CN120113270A

CN120113270A - Model updating method, terminal, network device, communication device and storage medium

Info

Publication number: CN120113270A
Application number: CN202380011411.8A
Authority: CN
Inventors: 牟勤
Original assignee: Beijing Xiaomi Mobile Software Co Ltd
Current assignee: Beijing Xiaomi Mobile Software Co Ltd
Priority date: 2023-09-28
Filing date: 2023-09-28
Publication date: 2025-06-06
Also published as: WO2025065662A1

Abstract

The disclosure relates to the technical field of communication, and in particular relates to a model updating method, a terminal, a network device, a communication device and a storage medium. The model updating method comprises the steps that a terminal receives model parameters which are sent by network equipment and used for updating a first model, and the model parameters are applied to a model structure of the first model to obtain the updated first model. According to the method and the device for updating the first model, when the first model deployed on the terminal is updated, the updated first model can be obtained by transmitting the updated model parameters between the terminal and the network equipment and applying the updated model parameters to the model structure of the first model by the terminal, and the model structure of the first model does not need to be recompiled by the terminal, so that the requirement on the terminal performance is reduced, and the terminal power consumption is saved.

Description

Model updating method, terminal, network device, communication device and storage medium

Technical Field

The present disclosure relates to the field of communication technologies, and in particular, to a model updating method, a terminal, a network device, a communication device, and a storage medium.

Background

The wide application of 5G technology brings great changes to the aspects of people's life. According to the prospect of the international telecommunications union (International Telecommunication Union, ITU), 5G will penetrate into various areas of future society, building an omnidirectional information ecosystem with users as the center. The method comprises the steps of enabling a 5G user experience rate to reach 100 Mbit/s-1 Gbit/s, supporting extreme service experiences such as mobile virtual reality, enabling a 5G peak rate to reach 10 Gbit/s-20 Gbit/s, enabling a flow density to reach 10Mbit/s/m ², enabling mobile service flow to be increased by more than thousand times in the future, enabling a 5G connection number density to reach 100 tens of thousands/m ², enabling massive Internet of things equipment to be effectively supported, enabling a 5G transmission delay to reach millisecond order, enabling severe requirements of Internet of vehicles and industrial control to be met, enabling 5G to support a 500km/h moving speed, and enabling good user experience to be met in a high-speed railway environment. It follows that 5G, as a new infrastructure representative, will reconstruct the future informative society.

In recent years, artificial intelligence (ARTIFICIAL INTELLIGENCE, AI) technology has been breaking through in a number of fields. The continuous development in the fields of intelligent voice, computer vision and the like not only brings various colorful applications for the intelligent terminal, but also has wide application in a plurality of fields of education, traffic, home, medical treatment, retail, security and the like, brings convenience to life of people, and simultaneously promotes industry upgrading of various industries. AI technology is also accelerating cross-penetration with other discipline fields, and its development incorporates knowledge of different disciplines while also providing new directions and methods for development of different disciplines.

In 3gpp Release 18 phase, a research project on artificial intelligence technology in wireless air interface is set up in RAN 1. The project aims at researching how to introduce artificial intelligence technology into a wireless air interface and discussing how the artificial intelligence technology assists in improving the transmission technology of the wireless air interface.

For example, application cases of artificial intelligence may include, but are not limited to, AI-based Channel State Information (CSI) enhancement, AI-based beam management, AI-based positioning, and the like.

Disclosure of Invention

The embodiment of the disclosure provides a model updating method, a terminal, a network device, a communication device and a storage medium, so as to solve the technical problems in the related art.

According to a first aspect of an embodiment of the present disclosure, a method for updating a model is provided and executed by a terminal, where the method includes receiving a model parameter sent by a network device and used for updating a first model, where the first model is deployed on the terminal, and applying the model parameter to a model structure of the first model to obtain an updated first model.

According to a second aspect of the embodiments of the present disclosure, a method for updating a model is provided, which is executed by a network device, and includes sending, to a terminal, a model parameter for updating a first model, where the first model is deployed on the terminal, and applying the model parameter to a model structure of the first model to obtain an updated first model.

According to a third aspect of the embodiments of the present disclosure, there is provided a model updating apparatus, the apparatus including:

The receiving and transmitting module is used for receiving model parameters sent by the network equipment and used for updating a first model, wherein the first model is deployed on the terminal;

And the processing module is used for applying the model parameters to the model structure of the first model to obtain an updated first model.

According to a fourth aspect of embodiments of the present disclosure, there is provided a model updating apparatus, the apparatus including:

And the receiving and transmitting module is used for transmitting the model parameters for updating the first model to the terminal, wherein the first model is deployed on the terminal, and the model parameters are used for being applied to the model structure of the first model to obtain the updated first model.

According to a fifth aspect of embodiments of the present disclosure, a terminal is provided, which includes one or more processors, where the terminal is configured to perform the model updating method of the first aspect.

According to a sixth aspect of the embodiments of the present disclosure, a network device is proposed, which includes one or more processors, where the network device is configured to perform the model updating method of the second aspect.

According to a seventh aspect of embodiments of the present disclosure, a communication device is proposed, which comprises one or more processors, wherein the processors are configured to invoke instructions to cause the communication device to perform the model updating method of the first aspect and/or the model updating method of the second aspect.

According to an eighth aspect of the embodiments of the present disclosure, a communication system is provided, which includes a terminal configured to implement the model updating method of the first aspect, and a network device configured to implement the model updating method of the second aspect.

According to a ninth aspect of the embodiments of the present disclosure, a storage medium is provided, the storage medium storing instructions that, when executed on a communication device, cause the communication device to perform the model updating method of the first aspect and/or the model updating method of the second aspect.

According to the embodiment of the disclosure, the terminal may receive the model parameters sent by the network device and used for updating the first model, and apply the received model parameters to the model structure of the first model, so as to obtain the updated first model. Accordingly, when the first model deployed on the terminal is updated, the updated first model can be obtained by transmitting updated model parameters between the terminal and the network equipment and applying the updated model parameters to the model structure of the first model by the terminal, and the whole model file related to the updated first model is not required to be transmitted between the terminal and the network equipment, and the model structure of the first model is not required to be recompiled by the terminal, so that the transmission data volume between the terminal and the network equipment is reduced, the requirement on the terminal performance is also reduced, and the terminal power consumption is saved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings required for the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present disclosure, and other drawings may be obtained according to these drawings without inventive effort to a person of ordinary skill in the art.

Fig.1 is a schematic architecture diagram of a communication system shown in accordance with an embodiment of the present disclosure.

Fig. 2 is an interactive schematic diagram illustrating a model update method according to an embodiment of the present disclosure.

Fig.3 is a schematic flow chart diagram illustrating a method of model update according to an embodiment of the present disclosure.

Fig.4 is a schematic flow chart diagram illustrating another model update method according to an embodiment of the present disclosure.

Fig. 5 is a schematic block diagram of a model updating apparatus shown according to an embodiment of the present disclosure.

Fig. 6 is a schematic block diagram of another model updating apparatus shown according to an embodiment of the present disclosure.

Fig.7 is a schematic structural diagram of a communication device according to an embodiment of the present disclosure.

Fig. 8 is a schematic structural diagram of a chip according to an embodiment of the disclosure.

Detailed Description

The embodiment of the disclosure provides a model updating method, a device, a terminal, a network device, a communication device and a storage medium.

In a first aspect, an embodiment of the present disclosure proposes a method for updating a model, which is executed by a terminal, where the method includes receiving a model parameter sent by a network device and used for updating a first model, where the first model is deployed on the terminal, and applying the model parameter to a model structure of the first model to obtain an updated first model.

In the above embodiment, the terminal may receive the model parameters sent by the network device and used for updating the first model, and apply the received model parameters to the model structure of the first model, so as to obtain the updated first model. Accordingly, when the first model deployed on the terminal is updated, the updated first model can be obtained by transmitting updated model parameters between the terminal and the network equipment and applying the updated model parameters to the model structure of the first model by the terminal, and the whole model file related to the updated first model is not required to be transmitted between the terminal and the network equipment, and the model structure of the first model is not required to be recompiled by the terminal, so that the transmission data volume between the terminal and the network equipment is reduced, the requirement on the terminal performance is also reduced, and the terminal power consumption is saved.

With reference to some embodiments of the first aspect. In some embodiments, the method further comprises sending a query request for the first model to the network device, wherein the query request is used for determining whether the network device has model parameters for updating the first model, and receiving a query response sent by the network device for the first model, wherein the query response is used for indicating whether the network device has model parameters for updating the first model.

With reference to some embodiments of the first aspect. In some embodiments, the query request contains at least one of an identification of a model that the terminal desires to update and a structural identification of a model that the terminal desires to update.

With reference to some embodiments of the first aspect. In some embodiments, the method further comprises receiving an update notification for the first model sent by the network device, wherein the update notification is used to indicate that model parameters of the first model are updated.

With reference to some embodiments of the first aspect. In some embodiments, the terminal supports model parameter update capabilities.

With reference to some embodiments of the first aspect. In some embodiments, the method further comprises receiving configuration information sent by the network device for updating the model parameters of the first model, wherein the configuration information is used for assisting the terminal in determining whether to receive the model parameters sent by the network device for updating the first model.

With reference to some embodiments of the first aspect. In some embodiments, the configuration information includes at least one of a computational complexity of the model parameters, a memory space required for the model parameters, and a usage scenario for the model parameters.

With reference to some embodiments of the first aspect. In some embodiments, the method further comprises determining whether to receive the model parameters based on the network device presence of model parameters for updating the first model.

With reference to some embodiments of the first aspect. In some embodiments, the method further comprises the step of successfully operating the updated first model and sending first information to the network device, wherein the first information is used for indicating that the updating of the first model is completed.

With reference to some embodiments of the first aspect. In some embodiments, the network device is expected to assign a new model identification to the updated first model.

