Applying conditions for input parameters for UE-sided prediction using an ML model

Abstract

A second network node received, from first network node, configuration for performing prediction by a ML model, including input condition(s) related to input parameters or values of the input parameters for the ML model. The second network node determines whether input parameters for the ML model do or do not fulfill the input condition(s). The second network node performs, responsive to the determining, one or both of prediction using the ML model or reporting to the first network node of information based on whether or not the prediction was performed. The first network node sends the configuration to the second network node, and receives one or both of output of a prediction performed using the ML model on the second network node or reporting from the second network node of information based on whether or not the prediction was performed.

Description

APPLYING CONDITIONS FOR INPUT PARAMETERS FOR UE-SIDED PREDICTION USING AN ML MODELTECHNICAL FIELD

[0001] Examples of embodiments herein relate generally to using Machine Learning (ML) models for beam prediction by a User Equipment (UE) in a cellular network, and, more specifically, relate to prediction by the UE applying conditions for input parameters for the ML models.BACKGROUND

[0002] Machine Learning (ML) models are being deployed in many locations in wireless networks such as cellular networks. These models can be trained by the network or other elements, and then used by the network or by User Equipment (UEs), which are devices that connect to the network. For instance, a UE may perform measurements on beams that the network uses to communicate. Those measurements on the beams (and potentially other input parameters) can be input to an ML model, which has been correspondingly trained to output an indication of a particular beam that should be used by the network for communication from the network to the UE. The ML models may be trained for many other communication parameters.BRIEF SUMMARY

[0003] This section is intended to include examples and is not intended to be limiting.

[0004] In an exemplary embodiment, a method is disclosed that includes receiving, from a first network node by a second network node, configuration for performing prediction by a machine-learning model, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; determining, by the second network node, whether input parameters for the machine-learning model do or do not fulfill the one or more input conditions; and performing, by the second network node responsive to the determining, one or both of prediction using the machine-learning model or reporting to the first network node of information based on whether or not the prediction was performed.

[0005] An additional exemplary embodiment includes a computer program, comprising instructions for performing the method of the previous paragraph, when the computer program is run on an apparatus. The computer program according to this paragraph, wherein the computer program is a computer program product comprising a computer-readable medium bearing the instructions embodied therein for use with the apparatus. Another example is the computer program according to this paragraph, wherein the program is directly loadable into an internal memory of the apparatus.

[0006] An exemplary apparatus includes one or more processors and one or more memories storing instructions that, when executed by the one or more processors, cause the apparatus at least to perform: receiving, from a first network node by a second network node, configuration for performing prediction by a machine-learning model, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; determining, by the second network node, whether input parameters for the machine-learning model do or do not fulfill the one or more input conditions; and performing, by the second network node responsive to the determining, one or both of prediction using the machine-learning model or reporting to the first network node of information based on whether or not the prediction was performed.

[0007] An exemplary computer program product includes a computer-readable storage medium bearing instructions that, when executed by an apparatus, cause the apparatus to perform at least the following: receiving, from a first network node by a second network node, configuration for performing prediction by a machine-learning model, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; determining, by the second network node, whether input parameters for the machine-learning model do or do not fulfill the one or more input conditions; and performing, by the second network node responsive to the determining, one or both of prediction using the machine-learning model or reporting to the first network node of information based on whether or not the prediction was performed.

[0008] In another exemplary embodiment, an apparatus comprises means for performing: receiving, from a first network node by a second network node, configuration for performing prediction by a machine-learning model, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; determining, by the second network node, whether input parameters for the machine-learning model do or do not fulfill the one or more input conditions; and performing, by the second network node responsive to the determining, one or both of prediction using the machine-learning model or reporting to the first network node of information based on whether or not the prediction was performed.

[0009] In an exemplary embodiment, a method is disclosed that includes sending, from a first network node to a second network node, configuration for performing prediction by a machine-learning model at the second network node, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machinelearning model; and receiving, by the first network node from the second network node, one or both of output of a prediction performed using the machine-learning model on the second network node or reporting from the second network node of information based on whether or not the prediction was performed.

[0010] An additional exemplary embodiment includes a computer program, comprising instructions for performing the method of the previous paragraph, when the computer program is run on an apparatus. The computer program according to this paragraph, wherein the computer program is a computer program product comprising a computer-readable medium bearing the instructions embodied therein for use with the apparatus. Another example is the computer program according to this paragraph, wherein the program is directly loadable into an internal memory of the apparatus.

[0011] An exemplary apparatus includes one or more processors and one or more memories storing instructions that, when executed by the one or more processors, cause the apparatus at least to perform: sending, from a first network node to a second network node, configuration for performing prediction by a machine-learning model at the second network node, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; and receiving, by the first network node from the second network node, one or both of output of a prediction performed using the machine-learning model on the second network node or reporting from the second network node of information based on whether or not the prediction was performed.

[0012] An exemplary computer program product includes a computer-readable storage medium bearing instructions that, when executed by an apparatus, cause the apparatus toperform at least the following: sending, from a first network node to a second network node, configuration for performing prediction by a machine-learning model at the second network node, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; and receiving, by the first network node from the second network node, one or both of output of a prediction performed using the machine-learning model on the second network node or reporting from the second network node of information based on whether or not the prediction was performed.

[0013] In another exemplary embodiment, an apparatus comprises means for performing: sending, from a first network node to a second network node, configuration for performing prediction by a machine-learning model at the second network node, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; and receiving, by the first network node from the second network node, one or both of output of a prediction performed using the machine-learning model on the second network node or reporting from the second network node of information based on whether or not the prediction was performed.BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In the attached drawings:

[0015] FIG. 1 is a flowchart of a method where conditions are applied for input data for UE-sided prediction using an ML model;

[0016] FIG. 2 is a block diagram showing a ML model and its input parameters and outputs, and also shows examples of input condition(s) that are related to input parameters or values of the input parameters for the ML model;

[0017] FIG. 3 is a signaling diagram and flowchart of applying conditions for input parameters for UE-sided prediction using an ML model; and

[0018] FIG. 4 is a block diagram of one possible and non-limiting exemplary system in which the exemplary embodiments may be practiced.DETAILED DESCRIPTION OF THE DRAWINGS

[0019] Abbreviations that may be found in the specification and / or the drawing figures are defined below, at the end of the detailed description section.

[0020] The word “exemplary” is used herein to mean “serving as an example, instance, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments. All of the embodiments described in this Detailed Description are exemplary embodiments provided to enable persons skilled in the art to make or use the invention and not to limit the scope of the invention which is defined by the claims.

[0021] When more than one drawing reference numeral, word, or acronym is used within this description withand in general as used within this description, the “ / ” may be interpreted as “or”, “and”, or “both”. As used herein, “at least one of the following: ” and “at least one of ” and similar wording, where the list of two or more elements are joined by “and” or “or,” mean at least any one of the elements, or at least any two or more of the elements, or at least all the elements.

[0022] As used herein, the singular forms “a”, “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises”, “comprising”, “has”, “having”, “includes” and / or “including”, when used herein, specify the presence of stated features, elements, and / or components etc., but do not preclude the presence or addition of one or more other features, elements, components and / or combinations thereof.

[0023] Any flow diagram (such as FIGS. 1 or 2) or signaling diagram (such as FIG. 3) herein is considered to be a logic flow diagram, and illustrates the operation of an exemplary method, results of execution of computer program instructions embodied on a computer readable memory, functions performed by logic implemented in hardware, and / or interconnected means for performing functions in accordance with an exemplary embodiment. Block diagrams (such as FIG. 4) also illustrate the operation of an exemplary method, results of execution of computer program instructions embodied on a computer readable memory, functions performed by logic implemented in hardware, and / or interconnected means for performing functions in accordance with an exemplary embodiment.

