[go: up one dir, main page]

WO2024169794A1 - Procédé de supervision de modèle et dispositif de communication - Google Patents

Procédé de supervision de modèle et dispositif de communication Download PDF

Info

Publication number
WO2024169794A1
WO2024169794A1 PCT/CN2024/076407 CN2024076407W WO2024169794A1 WO 2024169794 A1 WO2024169794 A1 WO 2024169794A1 CN 2024076407 W CN2024076407 W CN 2024076407W WO 2024169794 A1 WO2024169794 A1 WO 2024169794A1
Authority
WO
WIPO (PCT)
Prior art keywords
model
communication device
information
validity
input
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2024/076407
Other languages
English (en)
Chinese (zh)
Inventor
贾承璐
邬华明
潘翔
王园园
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Vivo Mobile Communication Co Ltd
Original Assignee
Vivo Mobile Communication Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Vivo Mobile Communication Co Ltd filed Critical Vivo Mobile Communication Co Ltd
Publication of WO2024169794A1 publication Critical patent/WO2024169794A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/09—Supervised learning
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information

Definitions

  • the present application belongs to the field of communication technology, and specifically relates to a model supervision method and communication equipment.
  • AI models are usually introduced to perform some reasoning tasks to improve network throughput, latency, and user capacity.
  • positioning models can be used to predict the location information of terminals
  • channel measurement models can be used for channel estimation.
  • model when using a model to perform a task, the model is deployed in the actual environment, and then relevant information in the current environment is collected as model input and model output.
  • the actual environment is usually dynamically changing, so the model may not be applicable (i.e., model failure), resulting in task execution failure.
  • model failure i.e., model failure
  • the embodiments of the present application provide a model supervision method and a communication device, which can solve the problem that the model is not applicable or fails when the model is deployed in an actual environment to perform tasks.
  • a model supervision method which is performed by a first communication device, and the method includes:
  • the first communication device receives model information of a first model from the second communication device; the first communication device determines the first model and at least one second model associated with the first model according to the model information of the first model; the first communication device supervises the validity of the at least one second model according to the first model; or,
  • the first communication device sends first information to the second communication device, where the first information is related to the first model or the at least one second model, and the first information is used by the second communication device to supervise the validity of the at least one second model based on the first model.
  • a model supervision method which is performed by a second communication device, and the method includes:
  • the second communication device sends model information of a first model to the first communication device, where the model information of the first model is used to determine the first model and at least one second model associated with the first model, where the first model is used to supervise the validity of the at least one second model; or,
  • the second communication device receives first information from the first communication device, where the first information is related to the first model or the at least one second model; the second communication device monitors the validity of the at least one second model based on the first information and the first model.
  • a model supervision device comprising:
  • a receiving module configured to receive model information of a first model from a second communication device; a determining module, configured to determine the first model and at least one second model associated with the first model according to the model information of the first model; a monitoring module, configured to monitor the validity of the at least one second model according to the first model; or,
  • a sending module is used to send first information to the second communication device, where the first information is related to the first model or the at least one second model, and the first information is used by the second communication device to monitor the validity of the at least one second model based on the first model.
  • a model supervision device comprising:
  • a sending module configured to send model information of a first model to a first communication device, wherein the model information of the first model is used to determine the first model and at least one second model associated with the first model, wherein the first model is used to supervise the validity of the at least one second model;
  • a receiving module is used to receive first information from the first communication device, where the first information is related to the first model or the at least one second model; a monitoring module is used to monitor the validity of the at least one second model based on the first information and the first model.
  • a communication device which includes a processor and a memory, wherein the memory stores a program or instruction that can be run on the processor, and when the program or instruction is executed by the processor, the steps of the method described in the first aspect are implemented, or the steps of the method described in the second aspect are implemented.
  • a communication device comprising a processor and a communication interface, wherein the communication interface is used to receive model information of a first model from a second communication device, and the processor is used to determine the first model and at least one second model associated with the first model according to the model information of the first model; supervise the validity of the at least one second model according to the first model; or,
  • the communication interface is used to send first information to the second communication device, where the first information is related to the first model or the at least one second model, and the first information is used by the second communication device to supervise the validity of the at least one second model according to the first model; or
  • the communication interface is used to send model information of a first model to a first communication device, where the model information of the first model is used to determine the first model and at least one second model associated with the first model, where the first model is used to supervise the validity of the at least one second model; or,
  • the communication interface is used to receive first information from the first communication device, where the first information is related to the first model or the at least one second model; the processor is used to monitor the validity of the at least one second model based on the first information and the first model.
  • a readable storage medium on which a program or instruction is stored.
  • the program or instruction is executed by a processor, the steps of the method described in the first aspect are implemented, or the steps of the method described in the second aspect are implemented.
  • a wireless communication system comprising: a first communication device and a second communication side device, wherein the first communication device can be used to execute the steps of the method described in the first aspect, and the second communication device can be used to execute the steps of the method described in the second aspect.
  • a chip comprising a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the method described in the first aspect, or to implement the method described in the second aspect.
  • a computer program/program product is provided, wherein the computer program/program product is stored in a storage medium, and the program/program product is executed by at least one processor to implement the steps of the model supervision method as described in the first aspect, or to implement the steps of the model supervision method as described in the second aspect.
  • the first communication device may receive model information of the first model from the second communication device, and then determine the first model and at least one second model associated with the first model based on the model information, and use the first model to supervise the validity of the at least one second model.
  • the first communication device sends first information to the second communication device, and the first information is related to the first model or at least one second model, which can be used by the second communication device to supervise the validity of at least one second model based on the first model. In this way, by using the first model to supervise the validity of at least one second model, the problem of task failure caused by still using the second model to perform the reasoning task when the second model is invalid can be avoided.
  • FIG1 is a schematic diagram of a wireless communication system according to an embodiment of the present application.
  • FIG2 is a schematic flow chart of a model supervision method according to an embodiment of the present application.
  • FIG3 is a schematic flow chart of a model supervision method according to an embodiment of the present application.
  • FIG4 is a schematic flow chart of a model supervision method according to an embodiment of the present application.
  • FIG5 is a schematic flow chart of a model supervision method according to an embodiment of the present application.
  • FIG6 is a schematic diagram of uneven distribution detection based on AI/ML according to an embodiment of the present application.
  • FIG7 is a schematic diagram of the results of uneven distribution detection based on AI/ML according to an embodiment of the present application.
  • FIG8 is a schematic diagram of the structure of a model supervision device according to an embodiment of the present application.
  • FIG9 is a schematic diagram of the structure of a model supervision device according to an embodiment of the present application.
  • FIG10 is a schematic diagram of the structure of a communication device according to an embodiment of the present application.
  • FIG11 is a schematic diagram of the structure of a communication device according to an embodiment of the present application.
  • FIG. 12 is a schematic diagram of the structure of a communication device according to an embodiment of the present application.
  • first, second, etc. of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the terms used in this way are interchangeable where appropriate, so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein, and the objects distinguished by “first” and “second” are generally of one type, and the number of objects is not limited, for example, the first object can be one or more.
  • “or” in the present application represents at least one of the connected objects.
  • “A or B” covers three schemes, namely, Scheme 1: including A but not including B; Scheme 2: including B but not including A; Scheme 3: including both A and B.
  • the character "/" generally indicates that the objects associated with each other are in an "or” relationship.
  • indication in this application can be a direct indication (or explicit indication) or an indirect indication (or implicit indication).
  • a direct indication can be understood as the sender explicitly informing the receiver of specific information, operations to be performed, or request results in the sent indication;
  • an indirect indication can be understood as the receiver determining the corresponding information according to the indication sent by the sender, or making a judgment and determining the operation to be performed or the request result according to the judgment result.
  • LTE Long Term Evolution
  • LTE-A Long Term Evolution-Advanced
  • CDMA Code Division Multiple Access
  • TDMA Time Division Multiple Access
  • FDMA Frequency Division Multiple Access
  • OFDMA Orthogonal Frequency Division Multiple Access
  • SC-FDMA Single-carrier Frequency-Division Multiple Access
  • NR New Radio
  • 6G 6th Generation
  • FIG1 shows a block diagram of a wireless communication system applicable to an embodiment of the present application.
  • the wireless communication system includes a terminal 11 and a network side device 12.
  • the terminal 11 may be a mobile phone, a tablet computer (Tablet Personal Computer), a laptop computer (Laptop Computer), a notebook computer, a personal digital assistant (Personal Digital Assistant, PDA), a handheld computer, a netbook, an ultra-mobile personal computer (Ultra-mobile Personal Computer, UMPC), a mobile Internet device (Mobile Internet Device, MID), an augmented reality (Augmented Reality, AR), a virtual reality (Virtual Reality, VR) device, a robot, a wearable device (Wearable Device), a flight vehicle, a vehicle user equipment (VUE), a shipborne equipment, a pedestrian terminal (Pedestrian User Equipment, PUE), a smart home (a home appliance with wireless communication function, such as a refrigerator, a television, a washing machine or furniture, etc.), a game console, a personal computer (Per
  • Terminal side equipment include: smart watches, smart bracelets, smart headphones, smart glasses, smart jewelry (smart bracelets, smart bracelets, smart rings, smart necklaces, smart anklets, smart anklets, etc.), smart wristbands, smart clothing, etc.
  • the vehicle-mounted equipment can also be called a vehicle-mounted terminal, a vehicle-mounted controller, a vehicle-mounted module, a vehicle-mounted component, a vehicle-mounted chip or a vehicle-mounted unit, etc. It should be noted that the specific type of the terminal 11 is not limited in the embodiment of the present application.
  • the network side device 12 may include an access network device or a core network device, wherein the access network device may also be referred to as a radio access network (Radio Access Network, RAN) device, a radio access network function or a radio access network unit.
  • the access network device may include a base station, a wireless local area network (Wireless Local Area Network, WLAN) access point (Access Point, AS) or a wireless fidelity (Wireless Fidelity, WiFi) node, etc.
  • WLAN wireless Local Area Network
  • AS Access Point
  • WiFi wireless Fidelity
  • the base station may be referred to as a Node B (NB), an evolved Node B (eNB), a next generation Node B (gNB), a New Radio Node B (NR Node B), an access point, a Relay Base Station (RBS), a Serving Base Station (SBS), a Base Transceiver Station (BTS), a radio base station, a radio transceiver, a Basic Service Set (BSS), an Extended Service Set (ESS), a Home Node B (HNB), a Home Evolved Node B, a Transmission Reception Point (TRP) or other appropriate terms in the field.
