US 12,217,192 B2
Distributed and contextualized artificial intelligence inference service
Francesc Guim Bernat, Barcelona (ES); Suraj Prabhakaran, Aachen (DE); Kshitij Arun Doshi, Tempe, AZ (US); Da-Ming Chiang, San Jose, CA (US); and Joe Cahill, Rathmore (IE)
Assigned to Intel Corporation, Santa Clara, CA (US)
Filed by Intel Corporation, Santa Clara, CA (US)
Filed on Dec. 30, 2022, as Appl. No. 18/091,874.
Application 18/091,874 is a continuation of application No. 17/668,844, filed on Feb. 10, 2022, granted, now 11,580,428.
Application 17/668,844 is a continuation of application No. 15/857,087, filed on Dec. 28, 2017, granted, now 11,250,336.
Prior Publication US 2023/0222363 A1, Jul. 13, 2023
Int. Cl. G06N 5/00 (2023.01); G06N 5/04 (2023.01)
CPC G06N 5/04 (2013.01) 20 Claims
OG exemplary drawing
 
1. A gateway computing device adapted for processing contextualized AI inference requests, the gateway computing device comprising:
communication circuitry to:
receive, from an edge device, an artificial intelligence (AI) inferencing operation request, wherein the AI inferencing operation request includes: a model identifier of an AI inferencing model, and contextual data; and
transmit, to a remote computing system, a request to execute a selected flavor of the AI inferencing model at the remote computing system; and
processing circuitry to:
select, in response to the AI inferencing operation request, a flavor of the AI inferencing model from among a plurality of flavors of the AI inferencing model based on the model identifier and the contextual data, wherein the flavors operate on respective hardware configurations at the remote computing system; and
transmit a communication to the remote computing system, the remote computing system configured to use the communication to initiate execution of the selected flavor of the AI inferencing model locally at the remote computing system, wherein inferencing results from the execution of the selected flavor of the AI inferencing model at the remote computing system are provided to the edge device via the gateway computing device;
wherein the gateway computing device is connected via a network to the edge device and the remote computing system.