[0024] Examples herein relate to a 3 GPP study item for AI / ML techniques in 3 GPP, see Qualcomm (Moderator), “Study on Artificial Intelligence (AI)ZMachine Learning (ML) for NR Air Interface” RP-213599, 3GPP TSGRAN Meeting #94e, Electronic Meeting, Dec. 6 - 17,2021. This study item has the following objectives, which are contained between opening and closing quotation marks:

[0025] “4. Objective

[0026] 4.1 Objective of SI or Core part WI or Testing part WI

[0027] Study the 3 GPP framework for AI / ML for air-interface corresponding to each target use case regarding aspects such as performance, complexity, and potential specification impact.

[0028] Use cases to focus on:

[0029] • Initial set of use cases includes:

[0030] o CSI feedback enhancement, e.g., overhead reduction, improved accuracy, prediction [RANI]

[0031] o Beam management, e.g., beam prediction in time, and / or spatial domain for overhead and latency reduction, beam selection accuracy improvement [RANI]

[0032] o Positioning accuracy enhancements for different scenarios including, e.g., those with heavy NLOS conditions [RANI]

[0033] • Finalize representative sub use cases for each use case for characterization and baseline performance evaluations by RAN#98

[0034] o The AI / ML approaches for the selected sub use cases need to be diverse enough to support various requirements on the gNB-UE collaboration levels

[0035] Note: the selection of use cases for this study solely targets the formulation of a framework to apply AI / ML to the air-interface for these and other use cases. The selection itself does not intend to provide any indication of the prospects of any future normative project.

[0036] AI / ML model, terminology and description to identify common and specific characteristics for framework investigations:

[0037] • Characterize the defining stages of AI / ML related algorithms and associated complexity:

[0038] o Model generation, e.g., model training (including input / output, pre- / post- process, online / offline as applicable), model validation, model testing, as applicable

[0039] o Inference operation, e.g., input / output, pre- / post-process, as applicable

[0040] • Identify various levels of collaboration between UE and gNB pertinent to the selected use cases, e.g.,

[0041] o No collaboration: implementation-based only AI / ML algorithms without information exchange [for comparison purposes]

[0042] o Various levels of UE / gNB collaboration targeting at separate or joint ML operation.

[0043] • Characterize lifecycle management of AI / ML model: e.g., model training, model deployment , model inference, model monitoring, model updating

[0044] • Dataset(s) for training, validation, testing, and inference

[0045] • Identify common notation and terminology for AI / ML related functions, procedures and interfaces

[0046] • Note: Consider the work done for FS_NR_ENDC_data_collect when appropriate

[0047] For the use cases under consideration:

[0048] 1) Evaluate performance benefits of AI / ML based algorithms for the agreed use cases in the final representative set:

[0049] o Methodology based on statistical models (from TR 38.901 and TR 38.857 [positioning]), for link and system level simulations.

[0050] ■ Extensions of 3 GPP evaluation methodology for better suitability toAI / ML based techniques should be considered as needed.

[0051] ■ Whether field data are optionally needed to further assess the performance and robustness in real-world environments should be discussed as part of the study.

[0052] ■ Need for common assumptions in dataset construction for training, validation and test for the selected use cases.

[0053] ■ Consider adequate model training strategy, collaboration levels and associated implications

[0054] ■ Consider agreed-upon base Al model(s) for calibration

[0055] ■ Al model description and training methodology used for evaluation should be reported for information and cross-checking purposes

[0056] o KPIs: Determine the common KPIs and corresponding requirements for the AI / ML operations. Determine the use-case specific KPIs and benchmarks of the selected usecases.

[0057] ■ Performance, inference latency and computational complexity of AI / ML based algorithms should be compared to that of a state-of-the-art baseline

[0058] ■ Overhead, power consumption (including computational), memory storage, and hardware requirements (including for given processing delays) associated with enabling respective AI / ML scheme, as well as generalization capability should be considered.

[0059] 2) Assess potential specification impact, specifically for the agreed use cases in the final representative set and for a common framework:

[0060] o PHY layer aspects, e g., (RANI)

[0061] ■ Consider aspects related to, e.g., the potential specification of the AlModel lifecycle management, and dataset construction for training, validation and test for the selected use cases

[0062] ■ Use case and collaboration level specific specification impact, such as new signaling, means for training and validation data assistance, assistance information, measurement, and feedback

[0063] o Protocol aspects, e.g., (RAN2) - RAN2 only starts the work after there is sufficient progress on the use case study in RANI

[0064] ■ Consider aspects related to, e.g., capability indication, configuration and control procedures (training / inference), and management of data and AI / ML model, per RANI input

[0065] ■ Collaboration level specific specification impact per use case

[0066] o Interoperability and testability aspects, e.g., (RAN4) - RAN4 only starts the work after there is sufficient progress on use case study in RANI and RAN2

[0067] ■ Requirements and testing frameworks to validate AI / ML based performance enhancements and ensuring that UE and gNB with AI / ML meet or exceed the existing minimum requirements if applicable

[0068] ■ Consider the need and implications for AI / ML processing capabilities definition

[0069] Note 1 : specific AI / ML models are not expected to be specified and are left to implementation. User data privacy needs to be preserved.

[0070] Note 2: The study on AI / ML for air interface is based on the current RAN architecture and new interfaces shall not be introduced.”

[0071] In addition to RP-213599, there are multiple agreements for 3GPP status on AI / ML These agreements include the following, and may be found in, e.g., Nokia, Nokia Shanghai Bell, “Other aspects on ML for beam management”, 3GPP TSG RAN WG1 #113, Incheon, Korea, May 22nd - May 26th, 2023, R1 -2304684.

[0072] Agreement

[0073] For BM-Casel with a UE-side AI / ML model, study the potential specification impact of LI signaling to report the following information of AI / ML model inference to NW

[0074] • The beam(s) that is based on the output of AI / ML model inference

[0075] • FFS: Predicted Ll-RSRP corresponding to the beam(s)

[0076] • FFS: other information

[0077] Agreement

[0078] For BM-Case2 with a UE-side AI / ML model, study the potential specification impact of LI signaling to report the following information of AI / ML model inference to NW

[0079] • The beam(s) of N future time instance(s) that is based on the output ofAI / ML model inference

[0080] o FFS: value of N

[0081] • FFS: Predicted Ll-RSRP corresponding to the beam(s)

[0082] • Information about the timestamp corresponding the reported beam(s)

[0083] o FFS: explicit or implicit

[0084] • FFS: other information

[0085] Agreement

[0086] For BM-Casel and BM-Case2 with a UE-side AI / ML model, study the following alternatives for model monitoring with potential down-selection:

[0087] • Altl. UE-side Model monitoring

[0088] o UE monitors the performance metric(s)

[0089] o UE makes decision(s) of model selection / activation / deactivation / switching / fallback operation

[0090] • Alt2. NW-side Model monitoring

[0091] o NW monitors the performance metric(s)

[0092] o NW makes decision(s) of model selection / activation / deactivation / switching / fallback operation

[0093] • Alt3. Hybrid model monitoring

[0094] o UE monitors the performance metric(s)

[0095] o NW makes decision(s) of model selection / activation / deactivation / switching / fallback operation

[0096] Agreement

[0097] Regarding NW-side model monitoring for a network-side AI / ML model of BM-Casel and BM-Case2, study the necessity and the potential specification impacts from the following aspects:

[0098] • UE reporting of beam measurement(s) based on a set of beams indicated by gNB

[0099] • Signaling, e.g., RRC-based, LI -based

[0100] • Note: Performance and UE complexity, power consumption should be considered

[0101] As can be seen by these agreements, there is active research in the AI / ML area. In this area, the UE may be configured with ML-based prediction using a network- deployed ML model. Since the model is trained by the functions of a wireless communication network (e.g., a model may be logically interfaced with RAN functions) or may be part of the RAN function, the required input parameters may be configured by the network (e.g., which input parameters are used for performing prediction such as beam index / Ll-RSRP, location, or the like).