With reference to some embodiments of the first aspect. In some embodiments, the method further comprises the step that the updated first model cannot be successfully operated, and second information is sent to the network equipment, wherein the second information is used for indicating that the updating of the first model fails.

With reference to some embodiments of the first aspect. In some embodiments, the second information is further used to indicate a cause of failure of the first model update.

With reference to some embodiments of the first aspect. In some embodiments, in combination with some embodiments of the first aspect. In some embodiments, the method further comprises indicating to the network device a model structure of at least one model deployed on the terminal.

With reference to some embodiments of the first aspect. In some embodiments, the method for indicating the model structure of the at least one model deployed on the terminal to the network equipment comprises the steps of sending third information to the network equipment, wherein the third information comprises model identification of the at least one model, the network equipment stores a first corresponding relation between the model identification and the model structure, and the third information is used for determining the model structure corresponding to the model identification of the at least one model according to the first corresponding relation by the network equipment.

With reference to some embodiments of the first aspect. In some embodiments, the method for indicating the model structure of the at least one model deployed on the terminal to the network equipment comprises the steps of sending fourth information to the network equipment, wherein the fourth information comprises a model structure identifier of the at least one model, the network equipment stores a second corresponding relation between the model structure identifier and the model structure, and the fourth information is used for determining the model structure corresponding to the model structure identifier of the at least one model according to the second corresponding relation by the network equipment.

In a second aspect, an embodiment of the disclosure proposes a method for updating a model, which is executed by a network device, and includes sending, to a terminal, a model parameter for updating a first model, where the first model is deployed on the terminal, and applying the model parameter to a model structure of the first model to obtain an updated first model.

In the above embodiment, the network device may send the model parameters for updating the first model to the terminal, so that the terminal applies the received model parameters to the model structure of the first model. Accordingly, when the first model deployed on the terminal is updated, the updated first model can be obtained by transmitting updated model parameters between the terminal and the network equipment and applying the updated model parameters to the model structure of the first model by the terminal, and the whole model file related to the updated first model is not required to be transmitted between the terminal and the network equipment, and the model structure of the first model is not required to be recompiled by the terminal, so that the transmission data volume between the terminal and the network equipment is reduced, the requirement on the terminal performance is also reduced, and the terminal power consumption is saved.

With reference to some embodiments of the second aspect. In some embodiments, the method further comprises receiving a query request for the first model sent by the terminal, wherein the query request is used for determining whether the network device has model parameters for updating the first model, and sending a query response for the first model to the terminal, wherein the query response is used for indicating whether the network device has the model parameters for updating the first model.

With reference to some embodiments of the second aspect. In some embodiments, the query request contains at least one of an identification of a model that the terminal desires to update and a structural identification of a model that the terminal desires to update.

With reference to some embodiments of the second aspect. In some embodiments, the method further comprises sending an update notification to the terminal for the first model, wherein the update notification is used to indicate that model parameters of the first model are updated.

With reference to some embodiments of the second aspect. In some embodiments, the method further comprises determining whether the terminal supports model parameter update capabilities;

the sending of the update notification for the first model to the terminal includes sending the update notification for the first model to the terminal in response to the terminal supporting model parameter update capabilities.

With reference to some embodiments of the second aspect. In some embodiments, the method further comprises sending configuration information for updating the model parameters of the first model to the terminal, wherein the configuration information is used for assisting the terminal in determining whether to receive the model parameters for updating the first model sent by the network device.

With reference to some embodiments of the second aspect. In some embodiments, the configuration information includes at least one of a computational complexity of the model parameters, a memory space required for the model parameters, and a usage scenario for the model parameters.

With reference to some embodiments of the second aspect. In some embodiments, the method further comprises determining whether the terminal receives the model parameters.

With reference to some embodiments of the second aspect. In some embodiments, the method further comprises receiving first information sent by the terminal, wherein the first information is used for indicating that the first model updating is completed.

With reference to some embodiments of the second aspect. In some embodiments, the method further comprises assigning a new model identification to the updated first model based on the first model update completion.

With reference to some embodiments of the second aspect. In some embodiments, the method further comprises receiving second information sent by the terminal, wherein the second information is used for indicating that the first model update fails.

With reference to some embodiments of the second aspect. In some embodiments, the second information is further used to indicate a cause of failure of the first model update.

With reference to some embodiments of the second aspect. In some embodiments, the method further comprises determining a model structure of at least one model deployed on the terminal according to the indication of the terminal.

With reference to some embodiments of the second aspect. In some embodiments, the network device stores a first correspondence of model identifications and model structures;

The method comprises the steps of determining a model structure of at least one model deployed on a terminal according to an indication of the terminal, and determining the model structure corresponding to the model identifier of the at least one model according to the first corresponding relation and the third information.

With reference to some embodiments of the second aspect. In some embodiments, the network device stores a second correspondence of model structure identifications and model structures;

The method for indicating the model structure of the at least one model deployed on the terminal to the network equipment comprises the steps of receiving fourth information sent by the terminal, wherein the fourth information comprises a model structure identifier of the at least one model, and determining the model structure corresponding to the model structure identifier of the at least one model according to the second corresponding relation and the fourth information.

In a third aspect, an embodiment of the present disclosure proposes a model updating apparatus, the apparatus including:

In a fourth aspect, an embodiment of the present disclosure proposes a model updating apparatus, including:

In a fifth aspect, an embodiment of the disclosure proposes a terminal, including one or more processors, where the terminal is configured to perform the model updating method according to the first aspect or the optional embodiment of the first aspect.

In a sixth aspect, an embodiment of the disclosure proposes a network device, including one or more processors, where the network device is configured to perform the model updating method according to the second aspect and the optional embodiments of the second aspect.

In a seventh aspect, embodiments of the present disclosure provide a communication device comprising one or more processors and one or more memories for storing instructions, wherein the processors are configured to invoke the instructions to cause the communication device to perform the methods as described in the first and second aspects, and optional embodiments of the first and second aspects.

In an eighth aspect, an embodiment of the present disclosure proposes a communication system, which includes a terminal configured to perform the methods described in the first and second aspects, the optional embodiments of the first and second aspects, and a network device configured to perform the methods described in the first and second aspects, the optional embodiments of the first and second aspects.

In a ninth aspect, embodiments of the present disclosure provide a storage medium storing instructions that, when executed on a communications device, cause the communications device to perform a method as described in the first and second aspects, and optional embodiments of the first and second aspects.

In a tenth aspect, embodiments of the present disclosure propose a program product which, when executed by a communication device, causes the communication device to perform the method as described in the first and second aspects, and optional embodiments of the first and second aspects.

In an eleventh aspect, embodiments of the present disclosure propose a computer program which, when run on a computer, causes the computer to carry out the method as described in the first and second aspects, the optional embodiments of the first and second aspects.

It will be appreciated that the above-described terminals, network devices, communication systems, storage media, program products, computer programs are all adapted to perform the methods set forth in the embodiments of the present disclosure. Therefore, the advantages achieved by the method can be referred to as the advantages of the corresponding method, and will not be described herein.

The embodiment of the disclosure provides a model updating method, a terminal, a network device, a communication device and a storage medium. In some embodiments, terms such as a model updating method, a model parameter updating method, an information processing method, a communication method, and the like may be replaced with each other, terms such as a terminal, a network device, a model updating apparatus, a model parameter updating apparatus, a communication apparatus, and the like may be replaced with each other, and terms such as a model updating system, a model parameter updating system, a communication system, and the like may be replaced with each other.

The embodiments of the present disclosure are not intended to be exhaustive, but rather are exemplary of some embodiments and are not intended to limit the scope of the disclosure. In the case of no contradiction, each step in an embodiment may be implemented as an independent embodiment, and steps may be arbitrarily combined, for example, a scheme in which part of steps are removed in an embodiment may be implemented as an independent embodiment, the order of steps may be arbitrarily exchanged in an embodiment, further, alternative embodiments in an embodiment may be arbitrarily combined, further, embodiments may be arbitrarily combined, for example, part or all of steps of different embodiments may be arbitrarily combined, and an embodiment may be arbitrarily combined with alternative embodiments of other embodiments.

In the various embodiments of the disclosure, terms and/or descriptions of the various embodiments are consistent throughout the various embodiments and may be referenced to each other in the absence of any particular explanation or logic conflict, and features from different embodiments may be combined to form new embodiments in accordance with their inherent logic relationships.

The terminology used in the embodiments of the disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure.

In the presently disclosed embodiments, elements that are referred to in the singular, such as "a," "an," "the," "said," etc., may mean "one and only one," or "one or more," "at least one," etc., unless otherwise indicated.

For example, where an article (article) is used in translation, such as "a," "an," "the," etc., in english, a noun following the article may be understood as a singular expression or as a plural expression.

In the presently disclosed embodiments, "plurality" refers to two or more.

In some embodiments, terms such as "at least one of", "one or more of", "multiple of" and the like may be substituted for each other.

In some embodiments, the recitations of "A, B at least one of", "A and/or B", "in one case A, in another case B", "in one case A", "in another case B", etc., may include the following, in some embodiments A (A being performed independently of B), in some embodiments B (B being performed independently of A), in some embodiments A and B being selected for execution (A and B being selectively executed), in some embodiments A and B (both A and B being executed). Similar to the above when there are more branches such as A, B, C.

In some embodiments, the description modes such as A or B can comprise the following technical scheme, namely A (A is executed independently of B) in some embodiments, B (B is executed independently of A) in some embodiments, and A and B are selected to be executed (A and B are selectively executed) in some embodiments according to the situation. Similar to the above when there are more branches such as A, B, C.

The prefix words "first", "second", etc. in the embodiments of the present disclosure are only for distinguishing different description objects, and do not limit the location, order, priority, number, content, etc. of the description objects, and the statement of the description object refers to the claims or the description of the embodiment context, and should not constitute unnecessary limitations due to the use of the prefix words.

