Evaluation and control of predictive machine learning models in mobile networks
There are provided measures for evaluation and control of predictive machine learning models in mobile networks. Such measures exemplarily comprise receiving information on a predictive model related to a radio resource management function, obtaining behavior information on an intended behavior of said predicted model, obtaining difference determination information on difference determination with respect to a predictive model prediction and said intended behavior, measuring a network condition, determining a prediction result based on said network condition and said information on said predictive model, determining a behavior result based on said network condition and said behavior information, and evaluating validity of said predictive model based on said prediction result, said behavior result, and said difference determination information.
1 . A method of a mobile terminal, the method comprising:
receiving information on a predictive model related to a radio resource management function,
obtaining behavior information on an intended behavior of said predicted model,
obtaining difference determination information on difference determination with respect to a predictive model prediction and said intended behavior,
measuring a network condition,
determining a prediction result based on said network condition and said information on said predictive model,
determining a behavior result based on said network condition and said behavior information,
receiving difference threshold information indicative of a difference threshold,
evaluating validity of said predictive model based on said prediction result, said behavior result, and said difference determination information by determining a value indicative of a difference between said prediction result and said behavior result based on said difference determination information, and comparing said value indicative of said difference with said difference threshold,
deciding reduced validity of said predictive model, if said value indicative of said difference exceeds said difference threshold, and
forwarding said measured network condition, said prediction result, said behavior result and said value indicative of said difference towards a network node in response to said deciding said reduced validity of said predictive model.
2 . The method according to claim 1 , further comprising
storing said measured network condition, said prediction result, said behavior result, and said value indicative of said difference in response to said deciding said reduced validity, and
transmitting said stored measured network condition, prediction result, behavior result, and value indicative of said difference towards a network node in response to a predetermined event.
3 . The method according to claim 2 , further comprising
activating utilization of an alternative control processing with respect to said radio resource management function based on said deciding said reduced validity of said predictive model.
4 . The method according to claim 3 , further comprising
de-activating utilization of said predictive model based on said deciding said reduced validity of said predictive model.
5 . The method according to claim 4 , further comprising receiving an instruction regarding control processing with respect to said radio resource management function.
6 . The method according to claim 1 , further comprising
notifying capability for determination of a behavior according to said behavior information, and
notifying capability for determination of a difference according to said difference determination information.
7 . The method according to claim 1 , further comprising
notifying capability for evaluation of validity with respect to said radio resource management function.
8 . The method according to claim 1 , further comprising
receiving said behavior information, and
receiving said difference determination information.
9 . The method according to claim 1 , wherein
said predictive model is a predictive mobile terminal handover model, and said network condition is a reference signal received power.
10 . A method of a network node, the method comprising:
maintaining a predictive model related to a radio resource management function,
transmitting, towards a mobile terminal, information on said predictive model, and
transmitting, towards said mobile terminal, difference threshold information indicative of a difference threshold for evaluation of validity of said predictive model,
the method further comprising:
receiving a notification of capability of said mobile terminal for determination of a behavior according to said behavior information,
receiving a notification of capability of said mobile terminal for determination of a difference according to said difference determination information,
receiving, from said mobile terminal, a measured network condition, a prediction result determined based on said network condition and said information on said predictive model, a behavior result determined based on said network condition and said behavior information, and a value indicative of a difference between said prediction result and said behavior result determined based on said difference determination information,
transmitting an instruction to activate utilization of an alternative control processing with respect to said radio resource management function based on said measured network condition, prediction result, behavior result, and value indicative of said difference,
transmitting an instruction to de-activate utilization of said predictive model based on said measured network condition, prediction result, behavior result, and value indicative of said difference.
11 . The method according to claim 10 , further comprising
receiving a notification of capability of said mobile terminal for evaluation of validity with respect to said radio resource management function.
12 . The method according to claim 10 , further comprising
transmitting, towards said mobile terminal, behavior information on an intended behavior of said predicted model, and
transmitting, towards said mobile terminal, difference determination information on difference determination with respect to a predictive model prediction and said intended behavior.
13 . The method according to claim 10 , further comprising
re-training said predictive model based on said measured network condition, prediction result, behavior result, and value indicative of said difference.
14 . The method according to claim 13 , further comprising
transmitting, towards said mobile terminal, information on said re-trained predictive model.
15 . The method according to claim 10 , further comprising
transmitting, towards said mobile terminal, information on actions to be triggered upon decided reduced validity of said predictive model.