[0102] The UE may be then provided with an ML model (e.g., by the RAN or ML functions associated with RAN / core network) and configured to perform prediction and provide the prediction output (which may also be referred to as a result) to the network.

[0103] However, although the NW has configured the input parameters, there may be further restrictions associated with the provided ML model and the input parameters. Incorrect or invalid input parameters, for example, may cause inaccurate or invalid predictions that may cause network to perform actions (such as control / data beam configuration) based on the invalid prediction and the result may not be optimal or may even cause link failures.

[0104] Embodiments herein address these and other problems. As an example, a UE that uses an ML model to perform prediction, such as beam prediction, may not performprediction using the ML model, e.g., based on whether certain condition(s) are not fulfilled. This is described in FIG. 1, which is used to provide an overview of the examples herein.

[0105] Referring to FIG. 1, this figure is a flowchart of a method where conditions are applied for input data for UE-sided prediction using an ML model. In block 110, a UE obtains an ML model from the network, or accesses an already-obtained ML mode, having input of parameters and output of one or more communication parameters. Herein, an ML model is a model that has been trained to recognize certain types of patterns for input parameters and to create a corresponding output of one or more communication parameters. One such example is for beam selection, based on input parameters. This model has been applied to a set of data during training, and implements an algorithm that used to reason over and learn from those data to produce an output. While the term “ML model” is mainly used herein, Al models or other models that are trained to recognize patterns and produce an output are considered to fall under the term “ML model”. It is further noted that the term “output” may encompass the following (where a beam selection example is used): a single output, such as an indication of a beam to select (as a communication parameter); multiple outputs, such as indication of a beam to select (as a communication parameter) and an indication of a likelihood this is the correct beam to select, or indications of multiple beams that could be selected (as multiple communication parameters) and corresponding likelihoods of these being the best beam for selection. As an example, ML model may refer to at least one or more of (but not limited to) neural network(s), convolutional neural network(s), recurrent neural network(s), deep reinforcement learning, transformer, convolutional-LSTM (where LSTM stands for long-short-term-memory), auto encoder-decoder LSTM and so on.

[0106] In block 120, the UE receives (e.g., further) configuration for performing ML- based prediction. The configuration comprises one or more input conditions related to input parameters or values of the input parameters for the ML model. In block 130, the UE determines whether input parameter(s) fulfill or do not fulfill the input condition(s) for a current prediction. This is referred to as a prediction because the models are based and previously trained on input parameters, and the output is based on those inputs, so that the output is predicted based on those previously examined input parameters. That is, an ML model is predicting that its output is correct based on its previous training.

[0107] In response to the input parameters fulfilling input condition(s) (reference 140), blocks 160 and 170 are performed. In block 160, the UE performs prediction using the ML model. The UE reports the prediction output to the network in block 170.

[0108] In response to the input parameters not fulfilling input condition(s) (reference 150), there are two possible alternative (Alt.) pathways 151 (Alt. 1) and 152 (Alt. 2). In block 180, for Alt. 1 151, the UE does not perform prediction using the ML model. Instead, the UE reports at least one indication indicated the criteria were not fulfilled and / or information on the criteria that were not fulfilled. This reporting occurs in block 190. In one embodiment, the UE may be configured to report (e.g., in a separate reporting instance) to the network which input parameter condition was not fulfilled for block 190. In this case, the UE may be configured not to include prediction (the information that was predicted as output) to the network.

[0109] Alternatively, in Alt. 2 152, the UE is configured to perform prediction (block185) and report (block 190) the predicted values with associated indication that the input parameter condition was not fulfilled. In one example, the UE is configured to still report (block186) but also indicate (block 190) that one or more input conditions were not valid. Another example is the UE performs prediction (block 185) but then performs reporting in block 190 without reporting the prediction output. See reference 187. This would, e.g., allow the network to query the UE later for the prediction, should the network determine that more information about the prediction result is useful. Alternative 2 152 allows the network to still receive (or later query for) a report, but also receive feedback that the report is based on the condition(s) not being fulfilled. This could be used, for instance, for additional training of the ML model or for other purposes.

[0110] Block 195 indicates possible techniques for reporting. That is, in one embodiment, any reporting may be performed using one of the following: a UCI (uplink control information) on PUSCH / PUCCH, MAC CE, or RRC. As a non-limiting example, the prediction performed with valid input parameters may be reported using PUSCHZPUCCH / MAC CE. In another non-limiting example, if non-valid input was used and prediction reported it may be reported using RRC or RRC and PUSCH / PUCCH / MAC CE. The RRC level reporting may be used for reporting the invalid parameter or parameters or the conditions that were considered invalid for the prediction.

[0111] Now that an overview has been provided, more details are provided.

[0112] Referring to FIG. 2, this figure is a block diagram showing a ML model and its input parameters and outputs, and also shows examples of input condition(s) that are related to input parameters or values of the input parameters for the ML model. The ML model 200, which may be a neural network, has a number of input parameters 210, in this case input parameters 210-1 through 210-N, N>1. The ML model 200 has been trained to produce an output 220 based on the input parameters 210. That output 220 may include the following as examples: 220-1, a communication parameter (CP), such as an indication of a beam; 220-2, a CP and a corresponding likelihood (e.g., a number indicating how likely this is the correct beam to select); 220-3, multiple CPs, e.g., with implied likelihood, e.g., Beam 1 and Beam 2 in that order, where Beam 1 has a higher likelihood of being correct than does Beam 2; or 220-4, multiple CPs and corresponding likelihoods.

[0113] It is noted that the terms “input parameters” and “input conditions” are merely used to describe corresponding elements used herein and other terms may be used. For instance, “input parameters” could be shortened to “inputs” or any other term indicating an input to an ML model 200. The “input conditions” could be referred to as input criteria or rules, as non-limiting examples, and correspond to any element that places a condition on an input parameter in order to determine whether or not prediction should be performed based on an input parameter meeting the condition.

[0114] The input conditions 230 are related to input parameters 210 or values of the input parameters 210 for the ML model 200. Multiple examples of the input conditions 230 are illustrated in FIG. 2 and described below.

[0115] In one embodiment, the UE is configured with at least one requirement as an input condition 230-1 for the input parameter(s) 210 for the UE ML-based prediction. If the requirement, as an input condition 230-1, is not fulfilled (as described in FIG. 1) for an input parameter 210, the UE is not allowed use the input parameter 210 for the prediction for the ML model, if the configuration (see block 120) indicates. See also reference 150, and Alt. 1 151 of FIG. 1.

[0116] Consider also input parameters and input conditions 230-2. In this example, there are N input parameters and UE has a requirement (as an input condition 230-2) for the current prediction one of the following:

[0117] a) all N input parameters for the current prediction, or

[0118] b) at least M < N input parameters; or

[0119] c) a specific parameter.

[0120] That is, if the UE does not meet, for the current prediction, the applicable (a), (b), or (c) requirement, the UE is not required to perform (or is not allowed to perform prediction that would be reported) the prediction (see Alt. 1 151 and block 180 in FIG. 1), or is configured not to report the prediction to the network, or is configured to report a value or indicator that prediction was not performed, or the reported prediction was performed using non-valid input values, or that not all the requirement(s) for input conditions were fulfilled (see, e.g., block 190 of FIG. 1). When the UE was not able to perform prediction (or was not required to perform) prediction, the UE may be configured to indicate such information to the network.

[0121] In one embodiment, the condition for performing beam prediction of a specific model may be a cell quality or a beam quality. See input condition 230-3 of FIG. 3. The following examples are possible under this embodiment. The quality may refer to, e.g., RSRP and / or SINK, (signal to interference and noise ratio) / RSRQ (reference signal received quality).

[0122] a) The model may require or is trained that the cell quality for which the prediction is associated is measured / observed by the UE, and is above a threshold level value (see input condition 230-3a).

