         4 . The method of  claim 2 , wherein:
 the physical interactions include pushing, grasping, and rotating.   
     
     
         5 . The method of  claim 2 , wherein:
 the master program unit comprises a Bayesian program.   
     
     
         6 . The method of  claim 2 , wherein:
 the machine learning program units comprise Bayesian program units.   
     
     
         7 . The method of  claim 2 , wherein:
 the machine learning program units include deep neural network (DNN) models.   
     
     
         8 . The method of  claim 2 , further comprising:
 accessing the physical interaction data from one or more databases.   
     
     
         9 . At least one non-transitory machine-readable medium comprising a plurality of instructions, which, when executed on a computing device, cause the computing device to perform a method comprising:
 applying physical interaction data to multiple machine learning program units to train the machine learning program units to simulate physical interactions, wherein each of the machine learning program units includes a different model based on the applied physical interaction data;   jointly approximating and modeling behaviors of each of the machine learning program units to train a master program unit; and   applying input to a master program unit to generate predicted physical interactions based on the training of the machine learning program units and the master program unit.   
     
     
         10 . The at least one non-transitory machine-readable medium of  claim 9 , wherein:
 the physical interaction data represents interactions between a human and a physical object.   
     
     
         11 . The at least one non-transitory machine-readable medium of  claim 9 , wherein:
 the physical interactions include pushing, grasping, and rotating.   
     
     
         12 . The at least one non-transitory machine-readable medium of  claim 9 , wherein:
 the master program unit comprises a Bayesian program.   
     
     
         13 . The at least one non-transitory machine-readable medium of  claim 9 , wherein:
 the machine learning program units comprise Bayesian program units.   
     
     
         14 . The at least one non-transitory machine-readable medium of  claim 9 , wherein:
 the machine learning program units include deep neural network (DNN) models.   
     
     
         15 . The at least one non-transitory machine-readable medium of  claim 9 , wherein the method further comprises:
 accessing the physical interaction data from one or more databases.   
     
     
         16 . A system comprising:
 one or more processors to:
 apply physical interaction data to multiple machine learning program units to train the machine learning program units to simulate physical interactions, wherein each of the machine learning program units includes a different model based on the applied physical interaction data; 
 jointly approximate and model behaviors of each of the machine learning program units to train a master program unit; and 
 apply input to a master program unit to generate predicted physical interactions based on the training of the machine learning program units and the master program unit. 
   
     
     
         17 . The system of  claim 16 , further comprising:
 logic to access the physical interaction data from one or more databases.   
     
     
         18 . The system of  claim 16 , wherein:
 the physical interaction data represents interactions between a human and a physical object.   
     
     
         19 . The system of  claim 16 , wherein:
 the physical interactions include pushing, grasping, and rotating.   
     
     
         20 . The system of  claim 16 , wherein:
 the one or more processors includes a graphics processing unit (GPU).   
     
     
         21 . The system of  claim 16 , further comprising one or more of:
 a display device, a network adapter coupled with the one or more processors, a storage device coupled with the one or more processors, and system memory coupled with the one or more processors.