Distributed training in communication networks

1 . An apparatus, comprising:

one or more processors, and

at least one memory storing instructions that, when executed by the one or more processors, cause the apparatus to:

monitor if a request to train a machine learning sub-model is received from a meta-training host;

generate training data;

train the machine learning sub-model using at least a first subset of the training data if the request is received and at least the first subset of the training data is generated;

supervise if the training of the machine learning sub-model fulfills a maturity condition;

check if a predefined condition related to the machine learning sub-model is fulfilled;

provide the trained machine learning sub-model and at least a second subset of the first subset of training data used to train the machine learning sub-model to the meta-training host if the predefined condition is fulfilled; and

inhibit the providing the trained machine learning sub-model if a maturity level does not fulfill the maturity condition,

wherein the maturity condition is received from the meta-training host, and

wherein the maturity condition comprises:

specified lowest thresholds for a test or a validation score,

a minimum number of observations used to train the machine learning sub-model, and

a minimum number of observed network events.

2 . The apparatus according to claim 1 , wherein the predefined condition comprises that the training of the machine learning sub-model fulfills the maturity condition.

3 . The apparatus according to claim 1 , wherein the predefined condition comprises that a poll from the meta-training host is received.

4 . The apparatus according to claim 1 , wherein the instructions, when executed by the one or more processors, further cause the apparatus to:

actively involve in a communication; wherein

the training data are generated based on the actively involving.

5 . The apparatus according to claim 1 , wherein the instructions, when executed by the one or more processors, further cause the apparatus to:

monitor if a trained meta-model is received from the meta-training host; and

use the trained meta-model for inference if the trained meta-model is received.

6 . The apparatus according to claim 5 , wherein the instructions, when executed by the one or more processors, further cause the apparatus to:

monitor if a plurality of further sub-models are received from the meta-training host; and

use the received further sub-models and the trained meta-model for inference if the trained meta-model and the further sub-models are received.