[0123] b) At least X (e.g., of N) beam quality values have to be above a threshold value that are used as input for the model. See input condition 230-3b. In an example, X < N. The threshold value may be configured by the network or is associated with the ML model 200.

[0124] c) As an example, the prediction may be, e.g., LI -RSRP, or beam ID(s) (e.g., either beam quality or cell quality) / spatial domain (e.g., beam prediction) for which the UE may fail to receive measurement samples, e.g., the at least one measurement sample (for either cell quality or beam quality for beam prediction) has lower quality than an input threshold configured for the model. See input condition 230-3c. In further detail, the prediction could be LI -RSRP or beam ID(s) or top-K beam ID(s), which indicate beam quality. When the NW uses the predicted Ll-RSRP / beam ID(s) in a system throughput calculation, the NW could obtain cell quality.

[0125] d) As an example, the UE was not able to obtain at least one or one or more of required measurements due to reception of other signals and channels (e.g., system information). In this example (see input condition 230-3 d), the UE has to obtain all required measurements (as input parameters) in order for the input parameters to be fulfilled (see reference 140 of FIG. 1).

[0126] e) As a further example, the UE was not able to feed measurements on one or more (or all) the required Rx panels (or beams). For instance, the model may require measurements on at least A out of B antenna panels (or beams) at UE, (A <= B) fed to the network. This is illustrated as input condition 230-3e: Require measurements (as input parameters) on at least A out of B antenna panels at the UE. The value of A may be configured by the network or defined in specification. B may be the UE capability for performing measurements. The value of B may be provided to the network as UE capability. In some examples, the value of B may be configured by network and / or associated with the ML model so that at least B panels (or beams) have to be used to obtain measurements (that are fed to the network).

[0127] f) As an example, the UE performed other tasks and was not able to perform measurements, other task maybe, e.g., random access. This can be handled under input condition 230-3d.

[0128] In one particular example, an input parameter may be a location (e.g., positioning) information input for the UE and there may be a specific threshold for this. See input condition 230-4. The specific threshold could be the following.

[0129] a) if the UE determines that there is an uncertainty associated with the location information used as input for the model that is greater than an uncertainty threshold, see input condition 230-4a:

[0130] i) e.g., the location may be satellite-based, there may be a limit on a minimum number of satellites that were visible / used for the location information, see input condition 230- 4ai (that is, a condition is not fulfilled if there are less than the minimum number of satellites, as the uncertainty is higher with a lower number of satellites); or

[0131] ii) if the location uncertainty is more than Z meters (or above value Z), see input condition 230-4aii (that is, a condition is not fulfilled if the UE is more than Z meters, as the uncertainty is higher with more meters). This may be indicated by the (satellite) positioning software or the uncertainty information may be determined by the UE or provided by the positioning entity such as an LMF, location management function. Positioning accuracy may be enhanced, e.g., using additional positioning enhancement technologies such as wireless access point (Wi-Fi or 3 GPP positioning methods), and thus in some examples, the residual location uncertainty may not be the value given by the satellite positioning but a combination of multipletechniques. In some examples, the location uncertainty threshold (value Z) may be a configurable value (configured by network, e.g., using RRC). In another example, the location uncertainty threshold may be a value associated with the ML model. Example values could be 1 meter, 5 meters, 10 meters, or the like. In an example, the UE may be configured to determine that input condition for a location specific input parameter is valid when the location uncertainty is below (or equal to) the value Z. In some cases, the UE may be configured to perform the prediction when the location uncertainty is above the value Z. When the UE reports the prediction outcome with an uncertainty above Z, the UE may be configured to indicate that the reported outcome is performed. In some examples, the UE may be configured with a location uncertainty threshold value (e.g., configured by network and / or associated with the ML model) and while the location uncertainty is above the threshold value, the UE may consider that location specific input condition is not valid. In some cases, the UE is not required to predict and / or report the prediction outcome. Alternatively, while the location uncertainty is below the threshold value, the UE expected to perform the prediction or consider that the location specific input condition is valid.

[0132] Referring to FIG. 3, this figure is a signaling diagram and flowchart of applying conditions for input parameters for UE-sided prediction using an ML model. In this figure, a base station 70, which may also be a first network node, performs signaling with a UE 10, which may be a second network node. In block 310, the base station 70 sends the ML model 200 to the UE. As an example, what gets sent may be indications 315 that define the ML model 200. In block 320, the base station 70 sends (e.g., further) configuration for performing ML- based prediction (e.g., input condition(s)) to the UE, which is illustrated as a (e.g., further) configuration 325.

[0133] The UE 10 in block 321 applies conditions for input parameters using an ML model, as described in FIGS. 1 and 2. Blocks 326 and 340 indicate possible reporting the UE can perform. In block 326, the UE 10 reports prediction output (see also block 170 of FIG. 1) and signals indication of the prediction output 330. The base station 70 receives the indication and in block 335 performs action(s) based on the indicated prediction output, e.g., such as (block 336) using a beam indicated by the prediction output 330 for the UE.

[0134] As another possibility for reporting, the UE 10 in block 340 performs reporting as per block 190 of FIG. 1. Reporting 345 may include indication indicating thecondition(s) were not fulfilled and / or information on the condition(s) that were not fulfilled. The base station 70 receives the reporting 345 and in block 350 performs action(s) based on reporting where input parameter(s) did not fulfill input condition(s). Block 360 indicates possible actions, such as further training the ML model 200 or deploying alternative ML models 200 for the UE based on the reporting.

[0135] Turning to FIG. 4, this figure shows a block diagram of one possible and nonlimiting example of a cellular network 1 that is connected to a user equipment (UE) 10. Multiple network elements are shown in the cellular network of FIG. 4: a base station 70; and a core network 90.

[0136] In FIG. 4, a user equipment (UE) 10, e.g., a second network node, is in wireless communication via radio link 11 with the base station 70, e.g., a first network node, of the cellular network 1. A UE 10 is a wireless communication device, such as a mobile device, that is configured to access a cellular network. The UE 10 is illustrated with one or more antennas 28. The ellipses 2 indicate there could be multiple UEs 10 in wireless communication via radio links with the base station 70. The UE 10 includes one or more processors 13, one or more memories 15, and other circuitry 16. The other circuitry 16 includes one or more receivers (Rx(s)) 17 and one or more transmitters (Tx(s)) 18. A program 12 is used to cause the UE 10 to perform the operations described herein. The program 12 includes, has access to, or otherwise causes the ML model 200 to perform prediction based on input parameters 210. For a UE 10, the other circuitry 16 could include circuitry such as for user interface elements (not shown) like a display.

[0137] The base station 70, as, e.g., a second network node and a network element of the cellular network 1, provides the UE 10 access to cellular network 1 and to the data network 91 via the core network 90 (e.g., via a user plane function (UPF) of the core network 90). That is, the base station 70 is an access node providing access to the network 1 by the UE 10. The base station 70 is illustrated as having one or more antennas 58. In general, the base station 70 may be referred to as RAN node 70, although many will make reference to this as a gNB (gNode B, a base station for NR, new radio) instead. There are, however, many other examples of RAN nodes including an eNB (evolved Node B) or TRP (Transmission-Reception Point). The base station 70 includes one or more processors 73, one or more memories 75, and other circuitry 76. The other circuitry 76 includes one or more receivers (Rx(s)) 77 and one or more transmitters (Tx(s)) 78. Aprogram 72 is used to cause the base station 70 to perform the operations described herein. The program 72 also has access to a version of the ML model 200, which the program 200 can send to the UE 10 or update through additional training, as examples.