For example, if the description object is a "field", the ordinal words before the "field" in the "first field" and the "second field" do not limit the position or the order between the "fields", and the "first" and the "second" do not limit whether the "fields" modified by the "first" and the "second" are in the same message or not. For another example, describing an object as "level", ordinal words preceding "level" in "first level" and "second level" do not limit priority between "levels". As another example, the number of descriptive objects is not limited by ordinal words, and may be one or more, taking "first device" as an example, where the number of "devices" may be one or more. Further, the objects modified by different prefix words may be the same or different, for example, the description object is a "device", the "first device" and the "second device" may be the same device or different devices, the types of which may be the same or different, and, further, the description object is an "information", the "first information" and the "second information" may be the same information or different information, and the contents thereof may be the same or different.

In some embodiments, "comprising a", "containing a", "for indicating a", "carrying a", may be interpreted as carrying a directly, or as indicating a indirectly.

In some embodiments, the terms "responsive to" and "responsive to determining" and "in the case of" in the first place "," when "," when "and" if "and the like may be substituted for each other.

In some embodiments, terms "greater than", "greater than or equal to", "not less than", "more than or equal to", "not less than", "above" and the like may be interchanged, and terms "less than", "less than or equal to", "not greater than", "less than or equal to", "not more than", "below", "lower than or equal to", "no higher than", "below" and the like may be interchanged.

In some embodiments, devices and the like may be interpreted as physical or virtual, the names of which are not limited to those described in the embodiments

The terms "apparatus," "device," "circuit," "network element," "node," "function," "unit," "component," "system," "network," "chip," "system-on-chip," "entity," "subject," and the like may be used interchangeably.

In some embodiments, a "network" may be interpreted as an apparatus (e.g., access network device, core network device, etc.) contained in a network.

In some embodiments, "access network device (access network device, AN device)", "radio access network device (radio access network device, RAN DEVICE)", "Base Station (BS)", "radio base station (radio base station)", "fixed station (fixed station)", "node", "access point", "transmission point (transmission point, TP)", "Reception Point (RP)", "transmission reception point (transmission/reception point, TRP)", "panel", "antenna panel (ANTENNA PANEL)", "antenna array (ANTENNA ARRAY)", "cell", "macro cell", "micro cell", "sector", "femto cell", "cell group", "service cell", "carrier", "bandwidth (carrier component 62", "bandwidth component bandwidth (bandwidth of the carrier, etc.

In some embodiments, terms such as "terminal" (terminal) "," terminal device (TERMINAL DEVICE) "," User Equipment (UE) "," user terminal "(MS)", "Mobile Station (MS)", mobile Terminal (MT) ", subscriber station (subscriber station), mobile unit (mobile unit), subscriber unit (subscriber unit), wireless unit (wireless unit), remote unit (remote unit), mobile device (mobile device), wireless device (WIRELESS DEVICE), wireless communication device (wireless communication device), remote device (remote device), mobile subscriber station (mobile subscriber station), access terminal (ACCESS TERMINAL), mobile terminal (mobile terminal), wireless terminal (WIRELESS TERMINAL), remote terminal (remote terminal), handheld device (handset), user agent (user agent), mobile client (mobile client), and the like may be used interchangeably.

In some embodiments, the access network device, core network device, or network device may be replaced with a terminal. For example, the embodiments of the present disclosure may also be applied to a configuration in which an access network device, a core network device, or communication between a network device and a terminal is replaced with communication between a plurality of terminals (for example, device-to-device (D2D), vehicle-to-everything (V2X), or the like). In this case, the terminal may have all or part of the functions of the access network device. In addition, terms such as "uplink", "downlink", and the like may be replaced with terms corresponding to communication between terminals (e.g., "side)". For example, uplink channels, downlink channels, etc. may be replaced with side-uplink channels, uplink, downlink, etc. may be replaced with side-downlink channels.

In some embodiments, the terminal may be replaced with an access network device, a core network device, or a network device. In this case, the access network device, the core network device, or the network device may have all or part of the functions of the terminal.

In some embodiments, the acquisition of data, information, etc. may comply with laws and regulations of the country of locale.

In some embodiments, data, information, etc. may be obtained after user consent is obtained.

Furthermore, each element, each row, or each column in the tables of the embodiments of the present disclosure may be implemented as a separate embodiment, and any combination of elements, any rows, or any columns may also be implemented as a separate embodiment.

As shown in fig. 1, the communication system 100 includes a terminal (terminal) 101 and a network device 102, wherein the network device includes at least one of an access network device, a core network device (core network device).

In some embodiments, the terminal 101 includes at least one of, for example, a mobile phone (mobile phone), a wearable device, an internet of things device, a communication enabled car, a smart car, a tablet (Pad), a wireless transceiver enabled computer, a Virtual Reality (VR) terminal device, an augmented reality (augmented reality, AR) terminal device, a wireless terminal device in industrial control (industrial control), a wireless terminal device in unmanned-driving (self-driving), a wireless terminal device in teleoperation (remote medical surgery), a wireless terminal device in smart grid (SMART GRID), a wireless terminal device in transportation security (transportation safety), a wireless terminal device in smart city (SMART CITY), a wireless terminal device in smart home (smart home), but is not limited thereto.

In some embodiments, the access network device is, for example, a node or device that accesses a terminal to a wireless network, and the access network device may include at least one of an evolved NodeB (eNB), a next generation evolved NodeB (next generation eNB, ng-eNB), a next generation NodeB (next generation NodeB, gNB), a NodeB (node B, NB), a Home NodeB (HNB), a home NodeB (home evolved nodeB, heNB), a wireless backhaul device, a radio network controller (radio network controller, RNC), a base station controller (base station controller, BSC), a base transceiver station (base transceiver station, BTS), a baseband unit (BBU), a mobile switching center, a base station in a 6G communication system, an Open base station (Open RAN), a Cloud base station (Cloud RAN), a base station in other communication systems, an access node in a Wi-Fi system, but is not limited thereto.

In some embodiments, the core network device may be a device, including one or more network elements, or may be a plurality of devices or a device group, including all or part of one or more network elements. The network element may be virtual or physical. The core network comprises, for example, at least one of an evolved packet core (Evolved Packet Core, EPC), a 5G core network (5G Core Network,5GCN), a next generation core (Next Generation Core, NGC).

In some embodiments, the technical solutions of the present disclosure may be applied to an Open RAN architecture, where an access network device or an interface in an access network device according to the embodiments of the present disclosure may become an internal interface of the Open RAN, and flow and information interaction between these internal interfaces may be implemented by using software or a program.

In some embodiments, the access network device may be composed of a Central Unit (CU) and a Distributed Unit (DU), where the CU may also be referred to as a control unit (control unit), and the structure of the CU-DU may be used to split the protocol layers of the access network device, where functions of part of the protocol layers are centrally controlled by the CU, and functions of the rest of all the protocol layers are distributed in the DU, and the DU is centrally controlled by the CU, but is not limited thereto.

It may be understood that, the communication system described in the embodiments of the present disclosure is for more clearly describing the technical solutions of the embodiments of the present disclosure, and is not limited to the technical solutions provided in the embodiments of the present disclosure, and those skilled in the art can know that, with the evolution of the system architecture and the appearance of new service scenarios, the technical solutions provided in the embodiments of the present disclosure are applicable to similar technical problems.

The embodiments of the present disclosure described below may be applied to the communication system 100 shown in fig. 1, or a part of the main body, but are not limited thereto. The respective bodies shown in fig. 1 are examples, and the communication system may include all or part of the bodies in fig. 1, or may include other bodies than fig. 1, and the number and form of the respective bodies may be arbitrary, and the respective bodies may be physical or virtual, and the connection relationship between the respective bodies is examples, and the respective bodies may not be connected or may be connected, and the connection may be arbitrary, direct connection or indirect connection, or wired connection or wireless connection.

Embodiments of the present disclosure may be applied to long term evolution (Long Term Evolution, LTE), LTE-Advanced (LTE-a), LTE-Beyond (LTE-B), SUPER3G, IMT-Advanced, fourth generation mobile communication system (4th generation mobile communication system,4G)), fifth generation mobile communication system (5th generation mobile communication system,5G), 5G New air interface (New Radio, NR), future Radio access (Future Radio Access, FRA), new Radio access technology (New-Radio Access Technology, RAT), new Radio (New Radio, NR), new Radio access (New Radio access, NX), future generation Radio access (Future generation Radio access, FX), global System for Mobile communications (GSM (registered trademark)), CDMA2000, ultra mobile broadband (Ultra Mobile Broadband, UMB), IEEE802.11 (Wi-Fi (registered trademark)), IEEE802.16 (WiMAX (registered trademark)), IEEE802.20, ultra-WideBand (UWB), bluetooth (registered trademark)), land public mobile network (Public Land Mobile Network, PLMN) network, device-to-Device (D2D) system, machine-to-machine (Machine to Machine, M2M) system, internet of things (Internet of Things, ioT) system, vehicle-to-eventing (V2X), system utilizing other communication methods, next generation system extended based on them, and the like. In addition, a plurality of system combinations (e.g., LTE or a combination of LTE-a and 5G, etc.) may be applied.

As shown in fig. 2, the model updating method includes:

in step S201, the network device transmits model parameters for updating the first model to the terminal.

In step S202, the terminal applies the model parameters to the model structure of the first model.

In some embodiments, one or more models may be deployed on a terminal.

In some embodiments, the model structures of the various models deployed on the terminal may be the same or different.

In some embodiments, the first model may be at least one model deployed on the terminal. In one possible implementation, the first model may be one or more models that the terminal desires to update, among at least one model deployed on the terminal.

In some embodiments, the terminal may receive model parameters sent by the network device for updating the first model.

In some embodiments, the process of updating the deployed first model by the terminal may be triggered by the terminal or may be triggered by the network device.

In some embodiments, the updating of the deployed first model by the terminal may be triggered by the terminal. In this case, before step S201, the method further includes the terminal sending a query request for the first model to the network device, where the query request is used to determine whether the network device has model parameters for updating the first model, and receiving a query response sent by the network device for the first model, where the query response is used to indicate whether the network device has model parameters for updating the first model.