  • NB Node B
  • eNB evolved Node B
  • gNB next generation Node B
  • NR Node B New Radio Node B
  • an access point a Relay Base Station
  • SBS Serving Base Station
  • BTS Base Transceiver Station
  • a radio base station a radio transceiver
  • BSS Basic Service Set
  • ESS Extended Service Set
  • HNB Home No
  • model in the embodiments of the present application can be an artificial intelligence (AI) model, including but not limited to neural networks, decision trees, support vector machines, Bayesian classifiers, etc.
  • AI artificial intelligence
  • an embodiment of the present application provides a model supervision method 200, which can be executed by a first communication device.
  • the model supervision method can be executed by software or hardware installed on the first communication device.
  • the model supervision method includes the following steps.
  • the first communication device receives model information of the first model from the second communication device.
  • the first communication device may receive model information of the first model from the second communication device when performing the model supervision.
  • the first communication device may be a terminal
  • the second communication device may be a network side device, which may be a base station, a location server, a network data analysis function (Network Data Analytics Function, NWDAF) or a model supervision device, etc.
  • NWDAF Network Data Analytics Function
  • the first model is a model for model supervision
  • the model information of the first model is information related to the first model, which can be used to determine (or construct) the first model and at least one second model related to the first model
  • the at least one second model is a supervised model.
  • the at least one second model may have the same function, for example, the at least one second model may all be positioning models.
  • the at least one second model may be different from the first model, that is, the first model may be used to perform model supervision on one or more models other than the first model.
  • the at least one second model may include the first model, that is, one of the at least one second model is the first model.
  • the first model in addition to supervising other models, can also perform self-supervision, for example, the first model has both model supervision and positioning functions, and the first model can supervise the effectiveness of itself and other positioning models.
  • the first model when the number of second models is multiple, the at least one second model is different from the first model, and when the number of second models is 1, the first model may be the same or different from the second model.
  • the first model can output both the inference result and the result indicating whether the first model is valid.
  • the model information of the first model may include at least one of the following:
  • Model parameter information of the first model is
  • the model parameter information of the first model may be, for example, the weight and bias of each model layer in the first model.
  • the model configuration information of the first model may include configuration information of the first model itself and information of a supervised model (i.e., at least one second model) associated with the first model.
  • the model configuration information of the first model may include at least one of the following (1) to (7):
  • the identification information here can be used to determine the first model.
  • the first communication device is pre-configured with multiple models that can be used for model supervision.
  • the identification information can be used to determine which model to use for model supervision, and this model is the first model.
  • the identification information here can be used to distinguish different model functions.
  • the model function of the first model can be used to indicate what the first model is used for, for example, the first model is used for model supervision, and can be used specifically for model supervision of the positioning model.
  • the identification information here is used to determine at least one second model.
  • the first communication device can determine which models are supervised by using the first model, that is, determine which models are supervised.
  • Model function of at least one second model or identification information of the model function
  • the identification information here can be used to distinguish different model functions, and the model function of the second model can be used to indicate what the second model is used for, for example, the second model is used for terminal positioning.
  • the model structure information of the first model can be used to construct the first model, and specifically may include the model input of the first model, the model output, the specific structure of each model layer, etc.
  • the model structure information of the first model may include the model input of the first model, the model output, the number of neuron layers, the number of neurons in each layer, the activation function used, etc.
  • the association relationship here may be an association relationship between the model structures of the first model and at least one second model.
  • the association relationship may be used to determine or construct the first model or at least one second model.
  • the association relationship may include at least one of the following:
  • the first model has the same model input as at least one second model
  • the first model and the at least one second model have partially the same model structure.
  • the structures of the first model and the at least one second model differ only in the last model layer, and the structures of the remaining model layers are the same.
  • Model complexity (parameter scale or parameter quantity) and computational complexity of the first model.
  • the first communication device may execute S204.
  • the first communication device determines the first model and at least one second model associated with the first model according to the model information of the first model.
  • the first communication device After receiving the model information of the first model from the second communication device, the first communication device can determine the first model and at least one second model according to the model information of the first model.
  • the first model is a model for model supervision
  • the at least one second model is a supervised model.
  • the first communication device can construct the first model according to the model structure information and model parameter information of the first model, and determine at least one corresponding second model from multiple preconfigured models according to the identification information of at least one second model.
  • the first communication device can construct the first model according to the model structure information and model parameter information of the first model, and construct at least one second model according to the association relationship between the first model and at least one second model and the first model.
  • the first communication device can also determine the first model and the first model in other ways according to the model information of the first model. There is one less second model, and I will not give examples here one by one.
  • S206 The first communication device supervises the validity of at least one second model according to the first model.
  • the first communication device may monitor the validity of the at least one second model based on the first model.
  • the method may include:
  • the first communication device collects N corresponding input samples according to the model input of the first model, where N is a positive integer greater than 0;
  • the first communication device inputs the input sample into the first model to obtain a model output of the first model, where the model output of the first model is used to indicate whether at least one second model is valid under the input sample.
  • the model input of the first model can be determined by the first communication device according to the model information of the first model.
  • N input samples can be collected by the first communication device in an actual application environment.
  • the model input of the first model is the channel state information of the UE.
  • the first model is used to supervise at least one second model with a positioning function.
  • the at least one second model will be applied to predict the position of the UE in environment 1.
  • the first communication device can measure the channel state information of the UE in environment 1 and obtain the channel state information measurement results of N UEs, that is, N input samples.
  • the N input sample representation can supervise at least one second model N times, and one input sample corresponds to one supervision and obtains an output result of the first model.
  • the specific value of N can be pre-configured by the first communication device, or indicated by the second communication device, or pre-defined by the protocol.
  • any input sample can be input into the first model to obtain the model output of the first model.
  • the model output of the first model can indicate whether at least one supervised second model is effective under the input sample.
  • the second model is valid under a certain input sample and may include at least one of the following (1) to (3):
  • the confidence that the second model is effective under the input sample is greater than or equal to a first threshold; wherein the confidence that the second model is effective under the input sample may be the confidence of the model output of the second model after the input sample is input into the second model, and the first threshold may be preconfigured by the first communication device, or indicated by the second communication device, or predefined by a protocol;
  • the accuracy of the model output of the second model is greater than or equal to a preset accuracy;
  • the preset accuracy may be preconfigured by the first communication device, or indicated by the second communication device, or predefined by a protocol;
  • the distribution of the input sample is consistent with the distribution of the training sample used to train the second model; wherein, when the distribution is consistent, after the input sample is input into the second model, the confidence of the model output of the second model is greater than or equal to the first threshold, or the accuracy of the model output of the second model is greater than or equal to the preset accuracy.
  • the first model is model A
  • the supervised model is model B (that is, the number of the second model is 1).
  • model B is a positioning model
  • the model input and the model input of model A are both channel impulse response (CIR)
  • the output of model A can indicate whether the positioning result obtained after the CIR sample is input into model B is accurate (that is, whether the accuracy is greater than or equal to the preset accuracy), or indicate whether model B is effective when the input is the CIR sample, or indicate whether the CIR sample is consistent with the distribution of the training set CIR of model B.
  • the first model is model A
  • the supervised model is model [B, C, D...] (i.e., there are multiple second models).
  • models [B, C, D...] are all positioning models
  • the model input and the model input of model A are both CIR
  • the CIR sample collected in the actual environment can be input into model A.
  • the output of model A can indicate whether the positioning result obtained after the CIR sample is input into model [B, C, D...] is accurate, or whether model [B, C, D...] is effective when the input is the CIR sample, or whether the CIR sample is consistent with the CIR distribution of the training set of each model [B, C, D...].
  • the first model is model C
  • the second model is also model C (that is, the first model and the second model are the same Model, model C can realize self-supervision)
  • the model input of model C is CIR
  • the output is the positioning result
  • the output of model C can also indicate whether the positioning result obtained after the CIR sample is input into model C is accurate, or indicate whether model C is effective when the model input is the CIR sample, or indicate whether the CIR is consistent with the distribution of the training set CIR of model C.
  • the model output of the first model may include at least one of the following:
  • the confidence value is in the range of [0,1].
  • the sum of the confidences of at least one second model being valid under the input sample may be equal to 1.
  • the number of second models is 4, and for a certain input sample, the confidence output by the first model may be [0.8, 0.1, 0.05, 0.02, 0.03], where 0.8, 0.1, 0.05, and 0.02 respectively represent the confidences of Model 1, Model 2, Model 3, and Model 4 being valid under the input sample, and 0.03 represents the confidence that these four models are all invalid under the input sample.
  • the validity of the second model under the input sample can be determined based on the confidence of the second model under the input sample. For example, the confidence output by the first model for the above four models is [0.8, 0.1, 0.05, 0.02, 0.03]. Assuming that the confidence threshold for the model validity is 0.7, it can be determined that Model 1 is valid and the remaining models are invalid.
  • the validity indication of at least one second model under the input sample can be determined according to the first indication information and the validity supervision result of at least one second model under the input sample.
  • the first indication information is used to indicate the association relationship between different validity indications and different validity supervision results of at least one second model.
  • the number of second models is 4, and the supervision results can include at least 5 types, namely, model 1 is valid and the other models are invalid, model 2 is valid and the other models are invalid, model 3 is valid and the other models are invalid, model 4 is valid and the other models are invalid, and models 1 to 4 are all invalid.
  • the first indication information can indicate that when model 1 is valid and the other models are invalid, the validity indication is 10000, when model 2 is valid and the other models are invalid, the validity indication is 01000, when model 3 is valid and the other models are invalid, the validity indication is 00100, when model 4 is valid and the other models are invalid, the validity indication is 00010, and when models 1 to 4 are all invalid, the validity indication is 00001. In this way, when only model 1 of the four models is valid and the other models are invalid under a certain input sample, the validity indication corresponding to these four models is 00010.