[0138] It is noted that the examples herein are not limited to particular wireless technology, although the examples use cellular technology. That is, the base station 70 may instead be implemented via other wireless technologies, such as Wi-Fi (a wireless networking protocol that devices use to communicate without direct cable connections). In the case of Wi-Fi, the link 11 could be characterized as a wireless link. For Wi-Fi, the base station 70 is generally referred to as an access point, and the UE 10 may be referred to as a wireless (e.g., mobile) device, and the network 1 is a wireless network. The core network 90 might be an Internet service provider or a wireless wide area network, and the data network 91 might be the Internet, as examples.

[0139] Two or more base stations 70 communicate using, e.g., link(s) 79. The link(s) 79 may be wired or wireless or both and may implement, e.g., an Xn interface for 5G (fifth generation), an X2 interface for LTE (Long Term Evolution), or other suitable interface for other standards.

[0140] The cellular network 1 may include a core network 90, as a third illustrated element or elements, that may include core network functionality, and which provide connectivity via a link or links 81 with a data network 91, such as a telephone network and / or a data communications network (e.g., the Internet). The core network 90 includes one or more processors 93, one or more memories 95, and other circuitry 96. The other circuitry 96 includes one or more receivers (Rx(s)) 97 and one or more transmitters (Tx(s)) 98. A program 92 is used to cause the core network 90 to perform the operations described herein.

[0141] The core network 90 could be a 5GC (5G core network). The core network 90 can implement or comprise multiple network functions (NF(s)) 99, and the program 92 may comprise one or more of the NFs 99. A 5G core network may use hardware such as memory and processors and a virtualization layer. It could be a single standalone computing system, a distributed computing system, or a cloud computing system. The NFs 99, as network elements, of the core network could be containers or virtual machines running on the hardware of the computing system(s) making up the core network 90.

[0142] Core network functionality for 5G may include access and mobility management functionality that is provided by a network function 99 such as an access and mobility management function (AMF(s)), session management functionality that is provided by a network function such as a session management function (SMF). Core network functionality for access and mobility management in an LTE (Long Term Evolution) network may be provided by an MME (Mobility Management Entity) and / or SGW (Serving Gateway) functionality, which routes data to the data network. Many others are possible, as illustrated by the examples in FIG. 4: AMF; SMF; MME; SGW; GMLC (Gateway Mobile Location Center); LMFs (Location Management Functions); UDM (Unified Data Management) / UDR (Unified Data Repository); NRF (Network Repository Function); and / or E-SMLC (Evolved Serving Mobile Location Center). These are merely exemplary core network functionality that may be provided by the core network 90, and note that both 5G and LTE core network functionality might be provided by the core network 90. The RAN node 70 is coupled via a backhaul link 31 to the core network 90. The RAN node 70 and the core network 90 may include an NG (Next Generation) interface for 5G, or an SI interface for LTE, or other suitable interface for other radio access technologies for communicating via the backhaul link 31.

[0143] Any methods presented herein are applicable to any (wireless) technology using ML, also any technologies such as 6G and beyond.

[0144] In the data network 91, there is a computer-readable medium 94. The computer-readable medium 94 contains instructions that, when downloaded and installed into the memories 15, 75, or 95 of the corresponding UE 10, base station 70, and / or core network element(s) 90, and executed by processor(s) 13, 73, or 93, cause the respective device to perform corresponding actions described herein. The computer-readable medium 94 may be implemented in other forms, such as via a compact disc or memory stick.

[0145] The programs 12, 72, and 92 contain instructions stored by corresponding one or more memories 15, 75, or 95. These instructions, when executed by the corresponding one or more processors 13, 73, or 93, cause the corresponding apparatus 10, 70, or 90, to perform the operations described herein. The computer readable memories 15, 75, or 95 may be of any type suitable to the local technical environment and may be implemented using any suitable data storage technology, such as semiconductor-based memory devices, flash memory, firmware, magnetic memory devices and systems, optical memory devices and systems, fixed memory andremovable memory. The computer readable memories 15, 75, and 95 may be means for performing storage functions. The one or more processors 13, 73, and 93, may be of any type suitable to the local technical environment, and may include one or more of general-purpose computers, special purpose computers, microprocessors, digital signal processors (DSPs) and processors based on a multi-core processor architecture, as non-limiting examples. The processors 13, 73, and 93 may be means for causing their respective apparatus to perform functions, such as those described herein. In particular, means for performing the functions herein may comprise the one or more processors 13, 73, and 93 and the corresponding one or more memories 15, 75, or 95, which individually store instructions that, when executed by the corresponding processor, cause performance of the corresponding apparatus.

[0146] The receivers 17, 77, and 97, and the transmitters 18, 78, and 98 may implement wired or wireless interfaces. The receivers and transmitters may be grouped together as transceivers.

[0147] The cellular network 1 may implement network virtualization, which is the process of combining hardware and software network resources and network functionality into a single, software-based administrative entity, a virtual network. Network virtualization involves platform virtualization, often combined with resource virtualization. Network virtualization is categorized as either external, combining many networks, or parts of networks, into a virtual unit, or internal, providing network-like functionality to software containers on a single system. Note that the virtualized entities (such as network functions 99) that result from the network virtualization are still implemented, at some level, using hardware such as processors 73 and / or 93 and memories 75 and / or 95, and also such virtualized entities create technical effects.

[0148] In general, the various embodiments of the user equipment 10 can include, but are not limited to, cellular telephones (such as smart phones, mobile phones, cellular phones, voice over Internet Protocol (IP) (VoIP) phones, and / or wireless local loop phones), tablets, portable computers, vehicles or vehicle-mounted devices for, e.g., wireless V2X (vehicle-to- everything) communication, image capture devices such as digital cameras, gaming devices, music storage and playback appliances, Internet appliances (including Internet of Things, loT, devices), loT devices with sensors and / or actuators for, e.g., automation applications, as well as portable units or terminals that incorporate combinations of such functions, laptop-embedded equipment (LEE), laptop-mounted equipment (LME), Universal Serial Bus (USB) dongles,smart devices, wireless customer-premises equipment (CPE), an Internet of Things (loT) device, a watch or other wearable, a head-mounted display (HMD), a vehicle, a drone, a medical device and applications (e.g., remote surgery), an industrial device and applications (e.g., a robot and / or other wireless devices operating in an industrial and / or an automated processing chain contexts), a consumer electronics device, a device operating on commercial and / or industrial wireless networks, and the like. That is, the UE 10 could be any end device that may be capable of wireless communication. By way of example rather than limitation, the UE may also be referred to as a communication device, terminal device (MT), a Subscriber Station (SS), a Portable Subscriber Station, a Mobile Station (MS), or an Access Terminal (AT).

[0149] Without in any way limiting the scope, interpretation, or application of the claims appearing below, a technical effect and / or advantage of one or more of the example embodiments disclosed herein is the prediction quality can be trusted by the network (e.g., for a beam management decision) when the UE performs ML based prediction in a communication network. Another technical effect and / or advantage of one or more of the example embodiments disclosed herein is the techniques provide further input for the network to determine why the prediction was not successful, e.g., it can be used to further train the model or deploy alternative models for the UE based on the reporting.

[0150] The following are additional examples.

[0151] Example 1. A method, comprising: receiving, from a first network node by a second network node, configuration for performing prediction by a machine-learning model, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; determining, by the second network node, whether input parameters for the machine-learning model do or do not fulfill the one or more input conditions; and performing, by the second network node responsive to the determining, one or both of prediction using the machine-learning model or reporting to the first network node of information based on whether or not the prediction was performed.

[0152] Example 2. The method according to example 1, wherein: the determining comprises determining that the input parameters for the machine-learning model fulfill the one or more input conditions; and the performing comprises: performing the prediction using the machine-learning model based on the determining that the input parameters for the machine-learning model fulfill the one or more input conditions; and reporting indication of output of the machine-learning model based on performing the prediction.