16 . An apparatus of a mobile terminal, the apparatus comprising:
at least one processor,
at least one memory including computer program code, and
at least one interface configured for communication with at least another apparatus,
the at least one processor, with the at least one memory and the computer program code, being configured to cause the apparatus to perform:
receiving information on a predictive model related to a radio resource management function,
obtaining behavior information on an intended behavior of said predicted model,
obtaining difference determination information on difference determination with respect to a predictive model prediction and said intended behavior,
measuring a network condition,
determining a prediction result based on said network condition and said information on said predictive model,
determining a behavior result based on said network condition and said behavior information,
receiving difference threshold information indicative of a difference threshold,
evaluating validity of said predictive model based on said prediction result, said behavior result, and said difference determination information, by determining a value indicative of a difference between said prediction result and said behavior result based on said difference determination information, and comparing said value indicative of said difference with said difference threshold,
deciding reduced validity of said predictive model, if said value indicative of said difference exceeds said difference threshold, and
forwarding said measured network condition, said prediction result, said behavior result and said value indicative of said difference towards a network node in response to said deciding said reduced validity of said predictive model.
17 . The apparatus according to claim 16 , wherein
the at least one processor, with the at least one memory and the computer program code, being configured to cause the apparatus to perform:
storing said measured network condition, said prediction result, said behavior result, and said value indicative of said difference in response to said deciding said reduced validity, and
transmitting said stored measured network condition, prediction result, behavior result, and value indicative of said difference towards a network node in response to a predetermined event.
18 . The apparatus according to claim 16 , wherein
the at least one processor, with the at least one memory and the computer program code, being configured to cause the apparatus to perform:
activating utilization of an alternative control processing with respect to said radio resource management function based on said deciding said reduced validity of said predictive model.
19 . The apparatus according to claim 16 , wherein
the at least one processor, with the at least one memory and the computer program code, being configured to cause the apparatus to perform:
de-activating utilization of said predictive model based on said deciding said reduced validity of said predictive model.
20 . The apparatus according to claim 16 , wherein
the at least one processor, with the at least one memory and the computer program code, being configured to cause the apparatus to perform:
receiving an instruction regarding control processing with respect to said radio resource management function.
21 . The apparatus according to claim 16 , wherein
the at least one processor, with the at least one memory and the computer program code, being configured to cause the apparatus to perform:
notifying capability for determination of a behavior according to said behavior information, and
notifying capability for determination of a difference according to said difference determination information.
22 . The apparatus according to claim 16 , wherein
the at least one processor, with the at least one memory and the computer program code, being configured to cause the apparatus to perform:
notifying capability for evaluation of validity with respect to said radio resource management function.
23 . The apparatus according to claim 16 , wherein
the at least one processor, with the at least one memory and the computer program code, being configured to cause the apparatus to perform:
receiving said behavior information, and
receiving said difference determination information.
24 . An apparatus of a network node, the apparatus comprising:
at least one processor,
at least one memory including computer program code, and
at least one interface configured for communication with at least another apparatus,
the at least one processor, with the at least one memory and the computer program code, being configured to cause the apparatus to perform:
maintaining a predictive model related to a radio resource management function,
transmitting, towards a mobile terminal, information on said predictive model, and
transmitting, towards said mobile terminal, difference threshold information indicative of a difference threshold for evaluation of validity of said predictive model,
the at least one processor, with the at least one memory and the computer program code, being configured to cause the apparatus to perform:
receiving a notification of capability of said mobile terminal for determination of a behavior according to said behavior information,
receiving a notification of capability of said mobile terminal for determination of a difference according to said difference determination information,
receiving a notification of capability of said mobile terminal for evaluation of validity with respect to said radio resource management function,
receiving, from said mobile terminal, a measured network condition, a prediction result determined based on said network condition and said information on said predictive model, a behavior result determined based on said network condition and said behavior information, and a value indicative of a difference between said prediction result and said behavior result determined based on said difference determination information,
transmitting an instruction to activate utilization of an alternative control processing with respect to said radio resource management function based on said measured network condition, prediction result, behavior result, and value indicative of said difference,
transmitting an instruction to de-activate utilization of said predictive model based on said measured network condition, prediction result, behavior result and value indicative of said difference.
25 . The apparatus according to claim 24 , wherein
the at least one processor, with the at least one memory and the computer program code, being configured to cause the apparatus to perform:
transmitting, towards said mobile terminal, behavior information on an intended behavior of said predicted model, and
transmitting, towards said mobile terminal, difference determination information on difference determination with respect to a predictive model prediction and said intended behavior.
26 . The apparatus according to claim 24 , wherein
the at least one processor, with the at least one memory and the computer program code, being configured to cause the apparatus to perform:
re-training said predictive model based on said measured network condition, prediction result, behavior result, and value indicative of said difference.
27 . The apparatus according to claim 26 , wherein
the at least one processor, with the at least one memory and the computer program code, being configured to cause the apparatus to perform:
transmitting, towards said mobile terminal, information on said re-trained predictive model.
28 . The apparatus according to claim 24 , wherein
the at least one processor, with the at least one memory and the computer program code, being configured to cause the apparatus to perform:
transmitting, towards said mobile terminal, information on actions to be triggered upon decided reduced validity of said predictive model.