Method and devices for transfer learning for inductive tasks in radio access network
AssigneeTelefonaktiebolaget LM Ericsson (publ)
InventorsEuhanna GHADIMI, Jonas WIOREK
Life of this patent
Estimating the term…
Abstract
A framework for inductive tasks in RANs includes a source learner, a target learner and an inference entity that can be merged into existing or future RAN functions. The signaling mechanism between different framework entities may also be merged into existing or future network interfaces. The framework can be reused by multiple inductive tasks to perform a variety of tasks such as prediction of some network phenomena or to improve certain KPIs in the network.
Claims21
A network operator device performing as a source learner in a radio access network, RAN, the network operator device comprising:
- a communication interface configured to intermediate communication with a target learner in the RAN;
- and
- a data processing unit connected to the communication interface and configured to prepare and supply source inductive parameters, SIPs, via the communication interface to the target learner, wherein the SIPs enable the target learner to generate a production model of a RAN inductive task for an inference entity in the RAN, and the SIPs include any one or more of:
- an inductive model of an inductive task different from the RAN inductive task;
- an inductive model of a task for a cell of a first cell type, and the RAN inductive task is for a cell of a second cell type;
- an inductive model generated from data related to multiple cells and the production model is for a specific cell;
- and
- an inductive model trained using first data and the production model is trained using second data which is different from the first data.
-5. (canceled)
-13. (canceled)
A network operator device performing as a target learner in a radio access network, RAN, the network operator device comprising:
- a communication interface configured to intermediate communication with a source learner and an inference entity in the RAN;
- and
- a data processing unit connected to the communication interface and configured to receive source inductive parameters, SIPs, via the communication interface from the source learner, to use the SIPs for generating a production model of a RAN inductive task, and to transmit the production model to the inference entity, wherein the SIPs include any one or more of:
- an inductive model of an inductive task different from the RAN inductive task;
- an inductive model of a task for a cell of a first cell t e and the RAN inductive task is for a cell of a second cell type;
- an inductive model generated from data related to multiple cells and the production model is for a specific cell;
- and
- an inductive model trained using first data and the production model is trained using second data which is different from the first data.
-18. (canceled)
A method for a network operator device in a radio access network, RAN, the method comprising:
- receiving source inductive parameters, SIPs, from a source learner in the RAN;
- generating a production model of a RAN inductive task using the SIPs;
- and
- transmitting the production model to an inference entity in the RAN;
- wherein the SIPs include any one or more of:
- an inductive model of an inductive task different from the RAN inductive task;
- an inductive model of a task for a cell of a first cell type, and wherein the RAN inductive task is for a cell of a second cell type;
- an inductive model generated from data related to multiple cells and the production model is for a specific cell;
- and
- an inductive model trained using first data and the production model is trained using second data which is different from the first data.
-27. (canceled)
A network operator device performing as an inference entity in a radio access network, RAN, the network operator device comprising:
- a communication interface configured to intermediate communication via the RAN;
- and
- a data processing unit connected to the interface and configured
- to receive a production model of a RAN inductive task from a target learner, wherein the production model of the RAN inductive task is generated using source inductive parameters including any one or more of:
- an inductive model of an inductive task different from the RAN inductive task;
- an inductive model of a task for a cell of a first cell type, and the RAN inductive task is for a cell of a second cell type;
- an inductive model generated from data related to multiple cells and the production model is for a specific cell;
- and
- an inductive model trained using first data and the production model is trained using second data which is different from the first data:
- to produce input for the production model from RAN-related parameters;
- to execute the production model of the RAN inductive task using the input to yield an output related to a RAN function;
- and
- to use the output to determine the RAN function.
-35. (canceled)
Dependencies are inferred from the claim text.