In some embodiments, the query request contains at least one of an identification of a model that the terminal desires to update and a structural identification of a model that the terminal desires to update.

In some embodiments, the network device stores at least one of a first correspondence of model identification and model structure, and a second correspondence of model structure identification and model structure.

In some embodiments, a network device may receive a query request sent by a terminal for the first model, where the query request is used to determine whether model parameters for updating the first model exist for the network device. The network device may send a query response to the terminal for the first model, wherein the query response is used to indicate whether the network device has model parameters for updating the first model.

In some embodiments, the network device may query whether model parameters exist for updating the first model based on the model identification and/or model structure identification contained in the query request for the first model.

In some embodiments, the updating of the deployed first model by the terminal may be triggered by the network device. In this case, before step S201, the method further includes the terminal receiving an update notification for the first model sent by the network device, where the update notification is used to indicate that the model parameters of the first model are updated.

In some embodiments, the terminal supports model parameter update capabilities.

In some embodiments, the network device may determine whether the terminal supports model parameter update capabilities, and in response to the terminal supporting model parameter update capabilities, the network device sends an update notification for the first model to the terminal.

In some possible implementations, the network device may request the terminal to report whether the terminal supports the model parameter updating capability, or the network device may determine whether the terminal supports the model parameter updating capability according to the terminal capability information reported when the terminal accesses the network, or the network device may acquire capability information supported by the terminal from the core network to determine whether the terminal supports the model parameter updating capability.

In some embodiments, the method further comprises the network device sending configuration information for updating the model parameters of the first model to the terminal, wherein the configuration information is used for assisting the terminal in determining whether to receive the model parameters for updating the first model sent by the network device.

In some embodiments, the method may further include the terminal receiving configuration information sent by the network device for updating the model parameters of the first model.

In some possible implementations, the configuration information may include at least one of a computational complexity of the model parameters, a memory space required for the model parameters, and a usage scenario for the model parameters.

In some embodiments, the terminal determines whether to receive the model parameters based on the network device having model parameters for updating the first model.

In some embodiments, the network device may determine whether the terminal receives model parameters for updating the first model.

In step S203, the terminal sends the first information or the second information to the network device.

In some embodiments, the updated first model is successfully operated, and the terminal sends first information to the network device, where the first information is used to indicate that updating of the first model is completed.

In some embodiments, the network device receives first information sent by the terminal, where the first information is used to indicate that the first model update is complete.

In one possible implementation, the network device is expected to assign a new model identification to the updated first model.

In one possible implementation manner, the network device allocates a new model identifier to the updated first model according to the completion of the update of the first model.

In some embodiments, the method further comprises that the updated first model cannot be successfully operated, and the terminal sends second information to the network device, wherein the second information is used for indicating that the updating of the first model fails.

In some embodiments, the network device receives second information sent by the terminal, where the second information is used to indicate that the first model update fails.

In a possible implementation manner, the second information is further used to indicate a cause of the failure of updating the first model.

In some embodiments, the method may further include the terminal indicating to the network device a model structure of at least one model deployed on the terminal.

In one possible implementation manner, the indicating the model structure of the at least one model deployed on the terminal to the network device specifically may include sending third information to the network device, where the third information includes a model identifier of the at least one model, the network device stores a first correspondence between model identifiers and model structures, and the third information is used for the network device to determine the model structure corresponding to the model identifier of the at least one model according to the first correspondence.

In the process of performing model identification (model identification), the terminal or a server associated with the terminal may report the model structure information of each model deployed on the terminal to the network device, the network device may allocate a unique model identifier, for example, a model ID, to each model, and further, each terminal may report the identifier of the model deployed on the terminal to the network device through third information, so that the network device determines the model structure supported by each terminal according to the first correspondence between the model identifier and the model structure.

In some embodiments, the network device stores a first correspondence of model identifications to model structures. The network equipment receives third information sent by the terminal, wherein the third information comprises a model identifier of the at least one model; and the network equipment determines a model structure corresponding to the model identification of the at least one model according to the first corresponding relation and the third information.

In another possible implementation manner, the indicating the model structure of the at least one model deployed on the terminal to the network device specifically may include sending fourth information to the network device, where the fourth information includes a model structure identifier of the at least one model, the network device stores a second correspondence between the model structure identifier and the model structure, and the fourth information is used for determining, by the network device, the model structure corresponding to the model structure identifier of the at least one model according to the second correspondence.

The network device may negotiate with the terminal or a server corresponding to the terminal, predefine a set of common model structures (common model structure), and assign unique model structure identifiers, such as structure IDs, to each common model structure; further, each terminal may report, to the network device, the model structure identifier of the model deployed on the terminal through the fourth information, so that the network device determines, according to the second correspondence between the model structure identifier and the model structure, the model structure supported by each terminal.

In some embodiments, the network device stores a second correspondence of model structure identifications to model structures. The network equipment receives fourth information sent by the terminal, wherein the fourth information comprises a model structure identifier of the at least one model, and the network equipment determines a model structure corresponding to the model structure identifier of the at least one model according to the second corresponding relation and the fourth information.

The communication method according to the embodiment of the present disclosure may include at least one of step S201 to step S203. For example, step S201 may be implemented as a separate embodiment, step S202 may be implemented as a separate embodiment, step S203 may be implemented as a separate embodiment, and step s201+s202 may be implemented as a separate embodiment, but is not limited thereto.

In some embodiments, steps S201, S202, S203 may be performed in exchange order or simultaneously.

In some embodiments, step S201 is optional, and one or more of these steps may be omitted or replaced in different embodiments.

In some embodiments, step S202 is optional, and one or more of these steps may be omitted or replaced in different embodiments.

In some embodiments, step S203 is optional, and one or more of these steps may be omitted or replaced in different embodiments.

In some embodiments, reference may be made to other alternative embodiments described before or after the description corresponding to fig. 2.

In some embodiments, an artificial intelligence model (ARTIFICIAL INTELLIGENCE model, AI model, hereinafter simply "model") may be trained by the network side, then transmitted from the network side to the terminal, and then applied by the terminal. After receiving the model, the terminal needs to perform corresponding processing on the model, for example, performing processing such as compiling and quantization on the model. The compiling process generally refers to converting a model file received by a terminal into executable code that can run on the terminal using a specific compiler, tool chain, dependency library, and the like. Quantization generally refers to a process of converting model parameters represented by floating point numbers into lower integer or fixed point numbers to reduce memory usage and computational overhead of the model.

In some embodiments, when the model needs to be updated, the new data set can be used for retraining the model at the network side, the model file corresponding to the updated model is regenerated, the model file corresponding to the updated model is transmitted to the terminal, and further, the terminal can compile, quantize and the like the received model file to obtain the updated model. In the above embodiment, when the terminal updates the model, additional processing power is required for the processing of compiling, quantizing, and the like of the model.

In some embodiments, to reduce the requirements on the processing power of the terminal, only model parameters corresponding to the model in which the terminal has been deployed may be considered for transmission when updating the model in which the terminal has been deployed. However, there are still some technical problems at present, for example, how to interact specifically between the terminal and the network device, how the network device knows the model structure supported by each terminal, etc.

In a first aspect, embodiments of the present disclosure propose a model update method. Fig. 3 is a schematic flow chart diagram illustrating a method of model update according to an embodiment of the present disclosure. The model updating method shown in the present embodiment may be executed by a terminal.

In some embodiments, one or more models may be deployed on a terminal. For example, model #1, model #2, and model #3 may be deployed on a terminal.

In some embodiments, the model structures of the various models deployed on the terminal may be the same or different. For example, the model structure of the model #1 deployed on the terminal is structure #1, and the model structures of the model #2 and the model #3 deployed on the terminal are structure #2.

In some embodiments, the first model may be at least one model deployed on the terminal. In one possible implementation, the first model may be one or more models that the terminal desires to update, among at least one model deployed on the terminal. In order to facilitate a better understanding of the present disclosure by those skilled in the art, the present disclosure is exemplarily described below using the first model as model # 1.

As shown in fig.3, the model updating method may include the steps of:

In step S301, model parameters sent by a network device for updating a first model are received, where the first model is deployed on a terminal.

In step S302, the model parameters are applied to the model structure of the first model to obtain an updated first model.

For example, the terminal is deployed with the model #1, and the terminal can receive the model parameters sent by the network device and used for updating the model #1, and further, the terminal can apply the received model parameters to the structure #1 to obtain the updated model #1.

It should be noted that the embodiment shown in fig. 3 may be implemented independently or in combination with at least one other embodiment in the disclosure, and specifically may be selected as needed, which is not limited by the disclosure.

Wherein a model structure (model structure) may define the overall framework and components of the model. The model structure may be used to determine the basic computational unit and data processing flow of the model. The model structure may describe the hierarchy of the model, the manner of connection, the data flow path, and so on. For example, the model structure may include a fully connected neural network (Fully Connected Neural Network), a convolutional neural network (Convolutional Neural Network), a recurrent neural network (Recurrent Neural Network), and the like.

The model structure to which the present disclosure relates may also be described as a model architecture or model architecture, to which the present disclosure is not limited.

The model parameters (model parameters) refer to variables that need to be optimized in the model training process. Model parameters may be used to adjust the behavior of the model to enable the model to efficiently model and predict input data. For example, model parameters of a neural network model may generally include weights (weights) that may be used to control the connection strength between neurons and biases (bias) that may be used to control the offset of the activation function. Model parameters are obtained by iterative computation using an optimization algorithm on the training data.

It should be noted that, in the embodiments related to the present disclosure, the first model is already deployed on the terminal, and it may be considered that, in the deployment process, the terminal has already translated the model structure of the first model, and the initial model parameters are applied to the model structure of the first model. Further, in step S302, the updated model parameters may be applied to the model structure of the first model. Wherein the application of the model parameters to the model structure may also be described as loading the model parameters to the model structure, which is not limiting in this disclosure.