  • the first indication information can be indicated by the second communication device to the first communication device, that is, when the first communication device performs model supervision, the first communication device can receive the first indication information from the second communication device.
  • the validity indication in the output result of the first model can be determined based on the first indication information and the supervision result.
  • the first communication device obtains the effectiveness supervision result of at least one second model through the above method, that is, after obtaining the output result of the first model under N input samples, optionally, at least one of the following operations may be performed:
  • the first communication device sends a validity supervision result of at least one second model to the second communication device;
  • the first communication device selects or activates a target model according to a result of monitoring the effectiveness of at least one second model.
  • the first communication device can select or activate the valid model, and the valid model is the target model.
  • the first communication device can use the valid model to perform the reasoning task and avoid the failure of the reasoning task.
  • the first communication device can also send the validity supervision result of at least one second model to the second communication device, so that the second communication device can determine the validity of the at least one second model based on the validity supervision result, and then select or activate the target model.
  • the first communication device can send the validity supervision result of at least one second model to the second communication device, so that the second communication device can configure the valid model after knowing the validity of at least one second model, and then select or activate the valid model and send the valid model to the first communication device.
  • the first communication device can The effective model performs reasoning tasks and avoids failure of reasoning tasks.
  • the first communication device can send the validity supervision result of at least one second model to the second communication device, so that the second communication device can select or activate other models after knowing that at least one second model is invalid, and send the selected or activated other models to the first communication device. In this way, the first communication device can use other models to perform reasoning tasks, thereby avoiding the problem of reasoning task failure caused by still using invalid second models to perform reasoning tasks.
  • the first communication device may first deactivate a currently activated invalid model, and then select or activate the target model after deactivating the invalid model.
  • the first communication device when the first communication device sends the effectiveness supervision result of at least one second model to the second communication device, the first communication device may include:
  • the first communication device receives second indication information from the second communication device, where the second indication information is used to indicate a reporting method or content of a validity supervision result of at least one second model;
  • the first communication device sends a validity supervision result of at least one second model to the second communication device according to the second indication information.
  • the reporting content indicated by the second instruction information may include at least one of the following (1) to (7):
  • the first model For each input sample, the first model has a model output, and each model output is used to indicate the effectiveness of at least one second model under an input sample.
  • the model output of the first model may include at least one of the confidence that at least one second model is effective under the input sample and the indication of the effectiveness of at least one second model under the input sample.
  • the validity indicated by the validity indication information here is the comprehensive result of the validity of at least one second model under N input samples.
  • the second model is valid under N input samples, which can be that the proportion of valid samples of the second model to the N input samples is greater than or equal to the proportion threshold.
  • the valid sample of the second model means that, for any input sample, if the second model is valid under the input sample, then the input sample is a valid sample of the second model.
  • the validity result of the second model under N input samples can be determined based on the model output of the first model under N input samples.
  • the proportion threshold can be pre-configured by the first communication device, or indicated by the second communication device, or pre-defined by the protocol.
  • the confidence here can be the confidence of the validity indicated by the validity indication information in (3), that is, the confidence that the second model is valid under N input samples.
  • the confidence can be equal to the ratio of valid samples of the second model to the N input samples.
  • identification information of a target second model where the target second model is one or more models in at least one second model whose effective confidence level is greater than or equal to a preset confidence level;
  • the confidence here may be the confidence in (4).
  • the first communication device may also determine one or more second models with higher confidence from the at least one second model, i.e., the target second model, and then send the identification information of the target second model to the second communication device.
  • the second communication device can refer to the target second model to configure a model for the first communication device. For example, when selecting or activating a target model, the target second model can be selected or activated.
  • each second model can correspond to such a ratio.
  • each second model can correspond to such a ratio.
  • the sent validity supervision result may include at least one item of (1) to (7) in the reported content.
  • the reporting method indicated by the second indication information may include at least one of the following:
  • M is a positive integer greater than 0, and the time unit may include: Orthogonal Frequency Division Multiplexing (OFDM) symbol, slot, subframe, frame, microsecond, millisecond, second, minute, hour, day, etc.;
  • OFDM Orthogonal Frequency Division Multiplexing
  • the validity supervision result of the second model is reported, that is, the validity supervision result of the second model is reported only when the second model is invalid, and when the second model is valid, you can choose not to report the validity supervision result of the second model. In this way, transmission resources can be saved.
  • the second communication device when the validity supervision result of a second model is not received, the second model can be assumed to be valid.
  • the first communication device may send the validity supervision result of at least one second model to the second communication device based on the reporting method.
  • the first communication device may receive model information of the first model from the second communication device, and then determine the first model and at least one second model associated with the first model according to the model information, and use the first model to supervise the validity of the at least one second model. In this way, by using the first model to supervise the validity of at least one second model, the problem of task failure caused by still using the second model to perform the reasoning task when the second model is invalid can be avoided.
  • an embodiment of the present application provides a model supervision method 300, which can be executed by a first communication device.
  • the model supervision method can be executed by software or hardware installed on the first communication device.
  • the model supervision method includes the following steps.
  • the first communication device sends first information to the second communication device, where the first information is related to the first model or at least one second model, and the first information is used by the second communication device to monitor the validity of the at least one second model according to the first model.
  • the first communication device may send the first information to the second communication device.
  • the first communication device may be a terminal
  • the second communication device may be a network side device, which may be a base station, a location server, a NWDAF or a model supervision device.
  • the first information is related to the first model and at least one second model.
  • the first model is a model for model supervision, and at least one second model is a supervised model.
  • the at least one second model may have the same function, for example, the at least one second model may all be positioning models.
  • the at least one second model may be different from the first model, that is, the first model may be used to perform model supervision on one or more models other than the first model.
  • the at least one second model may include the first model, that is, one of the at least one second model is the first model. In this case, the first model can supervise other models in addition to self-supervision.
  • the first model has both model supervision and positioning functions, and the first model can supervise the effectiveness of itself and other positioning models.
  • the first model may be the same or different from the second model.
  • the first model can output both the inference result and the result indicating whether the first model is valid.
  • the first information can be used by the second communication device to monitor the validity of at least one second model according to the first model.
  • the specific implementation method of the second communication device to monitor the validity can be found in the embodiment shown in FIG5 , which will not be described in detail here.
  • the first information may include at least one of the following (1) to (7):
  • the model input here can be the measurement quantity measured by the first communication device in the actual environment, which can be CIR, power delay profile (Power Delay Profile, PDP), reference signal receiving power (Reference Signal Receiving Power, RSRP), arrival time (Time of Arrival, TOA), etc.
  • CIR power delay profile
  • PDP Power Delay Profile
  • RSRP Reference Signal Receiving Power
  • TOA Time of Arrival
  • the first communication device is a terminal
  • at least one second model is a model for positioning the terminal
  • the model input is CIR. Then, when the second communication device performs model supervision, the first communication device needs to report the measurement value of CIR in the actual environment to the second communication device, so that the second communication device can input these CIR measurement values as input samples into the first model to obtain the supervision result of at least one second model.
  • the acquisition time of the model input may be the measurement time of the measured quantity of the model input in (1), and may be used to distinguish different measured quantities.
  • the identification information here may correspond to one or a group of model inputs in (1), that is, an ID is assigned to each measurement quantity or each group of measurement quantities.
  • the model output can be position, TOA, Reference Signal Time Difference (RSTD), Angle of arrival (AOA), Angle of Departure (AOD), etc., which can be determined by the function of the second model. For example, if the second model is a positioning model, the model output is position.
  • RSTD Reference Signal Time Difference
  • AOA Angle of arrival
  • AOD Angle of Departure
  • the identification information here can be used to determine at least one second model. For example, multiple models are pre-configured in the second communication device, and by sending the identification information to the second communication device, the second communication device can determine which of these models are the second models that need to be supervised according to the identification information.
  • Model function of at least one second model or identification information of the model function
  • the identification information here can be used to distinguish different model functions, and the model function of the second model can be used to indicate what the second model is used for, for example, the second model is used for terminal positioning.
  • TRPs transmission reception points
  • PRS positioning reference signal
  • the following steps may also be included:
  • the first communication device receives a validity supervision result of at least one second model from the second communication device.
  • the second communication device can supervise the validity of at least one second model based on the first information and using the first model. After obtaining the validity supervision result, the validity supervision result can be sent to the first communication device, and the first communication device can receive the validity supervision result from the second communication device.
  • the second communication device receiving the validity supervision result may include at least one of the following (1) to (5):
  • Sending identification information of at least one second model to the first communication device can facilitate the first communication device to determine the Supervised Second Model.
  • Model function of at least one second model or identification information of the model function (2) Model function of at least one second model or identification information of the model function.
  • the identification information here can be used to distinguish different model functions, and the model function of the second model can be used to indicate what the second model is used for, for example, the second model is used for terminal positioning.
  • the validity indicated by the validity indication here may be the validity of the second model under one input sample, or the overall validity of the second model under N input samples, where N is a positive integer greater than 0.
  • the validity indication method may be the same as the validity indication method included in the model output of the first model in the embodiment shown in FIG. 2, which is not described in detail here.