[0153] Example 3. The method according to example 1, wherein: the determining comprises determining that the input parameters for the machine-learning model do not fulfill the one or more input conditions; and the performing comprises: not performing the prediction using the machine-learning model based on the determining that the input parameters for the machinelearning model do not fulfill the one or more input conditions; and reporting at least one indication indicating one or both of the one or more input conditions were not fulfilled or information on the one or more input conditions that were not fulfilled.

[0154] Example 4. The method according to example 1, wherein: the determining comprises determining that the input parameters for the machine-learning model do not fulfill the one or more input conditions; and the performing comprises: performing the prediction using the machine-learning model based on the determining that the input parameters for the machinelearning model do not fulfill the one or more input conditions; reporting indication of output of the machine-learning model based on performing the prediction; and reporting at least one indication indicating one or both of the one or more input conditions were not fulfilled or information on the one or more input conditions that were not fulfilled.

[0155] Example 5. The method according to any of examples 1 to 4, wherein the one or more input conditions comprise at least one requirement for the input parameters for machinelearning-based prediction by the second network node, and the determining comprises determining whether input parameters for the machine-learning model do or do not fulfill the at least one requirement.

[0156] Example 6. The method according to any of examples 1 to 4, wherein: there are N input parameters and the second network node has, for a current prediction, a requirement as an input condition of one the following: a) all N input parameters for the current prediction, or b) at least M < N input parameters; or c) a specific input parameter; and the determining comprises determining whether input parameters for the machine-learning model do or do not fulfill the requirement.

[0157] Example 7. The method according to any of examples 1 to 4, wherein the one or more input conditions comprise one or both of a cell quality or a beam quality, and thedetermining comprises determining whether input parameters for the machine-learning model do or do not fulfill the one or both of the cell quality or the beam quality.

[0158] Example 8. The method according to example 7, wherein a selected quality of the cell quality or the beam quality is fulfilled if the selected quality is above a threshold level value.

[0159] Example 9. The method according to example 7, wherein at least X of N beam quality values that are used as input parameters have to be above a threshold value for the machine-learning model in order to fulfill beam quality as an input condition.

[0160] Example 10. The method according to example 7, wherein at least one measurement sample for either the cell quality or the beam quality has a lower quality than an input threshold configured for the machine-learning model in order to not fulfill beam quality as an input condition.

[0161] Example 11. The method according to example 7, wherein the second network node has to obtain all required measurements as input parameters in order for the one or both of the cell quality or the beam quality to be fulfilled.

[0162] Example 12. The method according to example 7, wherein the second network node is required to take measurements as input parameters on at least A out of B, A <= B, antenna panels at the second network node or the input parameters do or do not fulfill the one or both of the cell quality or the beam quality.

[0163] Example 13. The method according to any of examples 1 to 4, wherein the one or more input conditions comprise location information input for the second network node.

[0164] Example 14. The method according to example 13, wherein uncertainty associated with location information used as an input parameter for the machine-learning model has to be greater than an uncertainty threshold in order for the input parameters for the machinelearning model to not fulfill the uncertainty as an input condition.

[0165] Example 15. The method according to example 14, wherein the uncertainty comprises a limit on a minimum number of satellites that were visible or used for the location information, Example 16. The method according to example 14, wherein the uncertainty comprises a location uncertainty of more than Z meters.

[0166] Example 17. A method, comprising: sending, from a first network node to a second network node, configuration for performing prediction by a machine-learning model atthe second network node, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; and receiving, by the first network node from the second network node, one or both of output of a prediction performed using the machine-learning model on the second network node or reporting from the second network node of information based on whether or not the prediction was performed.

[0167] Example 18. The method according to any of examples 1 to 17, wherein the first network node comprises an access node providing access by the second network node to a wireless network, and the second network node comprises a wireless device able to connect to the wireless network via the access node.

[0168] Example 19. The method according to any of examples 1 to 17, wherein the first network node comprises a base station providing access by the second network node to a cellular network, and the second network node comprises a user equipment able to connect to the cellular network via the base station.

[0169] Example 20. A computer program, comprising instructions for performing the methods of any of examples 1 to 19, when the computer program is run on an apparatus.

[0170] Example 21. The computer program according to example 20, wherein the computer program is a computer program product comprising a computer-readable medium bearing instructions embodied therein for use with the apparatus.

[0171] Example 22. The computer program according to example 20, wherein the computer program is directly loadable into an internal memory of the apparatus.

[0172] Example 23. An apparatus comprising means for performing: receiving, from a first network node by a second network node, configuration for performing prediction by a machine-learning model, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; determining, by the second network node, whether input parameters for the machine-learning model do or do not fulfill the one or more input conditions; and performing, by the second network node responsive to the determining, one or both of prediction using the machine-learning model or reporting to the first network node of information based on whether or not the prediction was performed.

[0173] Example 24. The apparatus according to example 23, wherein: the determining comprises determining that the input parameters for the machine-learning model fulfill the one or more input conditions; and the performing comprises: performing the prediction using the machine-learning model based on the determining that the input parameters for the machine-learning model fulfill the one or more input conditions; and reporting indication of output of the machine-learning model based on performing the prediction.

[0174] Example 25. The apparatus according to example 23, wherein: the determining comprises determining that the input parameters for the machine-learning model do not fulfill the one or more input conditions; and the performing comprises: not performing the prediction using the machine-learning model based on the determining that the input parameters for the machine-learning model do not fulfill the one or more input conditions; and reporting at least one indication indicating one or both of the one or more input conditions were not fulfilled or information on the one or more input conditions that were not fulfilled.

[0175] Example 26. The apparatus according to example 23, wherein: the determining comprises determining that the input parameters for the machine-learning model do not fulfill the one or more input conditions; and the performing comprises: performing the prediction using the machine-learning model based on the determining that the input parameters for the machine-learning model do not fulfill the one or more input conditions; reporting indication of output of the machine-learning model based on performing the prediction; and reporting at least one indication indicating one or both of the one or more input conditions were not fulfilled or information on the one or more input conditions that were not fulfilled.

[0176] Example 27. The apparatus according to any of examples 23 to 26, wherein the one or more input conditions comprise at least one requirement for the input parameters for machine-learning-based prediction by the second network node, and the determining comprises determining whether input parameters for the machine-learning model do or do not fulfill the at least one requirement.

[0177] Example 28. The apparatus according to any of examples 23 to 26, wherein: there are N input parameters and the second network node has, for a current prediction, a requirement as an input condition of one the following: a) all N input parameters for the current prediction, or b) at least M < N input parameters; or c) a specific input parameter; and thedetermining comprises determining whether input parameters for the machine-learning model do or do not fulfill the requirement.

[0178] Example 29. The apparatus according to any of examples 23 to 26, wherein the one or more input conditions comprise one or both of a cell quality or a beam quality, and the determining comprises determining whether input parameters for the machine-learning model do or do not fulfill the one or both of the cell quality or the beam quality.

[0179] Example 30. The apparatus according to example 29, wherein a selected quality of the cell quality or the beam quality is fulfilled if the selected quality is above a threshold level value.

[0180] Example 31. The apparatus according to example 29, wherein at least X of N beam quality values that are used as input parameters have to be above a threshold value for the machine-learning model in order to fulfill beam quality as an input condition.

[0181] Example 32. The apparatus according to example 29, wherein at least one measurement sample for either the cell quality or the beam quality has a lower quality than an input threshold configured for the machine-learning model in order to not fulfill beam quality as an input condition.

[0182] Example 33. The apparatus according to example 29, wherein the second network node has to obtain all required measurements as input parameters in order for the one or both of the cell quality or the beam quality to be fulfilled.

[0183] Example 34. The apparatus according to example 29, wherein the second network node is required to take measurements as input parameters on at least A out of B, A <= B, antenna panels at the second network node or the input parameters do or do not fulfill the one or both of the cell quality or the beam quality.

[0184] Example 35. The apparatus according to any of examples 23 to 26, wherein the one or more input conditions comprise location information input for the second network node.

