| CPC G06N 3/08 (2013.01) [G06F 40/20 (2020.01); G06F 40/30 (2020.01); G06N 3/044 (2023.01)] | 12 Claims |
|
1. A system for intelligent service intermediation, comprising:
one or more service edge devices comprising a plurality of programming instructions stored in a memory of, and operating on at least one processor of, a computing device, wherein the plurality of programming instructions, when operating on the at least one processor, causes the computing device to:
receive updated local machine and deep learning global model parameters;
apply the updated local machine and deep learning global model parameters to the local models stored in the service edge device;
receive local data from service edge device sensors;
feed the received local data as input into one or more of the updated local machine and deep learning models to generate output actions responsive to a service edge device user query;
execute output actions and set up processes necessary for fulfillment of the service edge device user query;
generate a voice or text message relayed by a virtual assistant indicating that the service edge device user query has been received, understood, and an action responsive to the user query has been initialized or completed;
train and update local machine and deep learning models using the received local data;
upload trained and updated local model parameters to a service intermediary server; and
execute service actions received from a service intermediary server; and
a plurality of service intermediary servers arranged as a federated network in which each service intermediary server of the plurality of service intermediary servers is linked to, and configured to share information with, each of the other service intermediary servers, and wherein each service intermediary server comprises a plurality of programming instructions stored in a memory of, and operating on at least one processor of, a computing device, wherein the plurality of programming instructions, when operating on the at least one processor, causes the computing device to:
select a subset of service edge devices to which to upload trained and updated local model parameters, wherein at least one of the service edge devices selected is subscribed to a different service intermediary server of the plurality of service intermediary servers arranged as a federated network from the service intermediary server of the plurality of service intermediary servers arranged as a federated network performing the selection;
aggregate the received updated local model parameters from the subset of service edge devices and compute the average value of the local model parameters;
update the local machine and deep learning global models using the computed average values of the local model parameters as new global model parameters;
send the updated local machine and deep learning global models to the one or more service edge devices;
receive and store global data from service edge devices and external sources;
feed the global data into the local machine and deep learning global models to generate as output a service action; and
send the service action to one or more service edge devices for execution;
wherein the local machine and deep learning global models comprise a graph-based neural network.
|