In some embodiments, the terminal may determine, from an indication of the network device, whether the network device has model parameters for updating the first model. The network device may be configured to update the model parameters of the first model, and it may be considered that after the terminal applies the initial model parameters of the first model to the model structure of the first model, the model parameters of the first model maintained by the network device are updated, and the terminal has available updates of the model parameters of the first model.

In a further embodiment, the terminal may receive the model parameters for updating the first model sent by the network device if it is determined that the network device has model parameters for updating the first model. In one possible implementation manner, after determining that the network device has the model parameters for updating the first model, the terminal may determine whether to receive the model parameters first, and if the terminal determines to receive the model parameters, then receive the model parameters sent by the network device for updating the first model.

Optionally, the network device sends model parameters for updating the first model to the terminal, wherein the first model is deployed on the terminal, and the model parameters are used for being applied to a model structure of the first model to obtain an updated first model.

In some embodiments, the updating of the deployed first model by the terminal may be triggered by the terminal. In this case, before step S301, the method further comprises sending a query request for the first model to the network device, wherein the query request is used for determining whether the network device has model parameters for updating the first model, and receiving a query response sent by the network device for the first model, wherein the query response is used for indicating whether the network device has model parameters for updating the first model.

For example, the terminal may send a query request for model #1 to the network device to cause the network device to determine whether model parameters for updating model #1 exist, further, the terminal may receive a query response sent by the network device for model #1, wherein the terminal may determine that model parameters of model #1 exist for updating model #1 based on the query response to indicate that model parameters for updating model #1 exist for the network device, or the terminal may determine that model parameters of model #1 do not exist for updating model #1 based on the query response to indicate that model parameters for updating model #1 do not exist for the network device.

Optionally, the network device stores at least one of a first correspondence of the model identification and the model structure and a second correspondence of the model structure identification and the model structure.

For example, the terminal desires to update the model #1, and the terminal may send a query request for the model #1 to the network device, where the query request may include a model id=1 of the model #1, so that after the network device receives the query request, according to the model id=1 and a first correspondence relationship between the model ID and the model structure included in the query request, it determines that the terminal supports a structure #1 corresponding to the model id=1, and determines that the terminal desires to update model parameters related to the structure # 1.

For another example, the terminal desires to update the model #1, and the terminal may send a query request for the model #1 to the network device, where the query request may include a structure id=1 of the structure #1 of the model #1, so that after the network device receives the query request, it determines that the structure #1 corresponding to the model id=1 is supported by the terminal according to the structure id=1 and the second correspondence between the model ID and the model structure included in the query request, and determines that the terminal desires to update the model parameters related to the structure # 1.

In the above embodiment, when the terminal desires to update the deployed first model, the terminal may actively query the network device whether there is an available update for the first model, and further may obtain the model parameters for updating the first model from the network device, so that the terminal may update the first model as required, and save terminal power consumption.

Optionally, the network device may receive a query request for the first model sent by the terminal, where the query request is used to determine whether the network device has model parameters for updating the first model. The network device may send a query response to the terminal for the first model, wherein the query response is used to indicate whether the network device has model parameters for updating the first model.

Optionally, the network device may query whether model parameters for updating the first model exist according to the model identifier and/or the model structure identifier included in the query request for the first model.

In some embodiments, the updating of the deployed first model by the terminal may be triggered by the network device. In this case, the method further comprises, prior to step S301, receiving an update notification for the first model sent by the network device, where the update notification is used to indicate that the model parameters of the first model are updated.

For example, when the network device determines that the model parameters of the model #1 are updated, the network device sends an update notification for the model #1 to the terminal, and in response to receiving the update notification for the model #1 sent by the network device, the terminal can determine that there is an available update for the model parameters of the model # 1.

In some embodiments, the terminal supports model parameter update capabilities.

Alternatively, the network device may determine whether the terminal supports model parameter update capabilities, and in response to the terminal supporting model parameter update capabilities, the network device sends an update notification for the first model to the terminal.

In the above embodiment, when the network device determines that there is an available update for the first model, the terminal may be actively notified, so that the terminal obtains, from the network device, the model parameters for updating the first model as soon as possible, thereby improving the updating efficiency of the terminal on the first model.

In some embodiments, before step S301, the method may further include receiving configuration information sent by the network device for updating the model parameters of the first model, where the configuration information is used to assist the terminal in determining whether to receive the model parameters sent by the network device for updating the first model.

For example, the terminal determines that the network device has model parameters for updating model #1, and the terminal may receive configuration information for updating model #1 sent by the network device, and determine whether to receive model parameters for updating model #1 sent by the network device according to the received configuration information.

In some embodiments, prior to step S301, the method further comprises determining whether to receive model parameters for updating the first model based on the network device presence of the model parameters.

For example, the terminal may determine that the network device has model parameters for updating model #1 based on a query response to model #1 or an update notification to model #1, further, the terminal may determine to receive model parameters for updating model #1 sent by the network device, or the terminal may determine not to receive model parameters for updating model #1 sent by the network device.

For another example, the terminal may determine that the network device has the model parameters for updating the model #1 according to the query response to the model #1 or the update notification to the model #1, and the terminal may receive configuration information for updating the model parameters of the model #1 transmitted by the network device to learn information about the calculation complexity of the model parameters, the storage space required for the model parameters, the usage scenario of the model parameters, and the like, and further determine whether to receive the model parameters for updating the model #1 transmitted by the network device according to the configuration information.

The terminal determines that the network device is not expected to send the model parameter for updating the first model to the terminal, or the network device sends the model parameter for updating the first model to the terminal and the terminal refuses to receive the model parameter for updating the first model sent by the network device, or the network device sends the model parameter for updating the first model to the terminal and the terminal does not apply the received model parameter to the model structure of the first model after receiving the model parameter for updating the first model sent by the network device.

In the above embodiment, when the terminal determines that the network device has available update for the first model, it may determine whether to update the deployed first model by itself, or may combine configuration information of the updated model parameters, and determine whether to update the deployed first model by itself, so as to improve flexibility of updating the model by the terminal.

Alternatively, the network device may determine whether the terminal receives model parameters for updating the first model.

In some embodiments, after step S302, the method further includes, after the updated first model is successfully executed, sending first information to the network device, where the first information is used to indicate that the updating of the first model is completed.

Optionally, the network device receives first information sent by the terminal, where the first information is used to indicate that the first model update is completed.

Optionally, the network device allocates a new model identifier to the updated first model according to the completion of the updating of the first model.

For example, the terminal is deployed with a model #1, the terminal can receive model parameters sent by the network device for updating the model #1, further, the terminal can apply the received model parameters to the structure #1 to obtain an updated model #1, if the updated model #1 runs successfully, the terminal can send first information to the network device to indicate to the network device that the update of the model #1 deployed on the terminal is completed, and further, the network device can assign a new model identifier of "model id=4" to the updated model # 1.

In the above embodiment, the terminal determines that the network device has available update of the model parameters for the first model, and after the update is completed, the terminal may feed back to the network device through the first information, so that the network device knows that the terminal has completed the update for the first model deployed by itself. Further, to implement version control, the network device may also assign a new model identification to the updated first model to distinguish which version of the first model is currently deployed on the terminal.

In some embodiments, after step S302, the method further includes sending second information to the network device, where the second information is used to indicate that the first model update fails, where the updated first model fails to operate successfully.

Optionally, the network device receives second information sent by the terminal, where the second information is used to indicate that the first model update fails.

In a possible implementation manner, the second information is further used to indicate a cause of the failure of updating the first model. Among them, the present disclosure is not particularly limited to this as to the specific cause of update failure.

For example, the terminal is deployed with a model #1, and the terminal can receive model parameters sent by the network device for updating the model #1, further, the terminal can apply the received model parameters to the structure #1 to obtain an updated model #1, and if the updated model #1 cannot run successfully, the terminal can send second information to the network device to indicate to the network device that the update of the model #1 deployed on the terminal failed, and indicate to the network device that the update of the model #1 deployed on the terminal failed.

In the above embodiment, the terminal determines that the network device has available update for the model parameters of the first model, and after the update fails, the terminal may feed back to the network device through the second information, so that the network device knows that the update of the first model deployed by the terminal fails. Further, the terminal can also feed back the reason of the update failure to the network device through the second information, so that the network device can acquire the reason of the update failure of the terminal to the first model deployed on the terminal, and the terminal is beneficial to the subsequent retrying of the update of the first model deployed on the terminal.

In the related art, it is not known by the network device which model structures are supported by each terminal, so when the network device determines that the model parameters related to a certain model structure are updated, it cannot be determined for which terminals there are available updates of the model parameters.

In some embodiments, the method may further include indicating to the network device a model structure of at least one model deployed on the terminal.

For example, the model #1, model #2, and model #3 are deployed on the terminal, and the terminal may indicate to the network device that the model structure of the model #1 deployed on the terminal is structure #1, and the model structures of the model #2 and model #3 deployed on the terminal are structure #2.

For example, the third information sent by the terminal a to the network device includes a model identification "model id=1" of the model #1, a model identification "model id=2" of the model #2, and a model identification "model id=3" of the model #3, and the network device may determine that the model structures supported by the terminal a include structures #1 and #2 according to the first correspondence and the third information sent by the terminal a. The third information sent by the terminal B to the network device includes a model id=1 of the model #1, and the network device may determine that the model structure supported by the terminal B includes structure #1 according to the first correspondence and the third information sent by the terminal B.

Optionally, the network device stores a first correspondence of model identifications and model structures. The network equipment receives third information sent by the terminal, wherein the third information comprises a model identifier of the at least one model; and the network equipment determines a model structure corresponding to the model identification of the at least one model according to the first corresponding relation and the third information.

