US 11,836,640 B2
Artificial intelligence modules for computation tasks
Zhengping Ji, Hinsdale, IL (US); Rachid M. Alameh, Crystal Lake, IL (US); and Robert S. Witte, Algonquin, IL (US)
Assigned to Motorola Mobility LLC, Chicago, IL (US)
Filed by Motorola Mobility LLC, Chicago, IL (US)
Filed on May 15, 2020, as Appl. No. 16/875,279.
Prior Publication US 2021/0357780 A1, Nov. 18, 2021
Int. Cl. G06F 17/00 (2019.01); G06N 5/04 (2023.01); G06N 20/00 (2019.01)
CPC G06N 5/04 (2013.01) [G06N 20/00 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A method, comprising:
determining that a particular computation task is to be performed;
identifying multiple artificial intelligence (AI) modules that are combinable to perform different computation tasks, each AI module being configured to perform a function pertaining to the different computation tasks;
aggregating, based on parameters of the particular computation task, a set of AI modules from the multiple AI modules to generate a task module for performing the particular computation task;
ordering the AI modules of the set of AI modules in the task module based at least in part on an input type for each AI module of the set of AI modules and an output type of each AI module;
linking the task module to an instance of a client device based at least in part on a mapping of one or more client devices to sensor capabilities of the one or more client devices, the mapping indicating that the client device includes a sensor capable of providing sensor data corresponding to a first input type for a first AI module of the set of AI modules;
exposing the task module for use in performing the particular computation task; and
causing the computation task to be performed by the task module to generate an output of the computation task from the task module, including causing the sensor data to be obtained from the client device and causing the sensor data to be provided as the first input type to the first AI module.