  • the validity indication method may be indicated using 1 bit, for example, a bit value of 0 indicates invalidity, and a bit value of 1 indicates validity.
  • validity indication may also be performed in other ways, which are not described one by one here.
  • the second model is valid under N input samples, which may be that the ratio of the valid samples of the second model to the N input samples is greater than or equal to the ratio threshold.
  • the valid sample of the second model means that, for any input sample, when the second model is valid under the input sample, the input sample is a valid sample of the second model.
  • the ratio threshold may be preconfigured by the second communication device, or indicated by the first communication device, or predefined by the protocol.
  • the confidence here can be the confidence of the second model under one input sample, or the overall confidence of the second model under N input samples. If it is the confidence of the second model under one input sample, then the confidence can be the confidence contained in the model output after inputting one input sample into the first model. If it is the confidence of the second model under N input samples, then the confidence can be the ratio of the valid samples of the second model to the N input samples.
  • Model indication information is used to indicate model activation or model selection.
  • the model indication information may include identification information of the target model.
  • the identification information of the target model may be the identification information of the target second model.
  • the first communication device may refer to the target second model for model activation or model selection, such as selecting or activating the target second model. If at least one second model is invalid, the identification information of the target model may be the identification information of a valid model (having the same model function as the second model, such as both are positioning models) configured and activated or selected by the second communication device. After receiving the model indication information, the first communication device may select or activate the target model.
  • the first communication device may further include:
  • the first communication device activates or selects a target model according to a result of monitoring the effectiveness of at least one second model.
  • the first communication device can directly perform model selection or model activation according to the model indication information, and the target model selected or activated can be the target model indicated in the model indication information. If the validity supervision result does not include the above-mentioned model indication information, the first communication device can determine which second model among at least one second model is valid according to other information in the validity supervision result, and then select or activate the valid second model. Optionally, when selecting or activating the target model, the first communication device can first deactivate the invalid model that is currently in an activated state, and then select or activate the target model after deactivating the invalid model.
  • the first communication device After activating or selecting the target model, the first communication device can use the target model to perform the reasoning task. Since the target model is a valid model, the first communication device can avoid the failure of the reasoning task execution by using the target model to ensure the successful execution of the reasoning task.
  • the first communication device may send first information to the second communication device, where the first information is related to the first model or at least one second model and can be used by the second communication device to supervise the validity of at least one second model based on the first model.
  • the first model to supervise the validity of at least one second model, This can avoid the problem of task failure caused by using the second model to perform the reasoning task when the second model is invalid.
  • an embodiment of the present application provides a model supervision method 400, which can be executed by a second communication device.
  • the model supervision method can be executed by software or hardware installed on the second communication device.
  • the model supervision method includes the following steps.
  • the second communication device sends model information of the first model to the first communication device, where the model information of the first model is used to determine the first model and at least one second model associated with the first model, and the first model is used to supervise the validity of the at least one second model.
  • the second communication device may send the model information of the first model to the first communication device.
  • the first communication device may be a terminal
  • the second communication device may be a network side device, which may be a base station, a location server, an NWDAF or a model supervision device, etc.
  • the first model is a model for model supervision
  • the model information of the first model is information related to the first model, which can be used to determine (or construct) the first model and at least one second model related to the first model
  • the at least one second model is a supervised model.
  • the at least one second model may have the same function, for example, the at least one second model may all be positioning models.
  • the at least one second model may be different from the first model, that is, the first model may be used to perform model supervision on one or more models other than the first model.
  • the at least one second model may include the first model, that is, one of the at least one second model is the first model.
  • the first model in addition to supervising other models, can also perform self-supervision, for example, the first model has both model supervision and positioning functions, and the first model can supervise the effectiveness of itself and other positioning models.
  • the first model when the number of second models is multiple, the at least one second model is different from the first model, and when the number of second models is 1, the first model may be the same or different from the second model.
  • the first model can output both the inference result and the result indicating whether the first model is valid.
  • the model information of the first model may include at least one of the following:
  • Model parameter information of the first model is
  • the model parameter information of the first model may be, for example, the weight and bias of each model layer in the first model.
  • the model configuration information of the first model may include configuration information of the first model itself and information of a supervised model (i.e., at least one second model) associated with the first model.
  • the model configuration information of the first model may include at least one of the following:
  • Model structure information of the first model
  • association relationship between the first model and at least one second model may include at least one of the following:
  • a model input to the first model is the same as a model input to at least one second model
  • the first model has the same partial model structure as the at least one second model.
  • the first communication device can determine the first model and at least one second model associated with the first model according to the model information of the first model, and then supervise the effectiveness of the at least one second model according to the first model. For specific implementation methods, see the embodiment shown in FIG. S204 and S206 are not described again here.
  • the second communication device may further include:
  • the second communication device receives a validity supervision result of at least one second model from the first communication device.
  • the first communication device can send the validity supervision result to the second communication device, and the second communication device can receive the validity supervision result of at least one second model from the first communication device.
  • the second communication device may instruct the first communication device how to report the validity supervision result of at least one second model to the first communication device, which may specifically include:
  • the second communication device sends second indication information to the first communication device, where the second indication information is used to indicate a reporting method or content of a validity supervision result of at least one second model.
  • the reporting content indicated by the second indication information may include at least one of the following:
  • Validity indication information used to indicate whether at least one second model is effective under N input samples
  • Identification information of a target second model where the target second model is one or more models in at least one second model whose effective confidence level is greater than or equal to a preset confidence level;
  • the sent validity supervision result may include at least one item in the reported content.
  • the reporting method indicated by the second indication information may include at least one of the following:
  • the time unit may include: OFDM symbol, time slot, subframe, frame, microsecond, millisecond, second, minute, hour, day, etc.;
  • the validity supervision result of the second model is reported, that is, the validity supervision result of the second model is reported when the second model is invalid, and the validity supervision result of the second model may not be reported when the second model is valid. In this way, transmission resources can be saved.
  • the second communication device when the validity supervision result of a second model is not received, the second model can be assumed to be valid.
  • the first communication device may send the validity supervision result of at least one second model to the second communication device based on the reporting method.
  • the method may further include:
  • the second communication device selects or activates a target model according to a result of monitoring the effectiveness of at least one second model.
  • the second communication device may determine a second model whose confidence level is greater than or equal to the confidence level threshold according to the confidence level threshold, and then select or activate the second model.
  • the selected or activated second model is the target model.
  • the second communication device may select or activate the target model with reference to the target second model.
  • the target second model may be directly used as the target model. model, and select or activate the target second model.
  • the second communication device can configure and select or activate other valid models, which have the same model function as the second model, such as positioning models.
  • the second communication device can also indicate the valid model to the first communication device after selecting or activating other valid models, so that the first communication device uses the valid model to perform the reasoning task and ensure the successful execution of the reasoning task.
  • the second communication device may first deactivate a currently activated invalid model, and then select or activate the target model after deactivating the invalid model.
  • the second communication device may also send the effectiveness supervision to other second communication devices.
  • the second communication device is a base station or a location server, after receiving the effectiveness supervision result of at least one second model, it may send it to the NWDAF or the model monitoring device to realize the sharing of the effectiveness supervision result.
  • the second communication device may send model information of the first model to the first communication device, so that the first communication device may determine the first model and at least one second model associated with the first model based on the model information, and use the first model to supervise the validity of the at least one second model.
  • the problem of task failure caused by still using the second model to perform the reasoning task when the second model is invalid can be avoided.
  • an embodiment of the present application provides a model supervision method 500, which can be executed by a second communication device.
  • the model supervision method can be executed by software or hardware installed on the second communication device.
  • the model supervision method includes the following steps.
  • the second communication device receives first information from the first communication device, where the first information is related to the first model or at least one second model.
  • the second communication device may receive the first information from the first communication device.
  • the first communication device may be a terminal
  • the second communication device may be a network side device, which may be a base station, a location server, a NWDAF or a model supervision device.
  • the first information is related to the first model and at least one second model.
  • the first model is a model for model supervision, and at least one second model is a supervised model.
  • the at least one second model may have the same function, for example, the at least one second model may all be positioning models.
  • the at least one second model may be different from the first model, that is, the first model may be used to perform model supervision on one or more models other than the first model.
  • the at least one second model may include the first model, that is, one of the at least one second model is the first model. In this case, the first model can supervise other models in addition to self-supervision.
  • the first model has both model supervision and positioning functions, and the first model can supervise the effectiveness of itself and other positioning models.
  • the first model may be the same or different from the second model.
  • the first model can output both the inference result and the result indicating whether the first model is valid.
  • the first information may include at least one of the following:
  • the second communication device monitors the validity of at least one second model according to the first information and the first model.
  • the second communication device can supervise the validity of at least one second model based on the first information and the first model.
  • the first model and at least one second model can be pre-configured in the second communication device, and the second communication device can directly use the first model to perform model supervision on at least one second model when performing model supervision.