For example, the fourth information transmitted by the terminal A to the network device includes a model structure identification of "structure ID=1" of structure #1 of the model #1 and a model structure identification of "structure ID=2" of structure #2 of the model #2 or the model #3, and the network device may determine that the model structures supported by the terminal A include structure #1 and structure #2 based on the second correspondence and the fourth information transmitted by the terminal A. The fourth information sent by the terminal B to the network device includes a structure id=1 of the model #1 of the structure #1, and the network device may determine that the model structure supported by the terminal B includes the structure #1 according to the second correspondence and the fourth information sent by the terminal B.

Optionally, the network device stores a second correspondence between the model structure identifier and the model structure. The network equipment receives fourth information sent by the terminal, wherein the fourth information comprises a model structure identifier of the at least one model, and the network equipment determines a model structure corresponding to the model structure identifier of the at least one model according to the second corresponding relation and the fourth information.

In the above embodiment, the terminal may indicate, to the network device, the model structure of at least one model deployed on the terminal through the third information or the fourth information, so that the network device knows which model structures supported by each terminal are, and thus is beneficial to determining, by the network device, for which terminals, there is an available update of the model parameters when the model parameters related to a certain model structure are updated. For example, the network device may actively notify terminal a and terminal B when it determines that there is an update of the model parameters related to structure #1, and may actively notify terminal a without notifying terminal B when it determines that there is an update of the model parameters related to structure # 2.

In a second aspect, embodiments of the present disclosure propose a model update method. Fig. 4 is a schematic flow chart diagram illustrating another model update method according to an embodiment of the present disclosure. The model updating method shown in the present embodiment may be performed by a network device.

As shown in fig.4, the model updating method may include the steps of:

In step S401, a model parameter for updating a first model is sent to a terminal, where the first model is deployed on the terminal, and the model parameter is used to be applied to a model structure of the first model to obtain an updated first model.

For example, the network device may send model parameters for updating model #1 to the terminal so that the terminal applies the received model parameters to structure #1 to obtain updated model #1.

It should be noted that the embodiment shown in fig. 4 may be implemented independently or in combination with at least one other embodiment in the disclosure, and specifically may be selected as needed, which is not limited by the disclosure.

According to embodiments of the present disclosure, the network device may send the model parameters for updating the first model to the terminal, so that the terminal applies the received model parameters to the model structure of the first model. Accordingly, when the first model deployed on the terminal is updated, the updated first model can be obtained by transmitting updated model parameters between the terminal and the network equipment and applying the updated model parameters to the model structure of the first model by the terminal, and the whole model file related to the updated first model is not required to be transmitted between the terminal and the network equipment, and the model structure of the first model is not required to be recompiled by the terminal, so that the transmission data volume between the terminal and the network equipment is reduced, the requirement on the terminal performance is also reduced, and the terminal power consumption is saved.

Optionally, the terminal may receive a model parameter sent by the network device and used for updating the first model, and further, the terminal may apply the model parameter to a model structure of the first model to obtain an updated first model.

In some embodiments, the updating of the deployed first model by the terminal may be triggered by the terminal. In this case, before step S401, the method further includes receiving a query request for the first model sent by the terminal, where the query request is used to determine whether the network device has model parameters for updating the first model, and sending a query response for the first model to the terminal, where the query response is used to indicate whether the network device has model parameters for updating the first model.

For example, the network device may receive a query request sent by the terminal for model #1 and determine whether model parameters for updating model #1 exist, further, the network device may send a query response to the terminal for model #1, wherein the query response is used to indicate that the network device exists for updating model #1 model parameters so that the terminal determines that model #1 model parameters have available updates, or the query response is used to indicate that the network device does not exist for updating model #1 model parameters so that the terminal determines that model #1 model parameters do not have available updates

In some possible implementations, the network device stores at least one of a first correspondence of the model identification with the model structure and a second correspondence of the model structure identification with the model structure. In this case, the network device may query whether model parameters for updating the first model exist according to the model identification and/or the model structure identification contained in the query request for the first model.

For example, the terminal desires to update the model #1, the terminal may send a query request for the model #1 to the network device, where the query request may include a model id=1 of the model #1, and after the network device receives the query request, the network device may determine that the terminal supports a structure #1 corresponding to the model id=1 according to the model id=1 and the first correspondence between the model ID and the model structure included in the query request, and determine that the terminal desires to update the model parameters related to the structure # 1.

For another example, the terminal desires to update the model #1, the terminal may send a query request for the model #1 to the network device, where the query request may include a structure id=1 of the structure #1 of the model #1, and after the network device receives the query request, it may determine that the terminal supports the structure #1 corresponding to the structure id=1 of the model and determine that the terminal desires to update the model parameters related to the structure #1 according to the structure id=1 of the model and the second correspondence between the structure ID and the model structure included in the query request.

Optionally, the terminal sends a query request for the first model to the network device, and further, the terminal receives a query response for the first model sent by the network device.

In some embodiments, the updating of the deployed first model by the terminal may be triggered by the network device. In this case, before step S401, the method further comprises sending an update notification for the first model to the terminal, wherein the update notification is used to indicate that the model parameters of the first model are updated.

In some embodiments, the method further comprises determining whether the terminal supports model parameter update capabilities. In this case, the sending of the update notification for the first model to the terminal includes sending the update notification for the first model to the terminal in response to the terminal supporting model parameter update capabilities.

In some embodiments, before step S401, the method further comprises sending configuration information for updating the model parameters of the first model to the terminal, wherein the configuration information is used for assisting the terminal in determining whether to receive the model parameters for updating the first model sent by the network device.

For example, the network device may transmit configuration information for updating the model parameters of the model #1 to the terminal, so that the terminal determines whether to receive the model parameters for updating the model #1 transmitted by the network device according to the configuration information.

In some embodiments, prior to step S401, the method further comprises determining whether the terminal receives the model parameters.

In one possible implementation, the terminal may determine whether to receive the model parameters sent by the network device for updating the first model according to whether the network device has available updates for the model parameters of the first model and/or configuration information of the model parameters of the first model sent by the network device.

In some embodiments, after step S401, the method further includes receiving first information sent by the terminal, where the first information is used to indicate that the first model update is complete.

Optionally, the updated first model runs successfully, and the terminal sends first information to the network device, wherein the first information is used for indicating that the updating of the first model is completed.

In one possible implementation, the method further comprises assigning a new model identification to the updated first model based on the first model update completion.

In some embodiments, after step S401, the method further includes receiving second information sent by the terminal, where the second information is used to indicate that the first model update fails.

Optionally, the updated first model cannot be successfully operated, and second information is sent to the network device, where the second information is used to indicate that the updating of the first model fails.

In some embodiments, the method further comprises determining a model structure of at least one model deployed on the terminal according to the indication of the terminal.

For example, the model #1, model #2, and model #3 are deployed on the terminal, the terminal may indicate to the network device that the model structure of the model #1 deployed on the terminal is structure #1 and the model structures of the model #2 and model #3 deployed on the terminal are structure #2, and the network device may determine that the model structure of the model #1 deployed on the terminal is structure #1 and the model structures of the model #2 and model #3 deployed on the terminal are structure #2 according to the indication of the terminal.

In one possible implementation, the network device stores a first correspondence of model identifications and model structures. In this case, the determining the model structure of the at least one model deployed on the terminal according to the indication of the terminal includes receiving third information sent by the terminal, wherein the third information includes a model identifier of the at least one model, and determining the model structure corresponding to the model identifier of the at least one model according to the first correspondence and the third information.

Optionally, the terminal sends the third information to the network device.

In another possible implementation manner, the network device stores a second correspondence between model structure identifiers and model structures. In this case, the indicating the model structure of the at least one model deployed on the terminal to the network device includes receiving fourth information sent by the terminal, where the fourth information includes a model structure identifier of the at least one model, and determining a model structure corresponding to the model structure identifier of the at least one model according to the second correspondence and the fourth information.

Optionally, the terminal sends the fourth information to the network device.

In some embodiments, the names of information and the like are not limited to the names described in the embodiments, and terms such as "information", "message", "signal", "signaling", "report", "configuration", "instruction", "command", "channel", "parameter", "field", "symbol", "codebook", "code word", "code point", "codepoint", "bit", "data", "program", "chip", and the like may be replaced with each other.

In some embodiments, terms such as "time of day," "point of time," "time location," and the like may be interchanged, and terms such as "duration," "period," "time window," "time," and the like may be interchanged.

In some embodiments, terms of "component carrier (component carrier, CC)", "cell", "frequency carrier (frequency carrier)", "carrier frequency (carrier frequency)", and the like may be interchanged.

In some embodiments, "acquire," "obtain," "receive," "transmit," "bi-directional transmit," "send and/or receive" may be used interchangeably and may be interpreted as receiving from other principals, acquiring from protocols, acquiring from higher layers, processing itself, autonomous implementation, etc.

In some embodiments, terms such as "send," "transmit," "report," "send," "transmit," "bi-directional," "send and/or receive," and the like may be used interchangeably.

The present disclosure also provides an embodiment of a model updating apparatus, corresponding to the foregoing embodiment of the model updating method.

Fig. 5 is a schematic block diagram of a model updating apparatus shown according to an embodiment of the present disclosure. As shown in fig. 5, the model updating apparatus 500 includes a transceiver module 501 and a processing module 502.

In some embodiments, the transceiver module is configured to receive a model parameter sent by a network device and used for updating a first model, where the first model is deployed on a terminal;

In some embodiments, the transceiver module is further configured to send a query request for the first model to the network device, where the query request is used to determine whether model parameters exist for updating the first model by the network device;

the transceiver module is further configured to receive a query response sent by the network device for the first model, where the query response is used to indicate whether the network device has a model parameter for updating the first model.

In some embodiments, the transceiver module is further configured to receive an update notification sent by the network device for the first model, where the update notification is used to indicate that a model parameter of the first model is updated.

In some embodiments, the terminal supports model parameter update capabilities.