[0185] Example 36. The apparatus according to example 35, wherein uncertainty associated with location information used as an input parameter for the machine-learning model has to be greater than an uncertainty threshold in order for the input parameters for the machinelearning model to not fulfill the uncertainty as an input condition.

[0186] Example 37. The apparatus according to example 36, wherein the uncertainty comprises a limit on a minimum number of satellites that were visible or used for the location information, Example 38. The apparatus according to example 36, wherein the uncertainty comprises a location uncertainty of more than Z meters.

[0187] Example 39. An apparatus comprising means for performing: sending, from a first network node to a second network node, configuration for performing prediction by a machine-learning model at the second network node, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machinelearning model; and receiving, by the first network node from the second network node, one or both of output of a prediction performed using the machine-learning model on the second network node or reporting from the second network node of information based on whether or not the prediction was performed.

[0188] Example 40. The apparatus according to any of examples 23 to 39, wherein the first network node comprises an access node providing access by the second network node to a wireless network, and the second network node comprises a wireless device able to connect to the wireless network via the access node.

[0189] Example 41. The apparatus according to any of examples 23 to 39, wherein the first network node comprises a base station providing access by the second network node to a cellular network, and the second network node comprises a user equipment able to connect to the cellular network via the base station.

[0190] Example 42. The apparatus of any preceding apparatus example, wherein the means comprises: at least one processor; and at least one memory storing instructions that, when executed by at least one processor, cause the performance of the apparatus.

[0191] Example 43. An apparatus, comprising: one or more processors; and one or more memories storing instructions that, when executed by the one or more processors, cause the apparatus at least to perform: receiving, from a first network node by a second network node, configuration for performing prediction by a machine-learning model, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; determining, by the second network node, whether input parameters for the machine-learning model do or do not fulfill the one or more input conditions; and performing, by the second network node responsive to the determining, one orboth of prediction using the machine-learning model or reporting to the first network node of information based on whether or not the prediction was performed.

[0192] Example 44. The apparatus according to example 43, wherein: the determining comprises determining that the input parameters for the machine-learning model fulfill the one or more input conditions; and the performing comprises: performing the prediction using the machine-learning model based on the determining that the input parameters for the machine-learning model fulfill the one or more input conditions; and reporting indication of output of the machine-learning model based on performing the prediction.

[0193] Example 45. The apparatus according to example 43, wherein: the determining comprises determining that the input parameters for the machine-learning model do not fulfill the one or more input conditions; and the performing comprises: not performing the prediction using the machine-learning model based on the determining that the input parameters for the machine-learning model do not fulfill the one or more input conditions; and reporting at least one indication indicating one or both of the one or more input conditions were not fulfilled or information on the one or more input conditions that were not fulfilled.

[0194] Example 46. The apparatus according to example 43, wherein: the determining comprises determining that the input parameters for the machine-learning model do not fulfill the one or more input conditions; and the performing comprises: performing the prediction using the machine-learning model based on the determining that the input parameters for the machine-learning model do not fulfill the one or more input conditions; reporting indication of output of the machine-learning model based on performing the prediction; and reporting at least one indication indicating one or both of the one or more input conditions were not fulfilled or information on the one or more input conditions that were not fulfilled.

[0195] Example 47. The apparatus according to any of examples 43 to 46, wherein the one or more input conditions comprise at least one requirement for the input parameters for machine-learning-based prediction by the second network node, and the determining comprises determining whether input parameters for the machine-learning model do or do not fulfill the at least one requirement.

[0196] Example 48. The apparatus according to any of examples 43 to 46, wherein: there are N input parameters and the second network node has, for a current prediction, a requirement as an input condition of one the following: a) all N input parameters for the currentprediction, or b) at least M < N input parameters; or c) a specific input parameter; and the determining comprises determining whether input parameters for the machine-learning model do or do not fulfill the requirement.

[0197] Example 49. The apparatus according to any of examples 43 to 46, wherein the one or more input conditions comprise one or both of a cell quality or a beam quality, and the determining comprises determining whether input parameters for the machine-learning model do or do not fulfill the one or both of the cell quality or the beam quality.

[0198] Example 50. The apparatus according to example 49, wherein a selected quality of the cell quality or the beam quality is fulfilled if the selected quality is above a threshold level value.

[0199] Example 51. The apparatus according to example 49, wherein at least X of N beam quality values that are used as input parameters have to be above a threshold value for the machine-learning model in order to fulfill beam quality as an input condition.

[0200] Example 52. The apparatus according to example 49, wherein at least one measurement sample for either the cell quality or the beam quality has a lower quality than an input threshold configured for the machine-learning model in order to not fulfill beam quality as an input condition.

[0201] Example 53. The apparatus according to example 49, wherein the second network node has to obtain all required measurements as input parameters in order for the one or both of the cell quality or the beam quality to be fulfilled.

[0202] Example 54. The apparatus according to example 49, wherein the second network node is required to take measurements as input parameters on at least A out of B, A <= B, antenna panels at the second network node or the input parameters do or do not fulfill the one or both of the cell quality or the beam quality.

[0203] Example 55. The apparatus according to any of examples 43 to 46, wherein the one or more input conditions comprise location information input for the second network node.

[0204] Example 56. The apparatus according to example 55, wherein uncertainty associated with location information used as an input parameter for the machine-learning model has to be greater than an uncertainty threshold in order for the input parameters for the machinelearning model to not fulfill the uncertainty as an input condition.

[0205] Example 57. The apparatus according to example 56, wherein the uncertainty comprises a limit on a minimum number of satellites that were visible or used for the location information, Example 58. The apparatus according to example 56, wherein the uncertainty comprises a location uncertainty of more than Z meters.

[0206] Example 59. An apparatus, comprising: one or more processors; and one or more memories storing instructions that, when executed by the one or more processors, cause the apparatus at least to perform: sending, from a first network node to a second network node, configuration for performing prediction by a machine-learning model at the second network node, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; and receiving, by the first network node from the second network node, one or both of output of a prediction performed using the machine-learning model on the second network node or reporting from the second network node of information based on whether or not the prediction was performed.

[0207] Example 60. The apparatus according to any of examples 43 to 59, wherein the first network node comprises an access node providing access by the second network node to a wireless network, and the second network node comprises a wireless device able to connect to the wireless network via the access node.

[0208] Example 61. The apparatus according to any of examples 43 to 59, wherein the first network node comprises a base station providing access by the second network node to a cellular network, and the second network node comprises a user equipment able to connect to the cellular network via the base station.

[0209] As used in this application, the term “circuitry” may refer to one or more or all of the following:

[0210] (a) hardware-only circuit implementations (such as implementations in only analog and / or digital circuitry) and

[0211] (b) combinations of hardware circuits and software, such as (as applicable): (i) a combination of analog and / or digital hardware circuit(s) with software / firmware and (ii) any portions of hardware processor(s) with software (including digital signal processor(s)), software, and memory(ies) that work together to cause an apparatus, such as a mobile phone or server, to perform various functions) and

[0212] (c) hardware circuit(s) and or processor(s), such as a microprocessor(s) or a portion of a microprocessor(s), that requires software (e.g., firmware) for operation, but the software may not be present when it is not needed for operation.

[0213] This definition of circuitry applies to all uses of this term in this application, including in any claims. As a further example, as used in this application, the term circuitry also covers an implementation of merely a hardware circuit or processor (or multiple processors) or portion of a hardware circuit or processor and its (or their) accompanying software and / or firmware. The term circuitry also covers, for example and if applicable to the particular claim element, a baseband integrated circuit or processor integrated circuit for a mobile device or a similar integrated circuit in server, a cellular network device, or other computing or network device.

[0214] Embodiments herein may be implemented in software (executed by one or more processors), hardware (e.g., an application specific integrated circuit), or a combination of software and hardware. In an example embodiment, the software (e.g., application logic, an instruction set) is maintained on any one of various conventional computer-readable media. In the context of this document, a “computer-readable medium” may be any media or means that can contain, store, communicate, propagate or transport the instructions for use by or in connection with an instruction execution system, apparatus, or device, such as a computer, with one example of a computer described and depicted, e.g., in FIG. 4. A computer-readable medium may comprise a computer-readable storage medium (e.g., memories 15, 75, and 95 or other device) that may be any media or means that can contain, store, and / or transport the instructions for use by or in connection with an instruction execution system, apparatus, or device, such as a computer. A computer-readable storage medium does not comprise propagating signals, and therefore may be considered to be non-transitory. The term “non-transitory”, as used herein, is a limitation of the medium itself (i.e., tangible, not a signal) as opposed to a limitation on data storage persistency (e.g., RAM, random access memory, versus ROM, readonly memory).

[0215] If desired, the different functions discussed herein may be performed in a different order and / or concurrently with each other. Furthermore, if desired, one or more of the above-described functions may be optional or may be combined.

[0216] Although various aspects of the invention are set out in the independent claims, other aspects of the invention comprise other combinations of features from the described embodiments and / or the dependent claims with the features of the independent claims, and not solely the combinations explicitly set out in the claims.

[0217] It is also noted herein that while the above describes example embodiments of the invention, these descriptions should not be viewed in a limiting sense. Rather, there are several variations and modifications which may be made without departing from the scope of the present invention as defined in the appended claims.

[0218] The following abbreviations that may be found in the specification and / or the drawing figures are defined as follows:

[0219] 3GPP third generation partnership project

[0220] 5G fifth generation

[0221] Al artificial intelligence

[0222] Alt alternative

[0223] AMF access and mobility management function

[0224] BM beam management

[0225] CP communication parameter

[0226] E-SMLC evolved serving mobile location center

[0227] GMLC Gateway Mobile Location Center

[0228] eNB (or eNodeB) evolved Node B (e.g., an LTE base station)

[0229] FFS for future study

[0230] gNB (or gNodeB) base station for 5G / NR

[0231] I / F interface

[0232] KPI key performance indicator

[0233] LI layer 1

[0234] LMF Location Management Function

[0235] LTE long term evolution

[0236] MAC Medium Access Control

[0237] ML machine language

[0238] MME mobility management entity

[0239] NF network function

[0240] ng or NG next generation

[0241] NR new radio

[0242] NRF Network Repository Function

[0243] N / W or NW network

[0244] PUCCH physical uplink control channel

[0245] PUSCH physical uplink shared channel

[0246] RAN radio access network

[0247] RRC radio-resource control

[0248] RSRP Reference Signal Received Power

[0249] Rx receiver

[0250] SGW serving gateway

[0251] SMF session management function

[0252] TRP transmission-reception point

[0253] Tx transmitter

[0254] UCI uplink control information

[0255] UDM unified data management

[0256] UDR unified data repository

[0257] UE user equipment (e.g., a wireless, typically mobile device)

[0258] UPF user plane function

Claims

What is claimed is:

1. A method, comprising: receiving, from a first network node by a second network node, configuration for performing prediction by a machine-learning model, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; determining, by the second network node, whether input parameters for the machinelearning model do or do not fulfill the one or more input conditions; and performing, by the second network node responsive to the determining, one or both of prediction using the machine-learning model or reporting to the first network node of information based on whether or not the prediction was performed.

2. An apparatus comprising means for performing: receiving, from a first network node by a second network node, configuration for performing prediction by a machine-learning model, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; determining, by the second network node, whether input parameters for the machinelearning model do or do not fulfill the one or more input conditions; and performing, by the second network node responsive to the determining, one or both of prediction using the machine-learning model or reporting to the first network node of information based on whether or not the prediction was performed.

3. The apparatus according to claim 2, wherein: the determining comprises determining that the input parameters for the machine-learning model fulfill the one or more input conditions; and the performing comprises:performing the prediction using the machine-learning model based on the determining that the input parameters for the machine-learning model fulfill the one or more input conditions; and reporting indication of output of the machine-learning model based on performing the prediction.

4. The apparatus according to claim 2, wherein: the determining comprises determining that the input parameters for the machine-learning model do not fulfill the one or more input conditions; and the performing comprises: not performing the prediction using the machine-learning model based on the determining that the input parameters for the machine-learning model do not fulfill the one or more input conditions; and reporting at least one indication indicating one or both of the one or more input conditions were not fulfilled or information on the one or more input conditions that were not fulfilled.

5. The apparatus according to claim 2, wherein: the determining comprises determining that the input parameters for the machine-learning model do not fulfill the one or more input conditions; and the performing comprises: performing the prediction using the machine-learning model based on the determining that the input parameters for the machine-learning model do not fulfill the one or more input conditions; reporting indication of output of the machine-learning model based on performing the prediction; and reporting at least one indication indicating one or both of the one or more input conditions were not fulfilled or information on the one or more input conditions that were not fulfilled.

6. The apparatus according to any of claims 2 to 5, wherein the one or more input conditions comprise at least one requirement for the input parameters for machinelearning-based prediction by the second network node, and the determining comprises determining whether input parameters for the machine-learning model do or do not fulfill the at least one requirement.

7. The apparatus according to any of claims 2 to 5, wherein: there are N input parameters and the second network node has, for a current prediction, a requirement as an input condition of one the following: a) all N input parameters for the current prediction, or b) at least M < N input parameters; or c) a specific input parameter; and the determining comprises determining whether input parameters for the machinelearning model do or do not fulfill the requirement.

8. The apparatus according to any of claims 2 to 5, wherein the one or more input conditions comprise one or both of a cell quality or a beam quality, and the determining comprises determining whether input parameters for the machine-learning model do or do not fulfill the one or both of the cell quality or the beam quality, and wherein: a selected quality of the cell quality or the beam quality is fulfilled if the selected quality is above a threshold level value; or at least X of N beam quality values that are used as input parameters have to be above a threshold value for the machine-learning model in order to fulfill beam quality as an input condition; or at least one measurement sample for either the cell quality or the beam quality has a lower quality than an input threshold configured for the machine-learning model in order to not fulfill beam quality as an input condition; or the second network node has to obtain all required measurements as input parameters in order for the one or both of the cell quality or the beam quality to be fulfilled; or the second network node is required to take measurements as input parameters on at least A out of B, A <= B, antenna panels at the second network node or the inputparameters do or do not fulfill the one or both of the cell quality or the beam quality.

9. The apparatus according to any of claims 2 to 5, wherein the one or more input conditions comprise location information input for the second network node.

10. The apparatus according to claim 9, wherein uncertainty associated with location information used as an input parameter for the machine-learning model has to be greater than an uncertainty threshold in order for the input parameters for the machine-learning model to not fulfill the uncertainty as an input condition.

11. The apparatus according to claim 10, wherein the uncertainty comprises a limit on a minimum number of satellites that were visible or used for the location information,12. The apparatus according to claim 10, wherein the uncertainty comprises a location uncertainty of more than Z meters.

13. An apparatus comprising means for performing: sending, from a first network node to a second network node, configuration for performing prediction by a machine-learning model at the second network node, the configuration comprising one or more input conditions related to input parameters or values of the input parameters for the machine-learning model; and receiving, by the first network node from the second network node, one or both of output of a prediction performed using the machine-learning model on the second network node or reporting from the second network node of information based on whether or not the prediction was performed.

14. The apparatus according to any of claims 2 to 13, wherein the first network node comprises an access node providing access by the second network node to a wirelessnetwork, and the second network node comprises a wireless device able to connect to the wireless network via the access node.

15. The apparatus according to any of claims 2 to 13, wherein the first network node comprises a base station providing access by the second network node to a cellular network, and the second network node comprises a user equipment able to connect to the cellular network via the base station.

Images