  • the specific implementation method can be the same as the specific implementation method of the first communication model supervision shown in FIG2, that is, the model input in the first information is input as an input sample into the first model to obtain the model output of the first model, and the model output can characterize whether the at least one second model is valid under the input sample.
  • the specific meaning of whether the second model is valid under the input sample can also refer to the specific description of the corresponding content in the embodiment shown in FIG2, which will not be described in detail here.
  • the second communication device may perform at least one of the following operations:
  • the second communication device activates or selects the target model according to the effectiveness supervision result of at least one second model
  • the second communication device sends a validity supervision result of at least one second model to the first communication device.
  • the second communication device selects or activates the target model, for example, if there is a valid second model in at least one second model, the valid second model can be selected or activated. If at least one second model is invalid, other valid models can be configured and selected or activated. The functions of the other valid models are the same as those of the second models, for example, they can all be positioning models.
  • the second communication device selects or activates the target model, it can first deactivate the invalid model that is currently in the activated state, and then select or activate the target model after deactivating the invalid model.
  • the second communication device may send the validity supervision result of at least one second model to the first communication device, and the sent validity supervision result may include at least one of the following:
  • Model indication information the model indication information is used to indicate model activation or model selection.
  • the target model After receiving the validity supervision result sent by the second communication device, the target model can be selected or activated.
  • the specific implementation method please refer to the specific implementation of the corresponding steps in the embodiment shown in Figure 3, which will not be repeated here.
  • the first communication device can use the target model to perform reasoning tasks. Since the target model is a valid model, the first communication device uses the target model to perform reasoning tasks, which can avoid the failure of reasoning task execution and ensure the successful execution of the reasoning task.
  • the second communication device can receive the first information sent by the first communication device, the first information is related to the first model or at least one second model, and the second communication device can supervise the validity of at least one second model based on the first information and the first model. In this way, by using the first model to supervise the validity of at least one second model, the problem of task failure caused by still using the second model to perform the reasoning task when the second model is invalid can be avoided.
  • AI/ML artificial intelligence/machine learning
  • a model monitoring solution based on offline training is proposed, namely, AI/ML-based uneven distribution detection.
  • the model is used to learn the difference between the original training dataset and the non-original training dataset.
  • the AI/ML model can indicate whether the CIR belongs to the distribution of the original training dataset, or the possibility of belonging to the distribution of the original training set.
  • the specific implementation method is as follows:
  • Step 1 Build a classifier training dataset.
  • the training dataset consists of two parts.
  • the first part is the original training dataset used to train the AI/ML model for positioning.
  • the second part is the non-original training dataset, which has a different distribution from the original training dataset.
  • the non-original training dataset can be collected from other real environments, or it can also be generated through simulation. Then, all samples in the original training dataset are labeled with the label "1", and all samples in the non-original training dataset are labeled with the label "0".
  • Step 2 Offline training.
  • the binary classification model is trained offline using the constructed training dataset. In this way, the boundaries of the distribution of the original training dataset can be learned.
  • Step 3 Model inference.
  • the CIR with unknown distribution collected from the real environment is input into the well-trained AI/ML model, and the AI/ML model can predict whether the CIR belongs to the distribution of the original training data set, or whether it belongs to the possibility of the original training set distribution.
  • a predefined threshold for example, 90%
  • the model can indicate whether the input is consistent with the distribution of the training set or the confidence that it is consistent with the distribution of the training set;
  • the model supervision method provided in the embodiment of the present application can be executed by a model supervision device.
  • the model supervision device executing the model supervision method is taken as an example to illustrate the model supervision device provided in the embodiment of the present application.
  • Fig. 8 is a schematic diagram of the structure of a model supervision device according to an embodiment of the present application, and the device may correspond to the first communication device in other embodiments. As shown in Fig. 8, the device 800 includes the following modules.
  • a receiving module 801 is used to receive model information of a first model from a second communication device; a determining module 802 is used to determine the first model and at least one second model associated with the first model according to the model information of the first model; a monitoring module 803 is used to monitor the validity of the at least one second model according to the first model; or,
  • the sending module 804 is used to send first information to the second communication device, where the first information is related to the first model or the at least one second model, and the first information is used by the second communication device to monitor the validity of the at least one second model based on the first model.
  • the model information of the first model includes at least one of the following:
  • Model parameter information of the first model is
  • the model configuration information of the first model includes at least one of the following:
  • the model function of the first model or identification information of the model function
  • the model function or identification information of the model function of the at least one second model is the model function or identification information of the model function of the at least one second model
  • the association relationship between the first model and the at least one second model includes at least one of the following:
  • a model input of the first model is the same as a model input of the at least one second model
  • a partial model structure of the first model is the same as a partial model structure of the at least one second model.
  • the at least one second model includes the first model.
  • the monitoring module 803 is used to:
  • N a positive integer greater than 0;
  • the input sample is input into the first model to obtain a model output of the first model, where the model output of the first model is used to indicate whether the at least one second model is effective under the input sample.
  • the model output of the first model includes at least one of the following:
  • the second model is effective under the input sample, including:
  • the confidence that the second model is effective under the input sample is greater than or equal to a first threshold
  • the accuracy of the model output of the second model is greater than or equal to a preset accuracy
  • the input samples are consistent with the distribution of training samples used when training the second model.
  • the first threshold is preconfigured by the first communication device, or indicated by the second communication device, or predefined by a protocol.
  • the validity indication is determined according to the first indication information and a validity supervision result of the at least one second model under the input sample;
  • the first indication information is used to indicate the association between different validity indications and different validity supervision results for the at least one second model.
  • the receiving module 801 is further used for:
  • the first indication information is received from the second communication device.
  • At least one of the following is further included:
  • the sending module 804 is used to send the effectiveness supervision result of the at least one second model to the second communication device;
  • the supervision module 803 is used to select or activate a target model according to the effectiveness supervision result of the at least one second model.
  • the monitoring module 803 is used to:
  • Second indication information is received from the second communication device, wherein the second indication information is used to indicate the at least one The reporting method or content of the effectiveness monitoring results of the second model;
  • the validity supervision result of the at least one second model is sent to the second communication device according to the second indication information.
  • the reported content includes at least one of the following:
  • Validity indication information used to indicate whether the at least one second model is valid under N input samples
  • Identification information of a target second model wherein the target second model is one or more models of the at least one second model whose effective confidence level is greater than or equal to a preset confidence level;
  • the second model is valid under N input samples, including:
  • the ratio of the valid samples of the second model to the N input samples is greater than or equal to the ratio threshold
  • the input sample is a valid sample of the second model.
  • the ratio threshold is preconfigured by the first communication device, or indicated by the second communication device, or predefined by a protocol.
  • the reporting method includes at least one of the following:
  • the first information includes at least one of the following:
  • the model function or identification information of the model function of the at least one second model is the model function or identification information of the model function of the at least one second model
  • the at least one second model includes the IDs of multiple transmitting and receiving points TRP associated with the locations of multiple TRPs, the positioning reference signal PRS ID, and at least one of the PRS resource set IDs.
  • the receiving module 801 is further used for:
  • a validity supervision result of the at least one second model is received from the second communication device.
  • the effectiveness supervision result of the at least one second model includes at least one of the following:
  • the model function or identification information of the model function of the at least one second model is the model function or identification information of the model function of the at least one second model
  • Model indication information where the model indication information is used to indicate model activation or model selection.
  • the monitoring module 803 is further used to:
  • the target model is activated or selected according to the effectiveness supervision result of the at least one second model.
  • the apparatus 800 may refer to the processes of the methods 200 and 300 corresponding to the embodiments of the present application. Furthermore, the various units/modules in the device 800 and the other operations and/or functions mentioned above are respectively for implementing the corresponding processes in methods 200 and 300, and can achieve the same or equivalent technical effects. For the sake of brevity, they will not be repeated here.
  • Fig. 9 is a schematic diagram of the structure of a model supervision device according to an embodiment of the present application, and the device may correspond to the second communication device in other embodiments. As shown in Fig. 9, the device 900 includes the following modules.
  • a sending module 901 is used to send model information of a first model to a first communication device, where the model information of the first model is used to determine the first model and at least one second model associated with the first model, where the first model is used to supervise the validity of the at least one second model; or,
  • the receiving module 902 is used to receive first information from the first communication device, where the first information is related to the first model or the at least one second model; the monitoring module 903 is used to monitor the validity of the at least one second model based on the first information and the first model.
  • the model information of the first model includes at least one of the following:
  • Model parameter information of the first model is
  • the model configuration information of the first model includes at least one of the following:
  • the model function of the first model or identification information of the model function
  • the model function or identification information of the model function of the at least one second model is the model function or identification information of the model function of the at least one second model
  • the association relationship between the first model and the at least one second model includes at least one of the following:
  • a model input of the first model is the same as a model input of the at least one second model
  • the first model has the same partial model structure as the at least one second model.
  • the at least one second model includes the first model.
  • the receiving module 902 is further configured to:
  • a validity supervision result of the at least one second model is received from the first communication device.
  • the sending module 901 is further used to:
  • the second indication information is used to indicate a reporting method or a reporting content of a validity supervision result of the at least one second model.
  • the reported content includes at least one of the following:
  • Validity indication information used to indicate whether the at least one second model is valid under N input samples
