Granted Patent B2

US 12,596,962 · App. 17/760,232 · Granted Apr 7, 2026

Inventors: Mats Folkesson (Täby, SE); Bin Sun (Luleå, SE); Rafia Inam (Västerås, SE); Elena Fersman (Stockholm, SE); Daniel Lindström (Luleå, SE)

Assignee: TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)

G06N20/20G06N20/00

Abstract

Claims (53)

1 . A method for providing data to a destination node connected to one or more source nodes via one or more networks, the method being performed by a feature identifier for selecting at least one feature from a set of features, the set of features comprising a first feature and a second feature, wherein

each feature included in the set of features is associated with a data set such that the first feature is associated with a first data set and the second feature is associated with a second data set,

the first data set associated with the first feature comprises values for the first feature,

the second data set associated with the second feature comprises values for the second feature, and

the method comprises:

for each feature included in the set of features, obtaining a cost value for the feature, the cost value indicating a cost of providing the data set associated with the feature from a source node that stores the data set to the destination node that is connected to the source node via a network;

for each feature included in the set of features, obtaining an importance value indicating an importance of the feature;

for each feature included in the set of features, calculating a priority value using the importance value for the feature and the cost value for the feature;

based on the calculated priority values, selecting a subset of the set of features; and

for each selected feature, initiating the transmission of the respective data set associated with the respective selected feature from the source node that stores the respective data set associated with the respective selected feature to the destination node via the network that connects the destination node with the source node that stores the respective data set associated with the respective selected feature, wherein

the destination node is configured to execute a machine learning process that is operable to use the respective data sets to produce a model, and/or

the destination node is configured to execute a model generated by a machine learning process, and the model is operable to use the respective data sets to produce an inference.

2 . The method of claim 1 , wherein

the step of obtaining the cost value for each feature included in the set of features comprises obtaining a first cost value for the first feature, and

the first cost value is a function of size of the first data set.

3 . The method of claim 2 , wherein

the method further comprises obtaining network performance information for the network that connects the destination node with the source node that stores the first data set associated with the first feature, and

the first cost value is also the function of network performance information.

4 . The method of claim 1 , wherein the first cost value is inversely proportional to the feature importance value.

5 . The method of claim 1 , wherein

the subset of features includes the first feature but not the second feature, and

the method further comprises:

after initiating the transmission of the data sets for the selected features, receiving from the destination node a model training result pertaining to a training of a model using the first data set associated with the first feature; and

after receiving the model training result, initiating the transmission of the second data set associated with the second feature from the source node that stores the second data set associated with the second feature to the destination node.

6 . The method of claim 5 , wherein the model training result comprises a first feature importance value for the first feature, the first feature importance value indicating the first feature's impact on the performance of the model.

7 . A non-transitory computer readable storage medium storing a computer program comprising instructions which when executed by processing circuitry causes the processing circuitry to perform the method of claim 1 .

8 . An apparatus for providing data to a destination node connected to one or more source nodes via one or more networks and for selecting at least one feature from a set of features, the set of features comprising a first feature and a second feature, wherein

each feature included in the set of features is associated with a data set such that the first feature is associated with a first data and the second feature is associated with a second data set,

the first data set comprises values for the first feature,

the second data comprises values for the second feature, and

the apparatus including:

a memory; and

processing circuitry, wherein the apparatus is configured to:

for each feature included in the set of features, obtain a cost value for the feature, the cost value indicating a cost of providing the data set associated with the feature from a source node that stores the data set to the destination node that is connected to the source node via a network;

for each feature included in the set of features, obtain an importance value indicating an importance of the feature;

for each feature included in the set of features, calculate a priority value using the importance value for the feature and the cost value for the feature;

based on the calculated priority values, select a subset of the set of features; and

for each selected feature, initiate the transmission of the respective data set associated with the respective selected feature from the source node that stores the respective data set associated with the respective selected feature to the destination node via the network that connects the destination node with the source node that stores the respective data set associated with the respective selected feature, wherein the destination node is configured to:

execute a machine learning process that is operable to use the respective data sets to produce a model, and/or

execute a model generated by a machine learning process, and the model is operable to use the respective data sets to produce an inference.

9 . The apparatus of claim 8 , wherein

the apparatus is further configured to obtain the value for each feature included in the set of features by performing a process that includes obtaining a first cost value for the first feature, and

the first cost value is a function of size of the first data set.

10 . The apparatus of claim 8 , wherein

the apparatus is further configured to obtain network performance information for the network that connects the destination node with the source node that stores the first data set associated with the first feature, and

the first cost value is also a function of network performance information.

11 . The apparatus of claim 8 , wherein the first cost value is inversely proportional to the feature importance value.

12 . The apparatus of claim 8 , wherein

the subset of features includes the first feature but not the second feature, and

the apparatus is further configured to:

after initiating the transmission of the data sets for the selected features, receive from the destination node a model training result pertaining to a training of a model using the first data set associated with the first feature; and

after receiving the model training result, initiate the transmission of the second data set associated with the second feature from the source node that stores the second data set associated with the second feature to the destination node.

13 . The apparatus of claim 12 , wherein the model training result comprises a first feature importance value for the first feature, the first feature importance value indicating the first feature's impact on the performance of the model.

Assignments (1)

Continuity (1)

References Cited (23)