In some embodiments, the apparatus further comprises:

The transceiver module is further configured to receive configuration information sent by the network device and used for updating the model parameters of the first model, where the configuration information is used to assist the terminal in determining whether to receive the model parameters sent by the network device and used for updating the first model.

In some embodiments, the configuration information includes at least one of a computational complexity of the model parameters, a memory space required for the model parameters, and a usage scenario for the model parameters.

In some embodiments, the processing module is further configured to determine whether to receive the model parameters based on the network device existence of model parameters for updating the first model.

In some embodiments, the transceiver module is further configured to send first information to the network device after the updated first model is successfully operated, where the first information is used to indicate that updating of the first model is completed.

In some embodiments, the network device is expected to assign a new model identification to the updated first model.

In some embodiments, the transceiver module is further configured to send second information to the network device, where the second information is used to indicate that the first model fails to be updated.

In some embodiments, the second information is further used to indicate a cause of failure of the first model update.

In some embodiments, the transceiver module is further configured to indicate to the network device a model structure of at least one model deployed on the terminal.

In some embodiments, the transceiver module is configured to send third information to the network device, where the third information includes a model identifier of the at least one model, the network device stores a first correspondence between model identifiers and model structures, and the third information is used by the network device to determine a model structure corresponding to the model identifier of the at least one model according to the first correspondence.

In some embodiments, the transceiver module is configured to send fourth information to the network device, where the fourth information includes a model structure identifier of the at least one model, the network device stores a second correspondence between model structure identifiers and model structures, and the fourth information is used by the network device to determine a model structure corresponding to the model structure identifier of the at least one model according to the second correspondence.

Fig. 6 is a schematic block diagram of another model updating apparatus shown according to an embodiment of the present disclosure. As shown in fig. 6, the model updating apparatus 600 includes a second transceiver module 601.

In some embodiments, the transceiver module is configured to send, to a terminal, model parameters for updating a first model, where the first model is deployed on the terminal, and the model parameters are configured to be applied to a model structure of the first model to obtain an updated first model.

In some embodiments, the transceiver module is further configured to receive a query request for the first model sent by the terminal, where the query request is used to determine whether the network device has model parameters for updating the first model;

The transceiver module is further configured to send a query response to the terminal for the first model, where the query response is used to indicate whether the network device has model parameters for updating the first model.

In some embodiments, the transceiver module is further configured to send an update notification for the first model to the terminal, where the update notification is used to indicate that a model parameter of the first model is updated.

In some embodiments, the processing module is configured to determine whether the terminal supports a model parameter update capability;

The transceiver module is configured to send an update notification for the first model to the terminal in response to the terminal supporting a model parameter update capability.

In some embodiments, the transceiver module is further configured to send configuration information for updating the model parameters of the first model to the terminal, where the configuration information is used to assist the terminal in determining whether to receive the model parameters for updating the first model sent by the network device.

In some embodiments, the processing module is further configured to determine whether the terminal receives the model parameters.

In some embodiments, the transceiver module is further configured to receive first information sent by the terminal, where the first information is used to indicate that the updating of the first model is completed.

In some embodiments, the processing module is further configured to allocate a new model identifier to the updated first model according to the completion of the update of the first model.

In some embodiments, the transceiver module is further configured to receive second information sent by the terminal, where the second information is used to indicate that the first model update fails.

In some embodiments, the processing module is further configured to determine a model structure of at least one model deployed on the terminal according to the indication of the terminal.

In some embodiments, the network device stores a first correspondence of model identifications and model structures;

The transceiver module is further configured to receive third information sent by the terminal, where the third information includes a model identifier of the at least one model;

The processing module is further configured to determine a model structure corresponding to a model identifier of the at least one model according to the first correspondence and the third information.

In some embodiments, the network device stores a second correspondence of model structure identifications and model structures;

The transceiver module is further configured to receive fourth information sent by the terminal, where the fourth information includes a model structure identifier of the at least one model;

The processing module is further configured to determine, according to the second correspondence and the fourth information, a model structure corresponding to a model structure identifier of the at least one model.

It should be noted that, the modules included in the model updating apparatus 500 and/or the model updating apparatus 600 are not limited to the modules described in the above embodiments, and may include other modules, such as a storage module, a display module, and the like.

For the device embodiments, reference is made to the description of the method embodiments for the relevant points, since they essentially correspond to the method embodiments. The apparatus embodiments described above are merely illustrative, wherein the modules illustrated as separate components may or may not be physically separate, and the components shown as modules may or may not be physical, i.e., may be located in one place, or may be distributed over a plurality of network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.

Embodiments of the present disclosure further provide a terminal, including one or more processors, where the processors are configured to invoke instructions to cause the terminal to execute the information transceiving method according to the first aspect, and the optional embodiments of the first aspect.

Embodiments of the present disclosure also provide a network device, including one or more processors, where the processors are configured to invoke instructions to cause the network device to execute the information transceiving method according to the second aspect, and the optional embodiments of the second aspect.

Embodiments of the present disclosure also provide a communication device comprising one or more processors, wherein the processors are configured to invoke instructions to cause the communication device to perform the model updating method according to the first aspect, the optional embodiments of the first aspect, and/or the model updating method according to the second aspect, the optional embodiments of the second aspect.

Embodiments of the present disclosure also propose a communication system comprising a terminal configured to implement the model updating method according to the alternative embodiments of the first aspect, and a network device configured to implement the model updating method according to the alternative embodiments of the second aspect.

Embodiments of the present disclosure also provide a storage medium storing instructions that, when executed on a communication device, cause the communication device to perform the model updating method of the first aspect, the alternative embodiments of the first aspect, and/or the model updating method of the second aspect, the alternative embodiments of the second aspect.

The embodiments of the present disclosure also provide an apparatus for implementing any of the above methods, for example, an apparatus is provided, where the apparatus includes a unit or a module for implementing each step performed by the terminal in any of the above methods. For another example, another apparatus is also proposed, which includes a unit or module configured to implement steps performed by a network device (e.g., an access network device, a core network function node, a core network device, etc.) in any of the above methods.

It should be understood that the division of each unit or module in the above apparatus is merely a division of a logic function, and may be fully or partially integrated into one physical entity or may be physically separated when actually implemented. Furthermore, the units or modules in the device may be implemented in the form of processor-invoked software, e.g. the device comprises a processor, which is connected to a memory, in which instructions are stored, the processor invoking the instructions stored in the memory for implementing any of the above methods or for implementing the functions of the units or modules of the device, wherein the processor is e.g. a general purpose processor, such as a central processing unit (Central Processing Unit, CPU) or a microprocessor, and the memory is a memory within the device or a memory outside the device. Or the units or modules in the apparatus may be implemented in the form of hardware circuits, where the functions of some or all of the units or modules may be implemented by a design of a hardware circuit, where the hardware circuit may be understood as one or more processors, for example, in one implementation, the hardware circuit is an application-specific integrated circuit (ASIC), the functions of some or all of the units or modules may be implemented by a design of logic relationships between elements within the circuit, and in another implementation, the hardware circuit may be implemented by a programmable logic device (programmable logic device, PLD), where a field programmable gate array (Field Programmable GATE ARRAY, FPGA) may include a number of logic gates, where the connection relationships between the logic gates are configured by a configuration file, so as to implement the functions of some or all of the units or modules. All units or modules of the above device may be realized in the form of invoking software by a processor, or in the form of hardware circuits, or in part in the form of invoking software by a processor, and in the rest in the form of hardware circuits.

In the embodiments of the present disclosure, the processor is a circuit having a signal processing capability, and in one implementation, the processor may be a circuit having an instruction reading and running capability, such as a central processing unit (Central Processing Unit, CPU), a microprocessor, a graphics processor (graphics processing unit, GPU) (which may be understood as a microprocessor), or a digital signal processor (DIGITAL SIGNAL processor, DSP), etc., and in another implementation, the processor may implement a function through a logic relationship of a hardware circuit, where the logic relationship of the hardware circuit is fixed or reconfigurable, for example, the processor is a hardware circuit implemented by an application-specific integrated circuit (ASIC) or a programmable logic device (programmable logic device, PLD), for example, FPGA. In the reconfigurable hardware circuit, the processor loads the configuration document, and the process of implementing the configuration of the hardware circuit may be understood as a process of loading instructions by the processor to implement the functions of some or all of the above units or modules. Furthermore, a hardware circuit designed for artificial intelligence may be also be considered as an ASIC, such as a neural network Processing Unit (Neural Network Processing Unit, NPU), tensor Processing Unit (Tensor Processing Unit, TPU), deep learning Processing Unit (DEEP LEARNING Processing Unit, DPU), and the like.

Fig. 7 is a schematic structural diagram of a communication device 7100 according to an embodiment of the present disclosure. The communication device 7100 may be a network device (e.g., an access network device, a core network device, etc.), a terminal (e.g., a user device, etc.), a chip system, a processor, etc. that supports the network device to implement any of the above methods, or a chip, a chip system, a processor, etc. that supports the terminal to implement any of the above methods. The communication device 7100 may be used to implement the methods described in the above method embodiments, and may be referred to in particular in the description of the above method embodiments.

As shown in fig. 7, the communication device 7100 includes one or more processors 7101. The processor 7101 may be a general-purpose processor or a special-purpose processor, etc., and may be, for example, a baseband processor or a central processing unit. The baseband processor may be used to process communication protocols and communication data, and the central processor may be used to control communication devices (e.g., base stations, baseband chips, terminal devices, terminal device chips, DUs or CUs, etc.), execute programs, and process data for the programs. The processor 7101 is operable to invoke instructions to cause the communication device 7100 to perform any of the above methods.

In some embodiments, the communication device 7100 also includes one or more memories 7102 for storing instructions. Alternatively, all or part of the memory 7102 may be external to the communication device 7100.

In some embodiments, the communication device 7100 also includes one or more transceivers 7103. When the communication device 7100 includes one or more transceivers 7103, communication steps such as transmission and reception in the above method are performed by the transceivers 7103, and other steps are performed by the processor 7101.

