Device and method for optimising model performance
AssigneeKabushiki Kaisha Toshiba
InventorsAftab KHAN, Parag Kulkarni, Michael Jones
Life of this patent
Estimating the term…
Abstract
A device for optimising model performance comprising: a memory resource storing a local model on an electronic device, the local model being obtained by the use of a machine learning process; a first processor configured to: obtain data at the electronic device; apply the local model to the data or further data derived from the data to obtain at least one local model result; generate a measure of the quality of the local model result, and send the data, further data derived from the data and/or the local model result through a communication device to a second processor for use in updating a global model based on the measure of the quality of the local model result.
Claims20
A device for optimising model performance comprising:
- a memory resource storing a local model on an electronic device, the local model being obtained by the use of a machine learning process;
- a first processor configured to:
- obtain data at the electronic device;
- apply the local model to the data or further data derived from the data to obtain at least one local model result;
- generate a measure of the quality of the local model result, and
- send the data, further data derived from the data and/or the local model result through a communication device to a second processor for use in updating a global model based on the measure of the quality of the local model result.
A method comprising:
- obtaining data at an electronic device;
- applying a local model on the electronic device to the data or further data derived from the data to obtain at least one local model result, wherein the local model is obtained by the use of a machine learning process;
- generating a measure of the quality of the local model result using a first processor;
- and
- sending the data, further data derived from the data and/or the local model result to a second processor for use in updating a global model based on the measure of the quality of the local model result.
A method comprising:
- receiving, by a second processor, data, further data derived from the data and/or at least one local model result sent from a first processor on an electronic device, the second processor including a global model, wherein the second processor is linked by a network to a plurality of electronic devices;
- activating, by the second processor, an update of the global model by a machine learning process based on a measure of the quality of the local model result from the local model to obtain an updated global model;
- sending updated global model data to at least one of the electronic devices to update their respective local models based on the quality of the local model result.
Dependencies are inferred from the claim text.