  • Identification information of a target second model wherein the target second model is one or more models of the at least one second model whose effective confidence level is greater than or equal to a preset confidence level;
  • the second model is valid under N input samples, including:
  • the ratio of the valid samples of the second model to the N input samples is greater than or equal to the ratio threshold
  • the input sample is a valid sample of the second model.
  • the ratio threshold is preconfigured by the first communication device, or indicated by the second communication device, or predefined by a protocol.
  • the reporting method includes at least one of the following:
  • the monitoring module 903 is further used to:
  • a target model is selected or activated according to a result of monitoring the effectiveness of the at least one second model.
  • the first information includes at least one of the following:
  • the model function or identification information of the model function of the at least one second model is the model function or identification information of the model function of the at least one second model
  • the at least one second model includes the IDs of multiple transmitting and receiving points TRP associated with the locations of multiple TRPs, the positioning reference signal PRS ID, and at least one of the PRS resource set IDs.
  • At least one of the following is further included:
  • the supervision module 903 is used to activate or select a target model according to the effectiveness supervision result of the at least one second model
  • the sending module 901 is used to send the effectiveness supervision result of the at least one second model to the first communication device.
  • the effectiveness supervision result of the at least one second model includes at least one of the following:
  • the model function or identification information of the model function of the at least one second model is the model function or identification information of the model function of the at least one second model
  • Model indication information where the model indication information is used to indicate model activation or model selection.
  • the various units/modules in the device 900 and the above-mentioned other operations and/or functions are respectively for implementing the corresponding processes in the methods 400 and 500, and can achieve the same or equivalent technical effects, which will not be elaborated here for the sake of brevity.
  • the model supervision device in the embodiment of the present application can be an electronic device, such as an electronic device with an operating system, or a component in an electronic device, such as an integrated circuit or a chip.
  • the electronic device can be a terminal, or it can be other devices other than a terminal.
  • the terminal can include but is not limited to the types of terminal 11 listed above, and other devices can be servers, network attached storage (NAS), etc., which are not specifically limited in the embodiment of the present application.
  • the model supervision device provided in the embodiment of the present application can implement the various processes implemented by the method embodiments of Figures 2 to 5 and achieve the same technical effects. To avoid repetition, they will not be described here.
  • the embodiment of the present application further provides a communication device 1000, including a processor 1001 and a memory 1002, wherein the memory 1002 stores a program or instruction that can be run on the processor 1001.
  • the communication device 1000 is a first communication device
  • the program or instruction is executed by the processor 1001 to implement the various steps of the above-mentioned model supervision method embodiment, and can achieve the same technical effect.
  • the communication device 1000 is a second communication device
  • the program or instruction is executed by the processor 1001 to implement the various steps of the above-mentioned model supervision method embodiment, and can achieve the same technical effect.
  • the technical effects are not described here.
  • the embodiment of the present application also provides a communication device, including a processor and a communication interface, the communication interface and the processor are coupled, and the processor is used to run a program or instruction to implement the steps in the method embodiment shown in Figures 2 and 3.
  • the communication device embodiment corresponds to the above-mentioned first communication device side method embodiment, and each implementation process and implementation method of the above-mentioned method embodiment can be applied to the communication device embodiment, and can achieve the same technical effect.
  • Figure 11 is a schematic diagram of the hardware structure of a communication device implementing an embodiment of the present application.
  • the communication device 1100 includes but is not limited to: a radio frequency unit 1101, a network module 1102, an audio output unit 1103, an input unit 1104, a sensor 1105, a display unit 1106, a user input unit 1107, an interface unit 1108, a memory 1109 and at least some of the components of a processor 1110.
  • the communication device 1100 can also include a power supply (such as a battery) for supplying power to each component, and the power supply can be logically connected to the processor 1110 through a power management system, so as to implement functions such as charging, discharging, and power consumption management through the power management system.
  • a power supply such as a battery
  • the communication device structure shown in FIG11 does not constitute a limitation on the communication device, and the terminal can include more or fewer components than shown in the figure, or combine certain components, or arrange components differently, which will not be described in detail here.
  • the input unit 1104 may include a graphics processing unit (GPU) 11041 and a microphone 11042, and the graphics processor 11041 processes the image data of the static picture or video obtained by the image capture device (such as a camera) in the video capture mode or the image capture mode.
  • the display unit 1106 may include a display panel 11061, and the display panel 11061 may be configured in the form of a liquid crystal display, an organic light emitting diode, etc.
  • the user input unit 1107 includes a touch panel 11071 and at least one of other input devices 11072.
  • the touch panel 11071 is also called a touch screen.
  • the touch panel 11071 may include two parts: a touch detection device and a touch controller.
  • Other input devices 11072 may include, but are not limited to, a physical keyboard, function keys (such as a volume control key, a switch key, etc.), a trackball, a mouse, and a joystick, which will not be repeated here.
  • the RF unit 1101 can transmit the data to the processor 1110 for processing; in addition, the RF unit 1101 can send uplink data to the network side device.
  • the RF unit 1101 includes but is not limited to an antenna, an amplifier, a transceiver, a coupler, a low noise amplifier, a duplexer, etc.
  • the memory 1109 can be used to store software programs or instructions and various data.
  • the memory 1109 can mainly include a first storage area for storing programs or instructions and a second storage area for storing data, wherein the first storage area can store an operating system, an application program or instruction required for at least one function (such as a sound playback function, an image playback function, etc.), etc.
  • the memory 1109 can include a volatile memory or a non-volatile memory.
  • the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory.
  • the volatile memory may be a random access memory (RAM), a static random access memory (SRAM), a dynamic random access memory (DRAM), a synchronous dynamic random access memory (SDRAM), a double data rate synchronous dynamic random access memory (DDRSDRAM), an enhanced synchronous dynamic random access memory (ESDRAM), a synchronous link dynamic random access memory (SLDRAM) and a direct memory bus random access memory (DRRAM).
  • RAM random access memory
  • SRAM static random access memory
  • DRAM dynamic random access memory
  • SDRAM synchronous dynamic random access memory
  • DDRSDRAM double data rate synchronous dynamic random access memory
  • ESDRAM enhanced synchronous dynamic random access memory
  • SLDRAM synchronous link dynamic random access memory
  • DRRAM direct memory bus random access memory
  • the processor 1110 may include one or more processing units; optionally, the processor 1110 integrates an application processor and a modem processor, wherein the application processor mainly processes operations related to an operating system, a user interface, and application programs, and the modem processor mainly processes wireless communication signals, such as a baseband processor. It is understandable that the modem processor may not be integrated into the processor 1110.
  • the radio frequency unit 1101 is used to receive model information of a first model from a second communication device, and the processor 1110 is used to determine the first model and the associated model according to the model information of the first model.
  • the processor 1110 is used to determine the first model and the associated model according to the model information of the first model.
  • at least one second model of the invention supervising the effectiveness of the at least one second model according to the first model; or,
  • the radio frequency unit 1101 is used to send first information to the second communication device, where the first information is related to the first model or the at least one second model, and the first information is used by the second communication device to monitor the validity of the at least one second model based on the first model.
  • the first communication device may receive model information of the first model from the second communication device, and then determine the first model and at least one second model associated with the first model based on the model information, and use the first model to supervise the validity of the at least one second model.
  • the first communication device sends first information to the second communication device, and the first information is related to the first model or at least one second model, which can be used by the second communication device to supervise the validity of at least one second model based on the first model. In this way, by using the first model to supervise the validity of at least one second model, the problem of task failure caused by still using the second model to perform the reasoning task when the second model is invalid can be avoided.
  • the embodiment of the present application also provides a communication device, including a processor and a communication interface, the communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the steps of the method embodiment shown in Figures 3 and 4.
  • This communication device embodiment corresponds to the above-mentioned second communication device method embodiment, and each implementation process and implementation method of the above-mentioned method embodiment can be applied to the communication device embodiment, and can achieve the same technical effect.
  • the communication device 1200 includes: an antenna 121, a radio frequency device 122, a baseband device 123, a processor 124 and a memory 125.
  • the antenna 121 is connected to the radio frequency device 122.
  • the radio frequency device 122 receives information through the antenna 121 and sends the received information to the baseband device 123 for processing.
  • the baseband device 123 processes the information to be sent and sends it to the radio frequency device 122.
  • the radio frequency device 122 processes the received information and sends it out through the antenna 121.
  • the method executed by the second communication device in the above embodiment may be implemented in the baseband device 123, which includes a baseband processor.
  • the baseband device 123 may include, for example, at least one baseband board, on which a plurality of chips are arranged, as shown in FIG12 , wherein one of the chips is, for example, a baseband processor, which is connected to the memory 125 through a bus interface to call a program in the memory 125 and execute the second communication device operation shown in the above method embodiment.
  • the communication device may also include a network interface 126, which is, for example, a Common Public Radio Interface (CPRI).
  • a network interface 126 which is, for example, a Common Public Radio Interface (CPRI).
  • CPRI Common Public Radio Interface
  • the communication device 1200 of the embodiment of the present invention also includes: instructions or programs stored in the memory 125 and executable on the processor 124.
  • the processor 124 calls the instructions or programs in the memory 125 to execute the methods executed by the modules shown in Figure 9 and achieve the same technical effect. To avoid repetition, it will not be repeated here.
  • An embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored.
  • a program or instruction is stored.
  • the various processes of the above-mentioned model supervision method embodiment are implemented and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.
  • the processor is the processor in the terminal described in the above embodiment.
  • the readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.
  • the readable storage medium may be a non-transient readable storage medium.
  • An embodiment of the present application further provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the various processes of the above-mentioned model supervision method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.
  • the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.
  • the present application embodiment further provides a computer program/program product, wherein the computer program/program product is stored In the storage medium, the computer program/program product is executed by at least one processor to implement the various processes of the above-mentioned model supervision method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.
  • the above-mentioned embodiment method can be implemented by means of a computer software product plus a necessary general hardware platform, and of course, it can also be implemented by hardware.
  • the computer software product is stored in a storage medium (such as ROM, RAM, disk, CD, etc.), including several instructions to enable the terminal or network side device to execute the method described in each embodiment of the present application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Signal Processing (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