In some embodiments, the transceiver may include a receiver and a transmitter, which may be separate or integrated. Alternatively, terms such as transceiver, transceiver unit, transceiver circuit, etc. may be replaced with each other, terms such as transmitter, transmitter circuit, etc. may be replaced with each other, and terms such as receiver, receiving unit, receiver, receiving circuit, etc. may be replaced with each other.

Optionally, the communication device 7100 further comprises one or more interface circuits 7104, the interface circuits 7104 being connected to the memory 7102, the interface circuits 7104 being operable to receive signals from the memory 7102 or other means, and being operable to transmit signals to the memory 7102 or other means. For example, the interface circuit 7104 may read an instruction stored in the memory 7102 and send the instruction to the processor 7101.

The communication device 7100 in the above embodiment description may be a network device or a terminal, but the scope of the communication device 7100 described in the present disclosure is not limited thereto, and the structure of the communication device 7100 may not be limited by fig. 7. The communication device may be a stand-alone device or may be part of a larger device. The communication device may be, for example, 1) a stand-alone integrated circuit IC, or chip, or a system or subsystem of chips, (2) a set of one or more ICs, which may optionally also include storage means for storing data, programs, (3) an ASIC, such as a Modem, (4) a module that may be embedded in other devices, (5) a receiver, terminal device, smart terminal device, cellular telephone, wireless device, handset, mobile unit, vehicle-mounted device, network device, cloud device, artificial smart device, etc., (6) and so forth.

Fig. 8 is a schematic structural diagram of a chip 8200 according to an embodiment of the disclosure. For the case where the communication device 7100 may be a chip or a chip system, reference may be made to a schematic structural diagram of the chip 8200 illustrated in fig. 8, but is not limited thereto.

The chip 8200 includes one or more processors 8201, the processors 8201 being configured to invoke instructions to cause the chip 8200 to perform any of the methods described above.

In some embodiments, the chip 8200 further includes one or more interface circuits 8202, the interface circuits 8202 being coupled to the memory 8203, the interface circuits 8202 being operable to receive signals from the memory 8203 or other devices, the interface circuits 8202 being operable to provide signals to the memory

8203 Or other means. For example, the interface circuit 8202 may read instructions stored in the memory 8203 and send the instructions to the processor 8201. Alternatively, the terms interface circuit, interface, transceiver pin, transceiver, etc. may be interchanged.

In some embodiments, chip 8200 further includes one or more memories 8203 for storing instructions. Alternatively, all or part of the memory 8203 may be external to the chip 8200.

The present disclosure also proposes a storage medium having stored thereon instructions that, when executed on a communication device 7100, cause the communication device 7100 to perform any of the above methods. Optionally, the storage medium is an electronic storage medium. Alternatively, the storage medium described above is a computer-readable storage medium, but is not limited thereto, and it may be a storage medium readable by other devices. Alternatively, the above-described storage medium may be a non-transitory (non-transitory) storage medium, but is not limited thereto, and it may also be a transitory storage medium.

The present disclosure also proposes a program product which, when executed by a communication device 7100, causes the communication device 7100 to perform any of the above methods. Optionally, the above-described program product is a computer program product.

The present disclosure also proposes a computer program which, when run on a computer, causes the computer to perform any of the above methods.

Claims

A model updating method, characterized in that it is executed by a terminal, and the method comprises:

Receiving a model parameter for updating a first model sent by a network device, wherein the first model is deployed on the terminal;

The model parameters are applied to the model structure of the first model to obtain an updated first model.

The method according to claim 1, characterized in that the method further comprises:

Sending a query request for the first model to the network device, wherein the query request is used to determine whether the network device has model parameters for updating the first model;

A query response for the first model sent by the network device is received, wherein the query response is used to indicate whether the network device has model parameters for updating the first model.

The method according to claim 2, wherein the query request includes at least one of the following:

an identifier of a model that the terminal desires to update;

The terminal expects the structure identifier of the model to be updated.

An update notification for the first model sent by the network device is received, wherein the update notification is used to indicate that a model parameter of the first model is updated.

The method according to claim 4 is characterized in that the terminal supports model parameter update capability.

The method according to any one of claims 2 to 5, characterized in that the method further comprises:

Receive configuration information sent by the network device for updating the model parameters of the first model, wherein the configuration information is used to assist the terminal in determining whether to receive the model parameters sent by the network device for updating the first model.

The method according to claim 6, characterized in that the configuration information includes at least one of the following:

The computational complexity of the model parameters;

The storage space required for model parameters;

Scenarios where model parameters are used.

The method according to any one of claims 1 to 7, characterized in that the method further comprises:

According to the network device having model parameters for updating the first model, it is determined whether to receive the model parameters.

The method according to any one of claims 1 to 8, characterized in that the method further comprises:

The updated first model runs successfully, and first information is sent to the network device, wherein the first information is used to indicate that the update of the first model is completed.

The method according to claim 9 is characterized in that the network device is expected to assign a new model identifier to the updated first model.

The updated first model cannot run successfully, and second information is sent to the network device, wherein the second information is used to indicate that the update of the first model has failed.

The method according to claim 11 is characterized in that the second information is also used to indicate the reason why the first model update failed.

The method according to any one of claims 1 to 12, characterized in that the method further comprises:

A model structure of at least one model deployed on the terminal is indicated to the network device.

The method according to claim 13, characterized in that indicating to the network device a model structure of at least one model deployed on the terminal comprises:

Sending third information to the network device, wherein the third information includes a model identifier of the at least one model; the network device stores a first correspondence between the model identifier and the model structure, and the third information is used by the network device to determine the model structure corresponding to the model identifier of the at least one model according to the first correspondence.

Sending fourth information to the network device, wherein the fourth information includes a model structure identifier of the at least one model; the network device stores a second correspondence between the model structure identifier and the model structure, and the fourth information is used by the network device to determine the model structure corresponding to the model structure identifier of the at least one model according to the second correspondence.

A model updating method, characterized in that it is executed by a network device, and the method comprises:

Model parameters for updating a first model are sent to a terminal, wherein the first model is deployed on the terminal; the model parameters are used to be applied to the model structure of the first model to obtain an updated first model.

The method according to claim 16, characterized in that the method further comprises:

receiving a query request for the first model sent by the terminal, wherein the query request is used to determine whether the network device has model parameters for updating the first model;

A query response for the first model is sent to the terminal, wherein the query response is used to indicate whether the network device has model parameters for updating the first model.

The method according to claim 17, wherein the query request includes at least one of the following:

an identifier of a model that the terminal desires to update;

The terminal expects the structure identifier of the model to be updated.

An update notification for the first model is sent to the terminal, wherein the update notification is used to indicate that a model parameter of the first model is updated.

The method according to claim 19, characterized in that the method further comprises:

Determining whether the terminal supports model parameter update capability;

The sending an update notification for the first model to the terminal includes:

In response to the terminal supporting the model parameter update capability, sending an update notification for the first model to the terminal.

The method according to any one of claims 17 to 20, characterized in that the method further comprises:

Configuration information for updating the model parameters of the first model is sent to the terminal, wherein the configuration information is used to assist the terminal in determining whether to receive the model parameters for updating the first model sent by the network device.

The method according to claim 21, characterized in that the configuration information includes at least one of the following:

The computational complexity of the model parameters;

The storage space required for model parameters;

Scenarios where model parameters are used.

The method according to any one of claims 16 to 22, characterized in that the method further comprises:

It is determined whether the terminal receives the model parameters.

The method according to any one of claims 16 to 23, characterized in that the method further comprises:

Receive first information sent by the terminal, wherein the first information is used to indicate that the first model update is completed.

The method according to claim 24, characterized in that the method further comprises:

According to the completion of updating the first model, a new model identifier is allocated to the updated first model.

Second information sent by the terminal is received, wherein the second information is used to indicate that the first model update fails.

The method according to claim 26 is characterized in that the second information is also used to indicate the reason why the first model update failed.

The method according to any one of claims 16 to 27, characterized in that the method further comprises:

According to an instruction of the terminal, a model structure of at least one model deployed on the terminal is determined.

The method according to claim 28, characterized in that the network device stores a first correspondence between the model identifier and the model structure;

The determining, according to the instruction of the terminal, a model structure of at least one model deployed on the terminal includes:

receiving third information sent by the terminal, wherein the third information includes a model identifier of the at least one model;

A model structure corresponding to the model identifier of the at least one model is determined according to the first corresponding relationship and the third information.

The method according to claim 28, characterized in that the network device stores a second correspondence between the model structure identifier and the model structure;

receiving fourth information sent by the terminal, wherein the fourth information includes a model structure identifier of the at least one model;

A model structure corresponding to the model structure identifier of the at least one model is determined according to the second corresponding relationship and the fourth information.

A model updating device, characterized in that the device comprises:

A transceiver module, configured to receive a model parameter for updating a first model sent by a network device, wherein the first model is deployed on a terminal;

A processing module is used to apply the model parameters to the model structure of the first model to obtain an updated first model.

A model updating device, characterized in that the device comprises:

A transceiver module is used to send model parameters for updating a first model to a terminal, wherein the first model is deployed on the terminal; the model parameters are used to be applied to the model structure of the first model to obtain an updated first model.

A terminal, characterized by comprising:

one or more processors;

Wherein, the terminal is used to execute the model updating method described in any one of claims 1-15.

A network device, comprising:

one or more processors;

Wherein, the network device is used to execute the model updating method described in any one of claims 16-30.

A communication device, comprising:

one or more processors;

The processor is used to call instructions so that the communication device executes the model updating method described in any one of claims 1-15 or 16-30.

A communication system, characterized in that it includes a terminal and a network device, wherein the terminal is configured to implement the model updating method described in any one of claims 1-15, and the network device is configured to implement the model updating method described in any one of claims 16-30.

A storage medium storing instructions, characterized in that when the instructions are executed on a communication device, the communication device executes the model updating method described in any one of claims 1-15 or 16-30.