La présente demande, qui appartient au domaine technique des communications, divulgue un procédé de supervision de modèle et un dispositif de communication. Le procédé de supervision de modèle dans les modes de réalisation de la présente demande comprend : la réception, par un premier dispositif de communication, d'informations de modèle d'un premier modèle en provenance d'un second dispositif de communication ; selon les informations de modèle du premier modèle, la détermination, par le premier dispositif de communication, du premier modèle et d'au moins un second modèle associé au premier modèle ; et la supervision, par le premier dispositif de communication, de la validité du ou des seconds modèles selon le premier modèle ; ou l'envoi, par le premier dispositif de communication, de premières informations au second dispositif de communication, les premières informations étant associées au premier modèle ou au ou aux seconds modèles, et les premières informations étant utilisées par le second dispositif de communication pour superviser la validité du ou des seconds modèles selon le premier modèle.
PCT/CN2024/076407 2023-02-14 2024-02-06 Procédé de supervision de modèle et dispositif de communication Ceased WO2024169794A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202310123803.5A CN118504647A (zh) 2023-02-14 2023-02-14 模型监督方法及通信设备
CN202310123803.5 2023-02-14

Publications (1)

Publication Number Publication Date
WO2024169794A1 true WO2024169794A1 (fr) 2024-08-22

Family

ID=92231668

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2024/076407 Ceased WO2024169794A1 (fr) 2023-02-14 2024-02-06 Procédé de supervision de modèle et dispositif de communication

Country Status (2)

Country Link
CN (1) CN118504647A (fr)
WO (1) WO2024169794A1 (fr)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114787793A (zh) * 2020-02-18 2022-07-22 Oppo广东移动通信有限公司 网络模型的管理方法及建立或修改会话的方法、装置
CN114881129A (zh) * 2022-04-25 2022-08-09 北京百度网讯科技有限公司 一种模型训练方法、装置、电子设备及存储介质
CN115699209A (zh) * 2020-04-03 2023-02-03 普雷萨根私人有限公司 用于人工智能(ai)模型选择的方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114787793A (zh) * 2020-02-18 2022-07-22 Oppo广东移动通信有限公司 网络模型的管理方法及建立或修改会话的方法、装置
CN115699209A (zh) * 2020-04-03 2023-02-03 普雷萨根私人有限公司 用于人工智能(ai)模型选择的方法
CN114881129A (zh) * 2022-04-25 2022-08-09 北京百度网讯科技有限公司 一种模型训练方法、装置、电子设备及存储介质

Also Published As

Publication number Publication date
CN118504647A (zh) 2024-08-16

Similar Documents

Publication Publication Date Title
EP4422248A1 (fr) Procédé de demande de modèle, procédé de traitement de demande de modèle et dispositif associé
US20250148295A1 (en) Artificial intelligence request analysis method and apparatus and device
US20240346113A1 (en) Model construction method and apparatus, and communication device
US20240354659A1 (en) Method for Updating Model and Communication Device
CN116419267A (zh) 通信模型配置方法、装置和通信设备
WO2024067281A1 (fr) Procédé et appareil de traitement de modèle d'ia, et dispositif de communication
WO2024008111A1 (fr) Procédé et dispositif d'acquisition de données
WO2023040888A1 (fr) Procédé et appareil de transmission de données
CN117540200A (zh) 改进用于基于时间序列预测的多变量方法的准确性
WO2024169794A1 (fr) Procédé de supervision de modèle et dispositif de communication
WO2024093997A1 (fr) Procédé et appareil de détermination d'applicabilité de modèle, et dispositif de communication
WO2024083004A1 (fr) Procédé de configuration de modèle d'ia, terminal et dispositif côté réseau
JP2025507145A (ja) 通信ネットワークにおけるデータ処理方法、ネットワーク側機器及び可読記憶媒体
CN118282899A (zh) 功能性生命周期管理的模型监测方法、装置、通信设备、系统及存储介质
WO2023174325A1 (fr) Procédé et dispositif de traitement de modèle d'ia
CN118504646A (zh) 模型监督的方法、装置及通信设备
WO2025140432A1 (fr) Procédé et appareil de création de compte-rendu d'informations
WO2024169797A1 (fr) Procédé d'indication de modèle d'ia et dispositif de communication
WO2025092999A1 (fr) Procédé et appareil de supervision de performance de modèle, et dispositif
WO2024169796A1 (fr) Procédé et appareil de supervision de modèle, dispositif de communication
WO2024235140A1 (fr) Procédé et appareil de traitement de positionnement, et dispositif
WO2024153013A1 (fr) Procédé et appareil de transmission d'informations et dispositif de communication
WO2025124348A1 (fr) Procédés et appareils de supervision des performances pour fonction ia, dispositif côté réseau et dispositif terminal
WO2025092998A1 (fr) Procédé de transmission d'informations, appareil et dispositif
WO2024235043A1 (fr) Procédé et appareil d'acquisition d'informations, et dispositif de communication

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 24756167

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE