Data transmission method, device, terminal, storage medium and program product

Abstract

Disclosed is a data transmission method, the method comprising: a first device sending a first message to a first terminal, the first message being used for establishing one or more first sessions, a first session comprising one or more radio bearers used for transmitting artificial intelligence or machine learning data of any size, and a radio bearer of the first session being identified by means of an identifier of an artificial intelligence or machine learning model; the first terminal, by means of the radio bearer corresponding to the first identifier, directly accessing the artificial intelligence or machine learning data corresponding to the first identifier in the first device side for training or inference; the first terminal receiving the first message, and sending a second message to the first device, the second message being used for representing completion of the establishment of the first session(s); and the first device receiving the second message. Further disclosed are a first device, a first terminal, a second device, a fourth device, a computer-readable storage medium, and a computer program product.

Description

A data transmission method, device, terminal, storage medium, and program product.

[0001] Cross-reference to related applications

[0002] This disclosure claims priority to Chinese Patent Application No. 202411578397.2, filed on November 6, 2024, entitled "A data transmission method, apparatus, terminal, storage medium and program product", the entire contents of which are incorporated herein by reference. Technical Field

[0003] This disclosure relates to, but is not limited to, the field of communications, and particularly to a data transmission method, a first device, a first terminal, a second device, a fourth device, a computer-readable storage medium, and a computer program product. Background Technology

[0004] Currently, artificial intelligence (AI) data between terminals and base stations can be transmitted in two ways: one is through radio connection control plane signaling transmission, and the other is through radio connection user plane signaling transmission.

[0005] Regarding the transmission of control plane signaling for radio connectivity, 3GPP specifies that the transmission limit for control signaling protocol data unit (PDU) sessions is 9000 bytes, and the transmission limit for AI / machine learning (ML) models is 45KB. This cannot meet the transmission requirements for AI / ML models with sizes ranging from tens to hundreds of MB.

[0006] For user plane transmission in wireless connections, in mobile communication networks, data transmission for terminal services is established by the core network (CN) establishing sessions and the radio access network (RAN) establishing radio bearers. In terms of session management, PDU sessions are established and maintained by the CN. For AI for Radio Access Network (RAN) scenarios, AI computation within the RAN domain typically needs to meet low latency requirements, such as being at the millisecond level. Therefore, having the CN process RAN domain AI data will face real-time data processing risks, leading to increased data processing latency, data caching, data accumulation and overflow, and reduced network reliability.

[0007] In summary, the methods for transmitting AI data provided in related technologies cannot meet the requirements. Therefore, there is an urgent need to provide a new transmission method for transmitting AI data between terminals and base stations. Summary of the Invention

[0008] This disclosure provides a data transmission method, a first device, a first terminal, a second device, a fourth device, a computer-readable storage medium, and a computer program product, providing a novel transmission method for transmitting AI data between a terminal and a base station.

[0009] The technical solution of this disclosure embodiment is implemented as follows:

[0010] Firstly, embodiments of this disclosure provide a data transmission method applied to a first device, the method comprising:

[0011] Send a first message to a first terminal; wherein the first message is used to establish one or more first sessions; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier in the first device side for training or inference through the wireless bearer corresponding to the first identifier;

[0012] Receive a second message sent by the first terminal; wherein the second message is used to indicate that the first session has been established.

[0013] Secondly, embodiments of this disclosure provide a data transmission method applied to a first terminal, the method comprising:

[0014] The first terminal receives a first message sent by a first device; wherein the first message is used to establish one or more first sessions; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier in the first device side for training or inference through the wireless bearer corresponding to the first identifier.

[0015] A second message sent to the first device; wherein the second message is used to indicate that the first session has been established.

[0016] Thirdly, embodiments of this disclosure provide a data transmission method applied to a second device, the method comprising:

[0017] The first device receives a context sent by the first device; the context includes the context of the first terminal on the first device side and the first session context established for the first session between the first device and the first terminal; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier on the first device side for training or inference through the wireless bearer corresponding to the first identifier.

[0018] Based on the aforementioned context, an artificial intelligence or machine learning data transmission session is established for the first terminal, a transmission channel for artificial intelligence or machine learning data is established between the first device and the second device, and an artificial intelligence or machine learning model used by the first device is deployed / activated.

[0019] Fourthly, embodiments of this disclosure provide a data transmission method applied to a fourth device, the method comprising:

[0020] A third instruction message is sent to the first device; wherein the third instruction message is used to instruct the establishment of an artificial intelligence or machine learning model lifecycle task; the third instruction information includes model identifier, model size, model training or inference, performance requirements of the artificial intelligence or machine learning model lifecycle task; model data acquisition configuration information; and acquisition terminal information;

[0021] Receive a fifth message sent by the first device; wherein the fifth message is used to notify that the lifecycle task of the artificial intelligence or machine learning model has been established.

[0022] Receive artificial intelligence or machine learning data sent by the first device; wherein the artificial intelligence or machine learning data is sent directly by the terminal through a session, or sent through a third device; the session includes the first session; the terminal includes the first terminal;

[0023] Based on the collected artificial intelligence or machine learning data, artificial intelligence or machine learning models are trained.

[0024] Fifthly, embodiments of this disclosure provide a first device, the first device comprising:

[0025] The first sending part is configured to send a first message to a first terminal; wherein the first message is used to establish one or more first sessions; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier in the first device side for training or inference through the wireless bearer corresponding to the first identifier.

[0026] The first receiving part is configured to receive a second message sent by the first terminal; wherein the second message is used to indicate that the first session has been established.

[0027] Sixthly, this disclosure provides a first terminal, the first terminal comprising:

[0028] The second receiving part is configured to receive a first message sent by the first device; wherein the first message is used to establish one or more first sessions; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier in the first device side for training or inference through the wireless bearer corresponding to the first identifier.

[0029] The second sending part is configured to send a second message to the first device; wherein the second message is used to indicate that the first session has been established.

[0030] Seventhly, embodiments of this disclosure provide a second device, the second device comprising:

[0031] The third receiving part is configured to receive a context sent by the first device; the context includes the context of the first terminal on the first device side and the first session context established for the first session between the first device and the first terminal; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier on the first device side for training or inference through the wireless bearer corresponding to the first identifier;

[0032] The third processing unit is configured to, based on the context, establish an artificial intelligence or machine learning data transmission session for the first terminal, establish a transmission channel for artificial intelligence or machine learning data between the first device and the second device, and deploy / activate an artificial intelligence or machine learning model used by the first device.

[0033] Eighthly, embodiments of this disclosure provide a fourth device, the fourth device comprising:

[0034] The fourth sending part is configured to send a third indication message to the first device; wherein, the third indication message is used to indicate the establishment of an artificial intelligence or machine learning model lifecycle task; the third indication information includes model identifier, model size, model training or inference, performance requirements of the artificial intelligence or machine learning model lifecycle task; model data acquisition configuration information; and acquisition terminal information;

[0035] The fourth receiving part is configured to receive a fifth message sent by the first device; wherein the fifth message is used to notify that the lifecycle task of the artificial intelligence or machine learning model has been established.

[0036] The fourth transmitting section is further configured to receive artificial intelligence or machine learning data transmitted by the first device; wherein the artificial intelligence or machine learning data is transmitted directly by the terminal through a session, or through a third device; the session includes the first session; the terminal includes the first terminal;

[0037] The fourth processing section is configured to train artificial intelligence or machine learning models based on the collected artificial intelligence or machine learning data.

[0038] Ninth aspect, embodiments of this disclosure provide a first device, the first device comprising:

[0039] The first memory is configured to store executable instructions;

[0040] When the first processor is configured to execute executable instructions stored in the first memory, it performs the steps of the data transfer method described above.

[0041] Tenthly, embodiments of this disclosure provide a first terminal, the first terminal comprising:

[0042] The second memory is configured to store executable instructions;

[0043] When the second processor is configured to execute executable instructions stored in the second memory, it performs the steps of the above-described data transfer method.

[0044] Eleventhly, embodiments of this disclosure provide a second device, the second device comprising:

[0045] The third memory is configured to store executable instructions;

[0046] When the third processor is configured to execute executable instructions stored in the third memory, it performs the steps of the above-described data transfer method.

[0047] In a twelfth aspect, embodiments of this disclosure provide a fourth device, the fourth device comprising:

[0048] The fourth memory is configured to store executable instructions;

[0049] When the fourth processor is configured to execute executable instructions stored in the fourth memory, it performs the steps of the above-described data transfer method.

[0050] In a thirteenth aspect, embodiments of this disclosure provide a computer-readable storage medium having stored thereon one or more computer programs, which can be executed by one or more processors to implement the steps of the above-described data transmission method.

[0051] Fourteenthly, embodiments of this disclosure provide a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described data transmission method.

[0052] This disclosure provides a data transmission session, i.e., a first session, established based on a first interactive message between a first terminal and a first device, such as a base station, allowing the terminal to directly access AI / ML data training / inference on the base station side. Simultaneously, since the base station can directly perform AI / ML data training / inference in this disclosure, in an AIforRAN scenario, enabling AIforRAN data to close the RAN domain means that AI / ML data will directly terminate between the terminal and the base station, reducing unnecessary protocol conversions and processing, and lowering latency. Attached Figure Description

[0053] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings used in the embodiments of this disclosure will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 is a schematic diagram of a communication system according to an embodiment of this disclosure;

[0055] Figure 2 is a schematic flowchart of the data transmission method provided in an embodiment of this disclosure;

[0056] Figure 3 is a schematic diagram of a single AI data session provided in an embodiment of this disclosure;

[0057] Figure 4 is a schematic diagram of multiple AI data sessions provided in the embodiments of this disclosure;

[0058] Figure 5 is a schematic flowchart of the data transmission method provided in this embodiment of the present disclosure;

[0059] Figure 6 is a schematic flowchart of the data transmission method provided in this embodiment of the present disclosure;

[0060] Figure 7 is a schematic flowchart of the data transmission method provided in this embodiment of the present disclosure;

[0061] Figure 8 is a flowchart illustrating the data transmission method provided in this embodiment of the present disclosure.

[0062] Figure 9 is a schematic flowchart of the data transmission method provided in an embodiment of this disclosure;

[0063] Figure 10 is a schematic flowchart of the data transmission method provided in this embodiment of the present disclosure;

[0064] Figure 11 is a schematic block diagram of a first device provided in an embodiment of this disclosure;

[0065] Figure 12 is a schematic block diagram of a first terminal provided in an embodiment of this disclosure;

[0066] Figure 13 is a schematic block diagram of a second device provided in an embodiment of this disclosure;

[0067] Figure 14 is a schematic block diagram of a fourth device provided in an embodiment of this disclosure;

[0068] Figure 15 is a schematic structural diagram of a communication device provided in an embodiment of this disclosure. Detailed Implementation

[0069] The technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0070] The embodiments disclosed herein can be applied to various communication systems, such as: Global System of Mobile communication (GSM) system, Code Division Multiple Access (CDMA) system, Wideband Code Division Multiple Access (WCDMA) system, General Packet Radio Service (GPRS), Long Term Evolution (LTE) system, Advanced Long Term Evolution (LTE-A) system, New Radio (NR) system, evolution of NR system, LTE-based access to unlicensed spectrum (LTE-U) system, NR-based access to unlicensed spectrum (NR-U) system, Universal Mobile Telecommunication System (UMTS), Wireless Local Area Networks (WLAN), Wireless Fidelity (WiFi), next-generation communication systems, or other communication systems, etc.

[0071] Figure 1 is a schematic diagram of a communication system according to an embodiment of the present disclosure.

[0072] As shown in Figure 1, the communication system 100 may include a terminal device 110 and a network device 120. The network device 120 can communicate with the terminal device 110 via an air interface. Multi-service transmission is supported between the terminal device 110 and the network device 120.

[0073] In the communication system 100 shown in Figure 1, the network device 120 can be an access network device that communicates with the terminal device 110. The access network device can provide communication coverage for a specific geographical area and can communicate with the terminal device 110 located within that coverage area.

[0074] Network device 120 may be an evolved Node B (eNB or eNodeB) in a Long Term Evolution (LTE) system, or a Next Generation Radio Access Network (NG RAN) device, or a base station (gNB or gNodeB) in an NR system, or a radio controller in a Cloud Radio Access Network (CRAN), or the network device 120 may be a relay station, access point, vehicle-mounted equipment, wearable device, hub, switch, bridge, router, or network equipment in a future evolved Public Land Mobile Network (PLMN), etc.

[0075] Terminal equipment 110 can also be referred to as user equipment (UE), access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, terminal, wireless communication equipment, user agent, or user device, etc. Terminal equipment can be a station (STAION, ST) in a WLAN, a cellular phone, cordless phone, Session Initiation Protocol (SIP) phone, Wireless Local Loop (WLL) station, Personal Digital Assistant (PDA) device, handheld device with wireless communication capabilities, computing device or other processing device connected to a wireless modem, in-vehicle equipment, wearable device, and next-generation communication systems, such as terminal equipment in NR networks or terminal equipment in future evolved Public Land Mobile Network (PLMN) networks, etc.

[0076] Figure 1 illustrates an exemplary base station and two terminals. Optionally, the communication system 100 may include multiple base stations, and each base station may include other numbers of terminals within its coverage area. This disclosure does not limit the scope of the embodiments.

[0077] It should be noted that Figure 1 is merely an example illustrating the system to which this disclosure applies. Of course, the methods shown in the embodiments of this disclosure can also be applied to other systems. Furthermore, the terms "system" and "network" are often used interchangeably herein. The term "and / or" in this document merely describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship. It should also be understood that "instruction" mentioned in the embodiments of this disclosure can be a direct instruction, an indirect instruction, or an indication of a relationship. For example, A instructing B can mean that A directly instructs B, for example, B can be obtained through A; it can also mean that A indirectly instructs B, for example, A instructs C, B can be obtained through C; or it can mean that there is a relationship between A and B. It should also be understood that "correspondence" mentioned in the embodiments of this disclosure can indicate a direct or indirect correspondence between two things, or an association between two things, or a relationship of instruction and being instructed, configuration and being configured, etc. It should also be understood that the "predefined" or "predefined rules" mentioned in the embodiments of this disclosure can be implemented by pre-storing corresponding codes, tables, or other means that can be used to indicate relevant information in the device (e.g., including terminal devices and network devices), and this disclosure does not limit the specific implementation method. For example, predefined can refer to those defined in a protocol. It should also be understood that in the embodiments of this disclosure, the "protocol" can refer to standard protocols in the field of communication, such as the LTE protocol, the NR protocol, and related protocols applied to future communication systems, and this disclosure does not limit it.

[0078] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. The terminology used herein is for the purpose of describing embodiments of this disclosure only and is not intended to be limiting of this disclosure.

[0079] Before explaining this disclosure, the following describes the switching scheme in the related art:

[0080] The continuous growth in data volume driven by emerging technologies, applications, and scenarios has created a more urgent demand for computing power and networks across various industries. This is especially true for business scenarios such as autonomous driving and VR, which place extremely high demands on computing and networks. Considering the requirements for low latency, high bandwidth, and high computing power, it is necessary to fully utilize wireless edge computing networks and ensure efficient collaboration and flow of computing power across cloud, edge, and device, meeting the on-demand computing power needs of businesses. Simultaneously, with industry applications placing extreme demands on end-to-end network quality, networks need to evolve from best-effort to end-to-end deterministic guarantees, and network protocols also require innovative development.

[0081] With the gradual convergence of Communication Technology (CT) and Information Technology (IT), 5G wireless is also evolving towards cloudification. Traditional dedicated hardware is evolving towards general-purpose servers, and the hardware resources of traditional dedicated wireless base stations will also gradually evolve towards wireless-side cloud platforms by adding accelerator cards. This will support the deployment of wireless data plane and control plane functions, and also allow for the agile deployment of local field-level or edge service applications. Based on the cloud foundation on the wireless side, the wireless side also has the ability to support diverse computing resources and unified computing power orchestration and scheduling. How to fully utilize the remaining computing resources of base stations to maximize their value, considering the distributed nature of base stations, requires unified management and control of dispersed computing resources to achieve a synergistic effect while meeting the requirements of latency-sensitive services such as Extended Reality (XR) and Vehicle-to-X (V2X). Further research is needed in this area. With the continuous development of new businesses such as intelligent services, XR, metaverse-type immersive services, vehicle networking and IoT, intelligent terminals are not only becoming more diversified in product form, but also rapidly developing in computing power thanks to the continuous maturation of semiconductor technology and the introduction of AI chips.

[0082] Considering the limitations of terminal-side computing power due to factors such as computing power, battery capacity, and heat generation, as well as the sensitivity of immersive services to experiments, this study investigates how base stations provide computing services to terminals, assisting them in completing computing tasks in real time. In AI for RAN scenarios, it is necessary to study transmission schemes for AI data between terminals and base stations.

[0083] In AI for RAN scenarios, AI / ML model lifecycle management includes processes such as data collection, model training, model delivery, model registration, model configuration / activation / deactivation, model inference, model monitoring, and model adjustment. In 3GPP, during AI / ML collaboration between terminals and base stations, data acquisition for AI / ML model inference has latency requirements when the required data comes from other entities. Similarly, for (real-time) AI / ML performance monitoring, data acquisition latency is also required when the required monitoring data (such as performance metrics) comes from other entities. In other words, AI for RAN use cases impose latency requirements on data transmission and model delivery for AI / ML model training / inference.

[0084] Figure 2 is a flowchart illustrating a data transmission method according to an embodiment of this disclosure. As shown in Figure 2, the method is applied to the communication system shown in Figure 1, and includes:

[0085] Step 201: The first device sends a first message to the first terminal.

[0086] The first message is used to establish one or more first sessions; the radio bearer of the first session is identified by an identifier of an artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier in the first device side for training or inference through the radio bearer corresponding to the first identifier.

[0087] It should be noted that each use case / feature corresponds to a first session.

[0088] It should be noted that the first session includes one or more radio bearers for transmitting AI or machine learning data of arbitrary size. This means that the first terminal (terminal device 110 in Figure 1) and the first device (network device 120 in Figure 1) can transmit AI or machine learning data of arbitrary size, indicating that the size of AI / ML data between the terminal and the base station is not limited. Considering the evolution of AI technology and its application in the RAN, the radio access network should provide flexible transmission bandwidth adaptability for AI / ML data transmission between the terminal and the base station. For example, this avoids forcing 9001B data to be transmitted using two PDUs, reducing the processing overhead and PDU waiting time on the base station side.

[0089] In this embodiment of the disclosure, AI / ML data between the first device and the first terminal, i.e. between the terminal and the base station, is in a closed loop within the radio access network. This ensures the latency of AI for RAN data transmission, and the AI / ML data can be processed by the base station without having to go through the core network routing, thus avoiding unnecessary increases in transmission latency.

[0090] In this embodiment of the disclosure, the first message is an indication message or a notification message, used to indicate or notify the radio bearer information associated with the first session.

[0091] Here, the first device pre-starts the artificial intelligence or machine learning model training lifecycle management process, performs artificial intelligence or machine learning model selection and model deployment; collects training data for the selected or to-be-deployed artificial intelligence or machine learning model; receives artificial intelligence or machine learning data sent directly by the terminal through a session or through a third device; the session includes the first session; the terminal includes the first terminal; and performs artificial intelligence or machine learning model training based on the artificial intelligence or machine learning data.

[0092] This disclosure proposes a novel session for AI / ML data transmission between terminals and base stations, such as an AI data session: this session can be understood as a way to allow terminal devices to directly access AI / ML training / inference on the wireless access network side via a wireless air interface.

[0093] Figure 3 is a schematic diagram of a single AI data session provided in an embodiment of this disclosure. As shown in Figure 3, a single AI data session is established between a terminal and a base station. This AI data session can be established for a terminal, and each terminal has an AI data session with the base station. The AI ​​data session includes one or more wireless bearers, which are used to transmit AI data. The wireless bearers in the AI ​​data session can be identified according to AI / ML model identifiers. For example, if the terminal and the base station simultaneously perform AI / ML model training / inference such as CSI feedback compression, beam management, and physical layer AI positioning, the base station can establish a wireless bearer for each use case AI / ML model training / inference / dataset transmission / model transmission, and identify it with an AI / ML model identifier. For example, model ID_1 is used to identify wireless bearer 1, model ID_2 is used to identify wireless bearer 2, ..., and model ID_n is used to identify wireless bearer n.

[0094] Figure 4 is a schematic diagram of multiple AI data sessions provided in the embodiments of this disclosure; as shown in Figure 3: multiple AI data sessions are established between each terminal and the base station, such as AI data session 1, ..., AI data session n; AI data sessions can be established for terminals, and different AI / ML model training / inference can be performed simultaneously between terminals and base stations based on different AI data sessions; one AI data session is established for each AIforRAN use case / feature. For example, an AI data session is established for the CSI feedback compression use case / feature, and an AI data session is established for beam management. Each AI data session contains one radio bearer, which is used to transmit AI data. The radio bearers in the AI ​​data session can be identified according to the AI / ML model identifier. For example, model ID_1 is used to identify radio bearer 1, ..., and model ID_n is used to identify radio bearer n.

[0095] Step 202: The first terminal receives the first message.

[0096] Step 203: The first terminal sends a second message to the first device.

[0097] The second message is used to indicate that the first session has been established.

[0098] It should be noted that the second message also includes the identifier of the first session; here, the identifier of the first session can be the identifier of the artificial intelligence or machine learning model in the radio bearer of the first session.

[0099] Step 204: The first device receives the second message.

[0100] This disclosure provides a data transmission method, comprising: a first device sending a first message to a first terminal; wherein the first message is used to establish one or more first sessions; the first session includes one or more radio bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the radio bearers of the first session are identified by an identifier of an artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier on the first device side for training or inference through the radio bearer corresponding to the first identifier. The first terminal receives the first message and sends a second message to the first device; wherein the second message is used to indicate that the first session has been established; the first device receives the second message. In other words, this disclosure provides a data transmission session, i.e., a first session, based on a first interaction message between a first terminal and a first device, such as a base station, allowing the terminal to directly access AI / ML data training / inference on the base station side; simultaneously, in this disclosure, the base station can directly perform AI / ML data training / inference, so in the AIforRAN scenario, enabling AIforRAN data to close the RAN domain means that the AI / ML data will directly terminate between the terminal and the base station, reducing unnecessary protocol conversions and processing, and lowering latency.

[0101] In this embodiment of the disclosure, in the AIforRAN scenario, since AI data is data that the wireless network can understand and AIforRAN data terminates at the terminal and the base station, in order to reduce unnecessary protocol conversion and processing, AI data sessions do not necessarily need to be based on the IP protocol. For example, they can communicate based on the terminal IP or the IP of the AI / ML training / inference application.

[0102] The first session includes one or more of the following: session identifier; identifier of the first terminal; identifier of the artificial intelligence or machine learning model; processing method of the model; whether the artificial intelligence or machine learning data to be transmitted needs integrity protection or encryption; integrity protection or encryption algorithm that can be used for artificial intelligence or machine learning data that needs integrity protection or encryption; whether the artificial intelligence or machine learning data to be transmitted needs compression or encoding; encoding or compression method that can be used for artificial intelligence or machine learning data that needs compression or encoding; radio bearer information; artificial intelligence or machine learning data.

[0103] It should be noted that if "Data compression?" is "yes", the AI ​​data will be compressed according to the compression method described above; otherwise, it will not be compressed.

[0104] Here, session identifiers include, but are not limited to, PDU Session IDs.

[0105] Terminal identifiers include, but are not limited to, Cell Radio Network Temporary Identifier (CRNTI), Mobile Station international Integrated Services Digital Network number (MSISDN), International Mobile Subscriber Identity (IMSI), International Mobile Equipment Identity (IMEI), service identifiers (such as the terminal's domain name), and service numbers for a user's specific applications (such as user identifiers for chat software, communication software, etc., installed on the terminal).

[0106] The model processing methods include training the model, inference, and data processing.

[0107] Encryption algorithms include, but are not limited to, encryption algorithms based on the keyed hashed message authentication code (HMAC) function, the MD5 message-digest algorithm, the Secure Hash Algorithm 1 (SHA1), the Cyclic Redundancy Check (CRC), the Data Encryption Standard (DES), and the Advanced Encryption Standard (AES).

[0108] Encoding or compression methods include ASN.1, Delta Modulation (DM), Differential and Adaptive Coding (ADPCM).

[0109] Wireless bearer information includes, but is not limited to, wireless bearer bandwidth information and wireless bearer rate information.

[0110] It should be noted that if the first session meets the first condition, the first device updates the first session.

[0111] The first session meeting the first condition includes one or more of the following: the session identifier is adjusted; the identifier of the first terminal is adjusted; the identifier of the artificial intelligence or machine learning model is adjusted, for example, the base station selects a model with computational complexity adapted to the inference latency requirements, triggering a model update, and carrying the model identifier through the message; artificial intelligence or machine learning data that requires integrity protection or encryption is adjusted to not require integrity protection or encryption; artificial intelligence or machine learning data that does not require integrity protection or encryption is adjusted to require integrity protection or encryption; artificial intelligence or machine learning data that requires compression or encoding is adjusted to not require compression or encoding; artificial intelligence or machine learning data that does not require compression or encoding is adjusted to require compression or encoding; the integrity protection or encryption algorithm used for artificial intelligence or machine learning data that requires integrity protection or encryption is adjusted; the encoding or compression method used for artificial intelligence or machine learning data that requires compression or encoding is adjusted; and the radio bearer resources are adjusted.

[0112] Here, the steps for the first device to update the first session include: the first device sending a first instruction message to the first terminal; wherein the first instruction message is used to instruct the first device to update the first bearer resource of artificial intelligence or machine learning data; the first terminal receiving the first instruction message and sending a third message to the first device; wherein the third message is used to notify that the update of the first bearer resource is complete; and the first device receiving the third message.

[0113] It should be noted that "update" refers to the base station adjusting session resources. For example, in the AI / ML inference process between the terminal and the base station, the base station needs to adjust AI data carrying resources due to increased service load, thus triggering an AI data session update between the base station and the terminal. Figure 5 is a schematic diagram of an AI data session update process provided in this disclosure; as shown in Figure 5:

[0114] Step 501: The base station sends an AI data session update instruction message to the terminal.

[0115] Step 502: The terminal sends an AI / MI data session update completion message to the base station.

[0116] In some embodiments, the method provided by this disclosure includes the following:

[0117] Step A1: The first device establishes the first session for the first terminal in the artificial intelligence or machine learning lifecycle management.

[0118] The first session includes first information; the first information includes quality requirements for artificial intelligence or machine learning training / inference, quality of service (QoS) requirements for wireless connectivity, artificial intelligence or machine learning model information, the base station identifier where the artificial intelligence or machine learning model is located, such as the global identifier of the gNB, and the radio bearer information of the artificial intelligence or machine learning data.

[0119] In this embodiment of the disclosure, the artificial intelligence or machine learning model information includes, but is not limited to, AI / ML model identifier, version, provider, and model type.

[0120] In this embodiment of the disclosure, the wireless bearer information of AI data includes, but is not limited to, the configuration of the packet data convergence protocol (PDCP), radio link control (RLC), and medium access control (MAC) layer bearer (RB).

[0121] Step A2: The first device records the first information into the context.

[0122] The context includes the context of the first terminal on the first device side and the first session context established for the first session.

[0123] In this embodiment of the disclosure, in a scenario where the first terminal needs to switch to a base station serving it due to mobility, the first device sends a context to the second device. Here, the first device is the source device, the device that provides services to the first terminal before the switch, and the second device is the destination device, the device that provides services to the first terminal after the switch. Further, the second device receives the context and, based on the context, establishes an artificial intelligence or machine learning data transmission session for the first terminal, establishes a transmission channel between the first and second devices for artificial intelligence or machine learning data, and deploys / activates the artificial intelligence or machine learning model used by the first device.

[0124] Here, the first device and the second device can establish a transmission channel for artificial intelligence or machine learning data between the first device and the second device based on the context.

[0125] Here, the context sent from the second device to the second device includes, but is not limited to, terminal identifier, model information, AI performance requirements, and AI data session identifier.

[0126] For example, in a cross-site AI data transmission scenario, such as when the source base station's AI / ML model provides inference for the terminal and the terminal's mobility is switched, the source base station (i.e., the first device) transmits information from the AI ​​context to the target base station (i.e., the second device). This serves two purposes: firstly, it enables the target base station to establish an AI data session and wireless bearer for the terminal, and secondly, it establishes an AI data transmission channel between base stations. On the other hand, if the cross-site AI / ML model training / inference data transmission cannot meet the AI / ML performance requirements, it may involve AI / ML model migration. This is achieved by transmitting the AI ​​data session context between the source and target base stations, enabling the target base station to deploy / activate the AI / ML model used by the source base station.

[0127] Figure 6 is a flowchart illustrating the data transmission method between two base stations provided in this disclosure; as shown in Figure 6:

[0128] Step 601: The source base station sends the AI ​​data session context to the target base station.

[0129] Step 602: The target base station establishes an AI data session for the terminal based on the context.

[0130] Step 603: The source base station and the target base station can establish an AI data transmission channel between the base stations based on the AI ​​data session context, such as an Xn air interface data transmission channel, and / or deploy an AI / ML model.

[0131] It should be noted that step 603 is an optional step.

[0132] In this embodiment of the disclosure, the key information elements of the first session context include one or more of the following: an artificial intelligence or machine learning model identifier; an artificial intelligence or machine learning model vendor; an artificial intelligence or machine learning model version; an artificial intelligence or machine learning model type; an artificial intelligence or machine learning model size, such as the number of parameters; an artificial intelligence or machine learning model algorithm complexity; artificial intelligence or machine learning model computing resource requirements, such as a central processing unit (CPU), a graphics processing unit (GPU), resource capabilities (such as CPU clock speed, GPU clock speed / memory, etc.), resource quantity requirements, etc.; the amount of training data for the artificial intelligence or machine learning model; the input parameter information for training the artificial intelligence or machine learning model; the amount of inference data for the artificial intelligence or machine learning model; the input parameter information for inference of the artificial intelligence or machine learning model; terminal information associated with the first session; and base station information associated with the first session.

[0133] Here, the amount of model training / inference data is used to allocate uplink bandwidth.

[0134] Here, the input parameter information includes parameter types, such as Reference Signal Receiving Power (RSRP), Received Signal Strength Indication (RSSI), Signal Noise Ratio (SNR), location information, and calculated load information.

[0135] Here, the associated terminal information or associated base station information includes the terminal or base station identifier and training / inference.

[0136] Here, for AI / ML models, the base station can establish a lifecycle context for the AI / ML model to record the various processes and resource usages of the AI / ML lifecycle management, so as to facilitate dynamic resource preparation across sites when the terminal moves.

[0137] Step A3: The first device allocates data wireless bearer resources based on the first information.

[0138] In some embodiments, the radio bearer may use the data radio bearer (DRB) mechanism in related technologies, which can better ensure compatibility with existing systems and reduce the cost of system transformation; the base station allocates DRB resources suitable for the AI ​​data size, transmission latency and other requirements according to the QoS requirements of the AI ​​data session.

[0139] It should be noted that steps A1 and A3 are the process of the first device establishing the first session.

[0140] In some embodiments, the method provided by this disclosure includes the following:

[0141] Step B1: The first device senses the artificial intelligence or machine learning capabilities of the terminal.

[0142] Among them, the terminal includes the first terminal.

[0143] Here, the first device first sends an AI / MI model capability acquisition request to the first terminal, the first terminal receives the request, and reports the AI / MI model capability to the base station.

[0144] Here, the first device can also use sensing devices to perceive the AI / MI capabilities of the first terminal.

[0145] Step B2: The first device establishes a transmission session and wireless bearer for the artificial intelligence or machine learning data for the second terminal with artificial intelligence or machine learning capabilities.

[0146] Here, the second terminal can be the first terminal, or it can be a different terminal from the first terminal.

[0147] In this embodiment of the disclosure, after the first device senses the AI / MI model capability of the second terminal, it checks whether the second terminal has the corresponding AI / MI model capability. If it does, it establishes a default AI data session for the second terminal.

[0148] For example, when a terminal registers with the network, the base station can sense the terminal's AI / ML capabilities and establish a default AI / ML data transmission session and wireless bearer for terminals with AI / ML capabilities. Figure 7 is a schematic diagram of the default establishment process of an AI data session provided in this disclosure; as shown in Figure 7:

[0149] Step 701: The base station sends a Radio Resource Control (RRC) connection setup complete message to the terminal.

[0150] Step 702: The base station sends an AI / MI model capability acquisition message to the terminal.

[0151] Step 703: The terminal reports its AI / MI model capabilities to the base station.

[0152] Step 704: The base station checks the terminal's model capability. If the terminal has model capability, the base station establishes a default AI data session for the terminal and executes step 705.

[0153] Step 705: The base station sends an AI / MI data session establishment instruction message to the terminal.

[0154] Here, the AI / MI data session establishment instruction message is used to establish a default AI data session.

[0155] Step 706: The terminal reports an AI / MI data session establishment completion message to the base station.

[0156] In some embodiments, the method provided by this disclosure includes the following:

[0157] Step C1: The first device sends the second instruction information to the first terminal.

[0158] The second instruction information is used to instruct the release of the first session and the release of the second bearer resource for artificial intelligence or machine learning data.

[0159] Step C2: The first terminal receives the second instruction information.

[0160] Step C3: The first terminal sends a fourth message to the first device.

[0161] The fourth message is used to notify the first session and the second bearer resource release is complete.

[0162] Step C4: The first device receives the fourth message.

[0163] Figure 8 is a schematic diagram of an AI data session release process provided in this disclosure; as shown in Figure 8:

[0164] Step 801: The base station sends an AI data session release instruction message to the terminal.

[0165] Step 802: The terminal sends an AI / MI data session release completion message to the base station.

[0166] In some embodiments, the method provided by this disclosure includes the following:

[0167] Step D1: The fourth device, i.e., RIC, sends a third instruction message to the first device.

[0168] The third instruction message is used to instruct the establishment of lifecycle tasks for artificial intelligence or machine learning models. The third instruction information includes model identifier, model size, model training or inference, performance requirements for lifecycle tasks of artificial intelligence or machine learning models, model data acquisition configuration information, and information of the acquisition terminal.

[0169] Here, the performance requirements for the lifecycle tasks of artificial intelligence or machine learning models include training latency, inference latency, model transmission latency, and dataset transmission latency.

[0170] Here, the model data acquisition configuration information includes the base station identifier, data type, sampling period, and reporting period.

[0171] Here, the collected terminal information includes the terminal identifier, the type of terminal data collected, and the sampling period.

[0172] Step D2: The first device receives the third instruction message.

[0173] Step D3: The first device sends the fifth message to the fourth device.

[0174] The fifth message is used to notify that the lifecycle tasks of an artificial intelligence or machine learning model have been completed.

[0175] Step D4: The fourth device receives the fifth message.

[0176] Step D5: The first device receives artificial intelligence or machine learning data sent directly by the terminal through a session, or sent through a third device.

[0177] The session includes the first session; the terminal includes the first terminal.

[0178] Step D6: The first device sends artificial intelligence or machine learning data to the fourth device.

[0179] Step D7: The fourth device receives artificial intelligence or machine learning data and collects the artificial intelligence or machine learning data to train an artificial intelligence or machine learning model.

[0180] The following will describe an exemplary application of the embodiments of this disclosure in a practical application scenario.

[0181] Scenario 1: Base stations utilize idle computing resources or computing power cards to train physical layer AI / ML models. AI / ML model training requires terminal-side data and information from neighboring base stations. In addition to the aforementioned AI / ML context, AI data session context, and AI data session establishment, this scenario also illustrates the process of base stations transmitting AI data session information for AI / ML model training. Figure 9 is a flowchart illustrating a data transmission method provided in this embodiment; as shown in Figure 9:

[0182] Step 901: Base station 1 initiates the AI / ML model training lifecycle management process, and performs AI / ML model selection and model deployment.

[0183] Step 902: Base station 1 collects training data for the AI / ML model.

[0184] Step 903: Base station 1 sends an AI / ML data session establishment instruction to terminals i to j, or through base stations o to q or o' to n or m' to n'.

[0185] Here, the AI / ML data session establishment instructions include the wireless bearer information to be established, such as the wireless bearer meeting the bandwidth and / or latency requirements for model training data acquisition.

[0186] Here, base station 1 can transmit the AI / ML data session establishment instruction to terminal m~n or m'~n' through the base station AI / ML data session establishment instruction established with base station o~q or o'~q'.

[0187] Step 904: Terminals m to n, or m' to n', send an AI / ML data message to base station 1 via base station o to q or o' to q', or terminal i to j. The AI / ML data session is now established.

[0188] Here, after base station o~q or o'~q' sends the AI / ML session establishment completion message to terminal m~n or m'~n', it sends the neighbor station AI / ML session establishment completion message to base station 1.

[0189] Step 905: Base station 1 sends AI / ML data acquisition configuration to terminals i to j, and through base stations o to q or o' to q' to terminals m to n or m' to n'.

[0190] Step 906: Terminals i to j report AI / ML data to base station 1.

[0191] Step 907: Terminals m to n collect and report data to base station 1 through base station o to q.

[0192] Furthermore, the terminal performs data processing and AI / ML model training based on the received information.

[0193] It should be noted that there are no strict time requirements for the process from data acquisition to model training between different base stations. The AI / ML model training process can begin as soon as the data from the terminal and the base station arrive. The input parameters of the AI / ML model training data must ensure that the data from one base station and one terminal are used as training data. For example, at a certain time t, the RSRP, RSSI and location information of terminal 1, and the PRB load and spectral efficiency of the base station accessed by terminal 1 are used as training input parameters to ensure the consistency of the extracted terminal and base station channel data.

[0194] Scenario 2: Cloud-based base station scenario. The O-RAN network provides model training through RIC. The RIC configures the AI / ML model training data acquisition requirements and training connection QoS requirements to the base station via the E2 interface. The base station establishes an AI data session with the terminal. At this time, the AI ​​data session is from the terminal to the base station via the air interface, and from the base station to the RIC via the E2 interface. The RIC can internally transmit training data to the computational tasks for AI / ML model training. Since the AI ​​data session establishment process between the base station and the terminal is similar to that above, this scenario mainly describes the E2 interface interaction process between the RIC and the base station. Figure 10 is a flowchart illustrating an AI / ML model training use case based on RIC provided in this embodiment of the disclosure; as shown in Figure 10:

[0195] Step 1001: RIC sends an AI / ML model lifecycle task establishment instruction to base station m or base station n.

[0196] The instruction carries information such as model identifier, model size, model training or inference, task performance requirements (e.g., training latency, inference latency, model transmission latency, dataset transmission latency); model data acquisition configuration information, such as the base station identifier, data type, sampling period, reporting period, etc.; and the acquired terminal data, which may carry the terminal identifier, acquired terminal data type, sampling period, etc.

[0197] Step 1002: Base station m or base station n sends an AI / ML model lifecycle task establishment instruction to the RIC.

[0198] Step 1003: Base station m or base station n establishes an AI data session with the terminal and collects data.

[0199] Step 1004: Base station m or base station n reports the collected data to the RIC.

[0200] Step 1005: RIC performs AI / ML model training.

[0201] The embodiments of this disclosure provide a first device that can be used to implement a data transmission method provided in the embodiment corresponding to FIG2. Referring to FIG11, the first device 1100 includes:

[0202] The first transmitting part 1101 is configured to send a first message to a first terminal; wherein the first message is used to establish one or more first sessions; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier in the first device side for training or inference through the wireless bearer corresponding to the first identifier.

[0203] The first receiving part 1102 is configured to receive a second message sent by the first terminal; wherein the second message is used to indicate that the first session has been established.

[0204] In other embodiments of this disclosure, the first processing unit 1103 is configured to establish a first session for artificial intelligence or machine learning lifecycle management for a first terminal; wherein the first session includes first information; the first information includes quality requirements for artificial intelligence or machine learning training / inference, quality of service requirements for wireless connectivity, artificial intelligence or machine learning model information, base station identifier where the artificial intelligence or machine learning model is located, and wireless bearer information for artificial intelligence or machine learning data; the first information is recorded in a context; wherein the context includes the context of the first terminal on the first device side and the first session context established for the first session; and data wireless bearer resources are allocated based on the first information.

[0205] In other embodiments of this disclosure, the first sending portion 1101 is configured to send a context to the second device so that the second device can establish an artificial intelligence or machine learning data transmission session for the first terminal based on the context.

[0206] In other embodiments of this disclosure, the first processing unit 1103 is configured to establish a transmission channel for artificial intelligence or machine learning data between the first device and the second device based on the context.

[0207] In other embodiments of this disclosure, the first processing unit 1103 is configured to sense the artificial intelligence or machine learning capabilities of the terminal; wherein the terminal includes a first terminal; and a transmission session and wireless bearer for artificial intelligence or machine learning data are established for a second terminal with artificial intelligence or machine learning capabilities.

[0208] In other embodiments of this disclosure, the first session includes one or more of the following: a session identifier; an identifier of the first terminal; an identifier of the artificial intelligence or machine learning model; the processing method of the model; whether the artificial intelligence or machine learning data to be transmitted needs integrity protection or encryption; the integrity protection or encryption algorithm that can be used for the artificial intelligence or machine learning data that needs integrity protection or encryption; whether the artificial intelligence or machine learning data to be transmitted needs compression or encoding; the encoding or compression method that can be used for the artificial intelligence or machine learning data that needs compression or encoding; wireless bearer information; and artificial intelligence or machine learning data.

[0209] In other embodiments of this disclosure, the first processing unit 1103 is configured to update the first session if the first session meets a first condition; wherein, the first session meeting the first condition includes one or more of the following: the session identifier is adjusted; the identifier of the first terminal is adjusted; the identifier of the artificial intelligence or machine learning model is adjusted; artificial intelligence or machine learning data that requires integrity protection or encryption is adjusted to data that does not require integrity protection or encryption; artificial intelligence or machine learning data that does not require integrity protection or encryption is adjusted to data that requires integrity protection or encryption; artificial intelligence or machine learning data that requires compression or encoding is adjusted to data that does not require compression or encoding; artificial intelligence or machine learning data that does not require compression or encoding is adjusted to data that requires compression or encoding; the integrity protection or encryption algorithm used for artificial intelligence or machine learning data that requires integrity protection or encryption is adjusted; the encoding or compression method used for artificial intelligence or machine learning data that requires compression or encoding is adjusted; and the wireless bearer resources are adjusted.

[0210] In other embodiments of this disclosure, the first sending portion 1101 is configured to send a first indication message to a first terminal; wherein, the first indication message is used to instruct the first device to update the first bearer resource for artificial intelligence or machine learning data;

[0211] The first receiving part 1102 is configured to receive a third message sent by the first terminal; wherein the third message is used to notify the first bearer resource update is complete.

[0212] In other embodiments of this disclosure, the first sending portion 1101 is configured to send second indication information to the first terminal; wherein the second indication information is used to indicate the release of the first session and the release of the second bearer resource for artificial intelligence or machine learning data;

[0213] The first receiving part 1102 is configured to receive a fourth message sent by the first terminal; wherein the fourth message is used to notify the first session and the completion of the release of the second bearer resources.

[0214] In other embodiments of this disclosure, the first session context includes one or more of the following: an artificial intelligence or machine learning model identifier; an artificial intelligence or machine learning model vendor; an artificial intelligence or machine learning model version; an artificial intelligence or machine learning model type; an artificial intelligence or machine learning model size; an artificial intelligence or machine learning model algorithm complexity; an artificial intelligence or machine learning model computing resource requirements; an artificial intelligence or machine learning model training data volume; artificial intelligence or machine learning model training input parameter information; an artificial intelligence or machine learning model inference data volume; artificial intelligence or machine learning model inference input parameter information; terminal information associated with the first session; and base station information associated with the first session.

[0215] In other embodiments of this disclosure, the first processing unit 1103 is configured to initiate an artificial intelligence or machine learning model training lifecycle management process, perform artificial intelligence or machine learning model selection and model deployment, and collect training data for the selected or to-be-deployed artificial intelligence or machine learning model.

[0216] The first receiving unit 1102 is configured to receive artificial intelligence or machine learning data sent directly by the terminal through a session or sent through a third device; the session includes a first session; the terminal includes a first terminal;

[0217] The first processing unit 1103 is configured to train an artificial intelligence or machine learning model based on artificial intelligence or machine learning data.

[0218] In other embodiments of this disclosure, the first receiving portion 1102 is configured to receive a third indication message sent by a fourth device; wherein the third indication message is used to indicate the establishment of an artificial intelligence or machine learning model lifecycle task; the third indication information includes model identifier, model size, model training or inference, performance requirements of the artificial intelligence or machine learning model lifecycle task; model data acquisition configuration information; and acquired terminal information;

[0219] The first transmitting part 1101 is configured to send a fifth message to the fourth device; wherein the fifth message is used to notify that the life cycle task of the artificial intelligence or machine learning model has been established.

[0220] The first receiving unit 1102 is configured to receive artificial intelligence or machine learning data sent directly by the terminal through a session or sent through a third device; the session includes a first session; the terminal includes a first terminal;

[0221] The first transmitting part 1101 is configured to send the collected artificial intelligence or machine learning data to the fourth device so that the fourth device can train an artificial intelligence or machine learning model based on the collected artificial intelligence or machine learning data.

[0222] The description of the above device embodiments is similar to that of the above method embodiments, and has similar beneficial effects. For technical details not disclosed in the device embodiments of this disclosure, please refer to the description of the method embodiments of this disclosure for understanding.

[0223] It should be noted that, in the embodiments of this disclosure, if the above-described data transmission method is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this disclosure, or the part that contributes to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a terminal device to execute all or part of the methods of the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this disclosure are not limited to any specific hardware and software combination.

[0224] This disclosure provides a first terminal that can be used to implement a data transmission method provided in the embodiment corresponding to FIG2. Referring to FIG12, the first terminal 1200 includes:

[0225] The second receiving part 1201 is configured to receive a first message sent by the first device; wherein the first message is used to establish one or more first sessions; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier in the first device side for training or inference through the wireless bearer corresponding to the first identifier.

[0226] The second sending section 1202 is configured to send a second message to the first device; wherein the second message is used to indicate that the first session has been established.

[0227] In other embodiments of this disclosure, the second receiving portion 1201 is configured to receive a first indication message sent by the first device; wherein the first indication message is used to instruct the first device to update the first bearer resource for artificial intelligence or machine learning data;

[0228] The second sending part 1202 is configured to send a third message to the first terminal; wherein the third message is used to notify the first bearer resource update is complete.

[0229] In other embodiments of this disclosure, the second receiving portion 1201 is configured to receive second indication information sent by the first device; wherein the second indication information is used to indicate the release of the first session and the release of a second bearer resource for artificial intelligence or machine learning data;

[0230] The second sending part 1202 is configured to send a fourth message to the first device; wherein the fourth message is used to notify the first session and the completion of the release of the second bearer resources.

[0231] The description of the above device embodiments is similar to that of the above method embodiments, and has similar beneficial effects. For technical details not disclosed in the device embodiments of this disclosure, please refer to the description of the method embodiments of this disclosure for understanding.

[0232] It should be noted that, in the embodiments of this disclosure, if the above-described data transmission method is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this disclosure, or the part that contributes to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a network device to execute all or part of the methods of the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, external hard drives, ROMs, magnetic disks, or optical disks. Thus, the embodiments of this disclosure are not limited to any specific hardware and software combination.

[0233] This disclosure provides a second device that can be used to implement a data transmission method provided in this disclosure. Referring to FIG13, the second device 1300 includes:

[0234] The third receiving section 1301 is configured to receive a context sent by the first device; the context includes the context of the first terminal on the first device side and the first session context established for the first session between the first device and the first terminal; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier on the first device side for training or inference through the wireless bearer corresponding to the first identifier;

[0235] The third processing unit 1302 is configured to establish an artificial intelligence or machine learning data transmission session for the first terminal, establish a transmission channel for artificial intelligence or machine learning data between the first device and the second device, and deploy / activate an artificial intelligence or machine learning model used by the first device, based on the context.

[0236] The description of the above device embodiments is similar to that of the above method embodiments, and has similar beneficial effects. For technical details not disclosed in the device embodiments of this disclosure, please refer to the description of the method embodiments of this disclosure for understanding.

[0237] It should be noted that, in the embodiments of this disclosure, if the above-described data transmission method is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this disclosure, or the part that contributes to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a network device to execute all or part of the methods of the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, external hard drives, ROMs, magnetic disks, or optical disks. Thus, the embodiments of this disclosure are not limited to any specific hardware and software combination.

[0238] This disclosure provides a fourth device that can be used to implement a data transmission method provided in this disclosure. Referring to FIG14, the fourth device 1400 includes:

[0239] The fourth sending part 1401 is configured to send a third indication message to the first device; wherein the third indication message is used to indicate the establishment of an artificial intelligence or machine learning model lifecycle task; the third indication information includes model identifier, model size, model training or inference, performance requirements of the artificial intelligence or machine learning model lifecycle task; model data acquisition configuration information; and acquisition terminal information;

[0240] The fourth receiving section 1402 is configured to receive a fifth message sent by the first device; wherein the fifth message is used to notify that the lifecycle task of the artificial intelligence or machine learning model has been established.

[0241] The fourth receiving section 1402 is configured to receive artificial intelligence or machine learning data sent by the first device; wherein the artificial intelligence or machine learning data is sent directly by the terminal through a session, or sent through a third device; the session includes a first session; the terminal includes a first terminal;

[0242] The fourth processing section 1403 is configured to train an artificial intelligence or machine learning model based on the collected artificial intelligence or machine learning data.

[0243] The description of the above device embodiments is similar to that of the above method embodiments, and has similar beneficial effects. For technical details not disclosed in the device embodiments of this disclosure, please refer to the description of the method embodiments of this disclosure for understanding.

[0244] It should be noted that, in the embodiments of this disclosure, if the above-described data transmission method is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this disclosure, or the part that contributes to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a network device to execute all or part of the methods of the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, external hard drives, ROMs, magnetic disks, or optical disks. Thus, the embodiments of this disclosure are not limited to any specific hardware and software combination.

[0245] Figure 15 is a schematic structural diagram of a communication device 1500 provided in an embodiment of this disclosure. This communication device can be a first terminal, a first device, a second device, or a fourth device. The communication device 1500 shown in Figure 15 includes a processor 1510, which can call and run computer programs from memory to implement the methods in the embodiments of this disclosure.

[0246] Optionally, as shown in FIG15, the communication device 1500 may further include a memory 1520. The processor 1510 may retrieve and run computer programs from the memory 1520 to implement the methods in the embodiments of this disclosure.

[0247] The memory 1520 can be a separate device independent of the processor 1510, or it can be integrated into the processor 1510.

[0248] Optionally, as shown in FIG15, the communication device 1500 may further include a transceiver 1530, and the processor 1510 may control the transceiver 1530 to communicate with other devices. Specifically, it may send information or data to other devices or receive information or data sent by other devices.

[0249] The transceiver 1530 may include a transmitter and a receiver. The transceiver 1530 may further include an antenna, and the number of antennas may be one or more.

[0250] Optionally, the communication device 1500 may specifically be the first device / first terminal / second device / fourth device in the embodiments of this disclosure, and the communication device 1500 may implement the corresponding processes implemented by the first device / first terminal / second device / fourth device in the various methods of the embodiments of this disclosure. For the sake of brevity, it will not be described in detail here.

[0251] This disclosure also provides a computer program product, including a computer program that can be executed by the processor 1510 of the communication device 1500 to perform the steps described in any of the foregoing methods.

[0252] It should be understood that the processor in this disclosure embodiment may be an integrated circuit chip with signal processing capabilities. In implementation, the steps of the above method embodiments can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor described above can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in this disclosure embodiment. The general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the methods disclosed in the embodiments of this disclosure can be directly embodied as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0253] As one embodiment, the processor may include one or more general-purpose central processing units (CPUs). Each of these processors may be a single-core processor or a multi-core processor. Here, "processor" may refer to one or more devices, circuits, and / or processing cores used for processing data (e.g., executing instructions).

[0254] It is understood that the memory in the embodiments of this disclosure can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be ROM, Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), or flash memory. The volatile memory can be Random Access Memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DR RAM). It should be noted that the memory used in the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0255] This disclosure also provides a computer-readable storage medium for storing computer programs.

[0256] Optionally, the computer-readable storage medium can be applied to the first device / first terminal / second device / fourth device in the embodiments of this disclosure, and the computer program causes the computer to execute the corresponding processes implemented by the first device / first terminal / second device / fourth device in the various methods of the embodiments of this disclosure. For the sake of brevity, it will not be described in detail here.

[0257] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product.

[0258] A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to embodiments of this disclosure is generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a server or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).

[0259] The data transmission method, first device, first terminal, second device, fourth device, computer-readable storage medium, and computer program product provided in the embodiments of this disclosure have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this disclosure. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this disclosure. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this disclosure. Therefore, the content of this specification should not be construed as a limitation of this disclosure.

[0260] It should be understood that the terms "an embodiment," "an embodiment," "an embodiment of this disclosure," "the foregoing embodiment," "some implementations," or "some embodiments" mentioned throughout the specification mean that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this disclosure. Therefore, the phrases "an embodiment," "an embodiment," "an embodiment of this disclosure," "the foregoing embodiment," "some implementations," or "some embodiments" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this disclosure, the sequence numbers of the above processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this disclosure. The sequence numbers of the above embodiments of this disclosure are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0261] Unless otherwise specified, any step performed by the first device / first terminal / second device / fourth device in the embodiments of this disclosure may be executed by the processor of the first device / first terminal / second device / fourth device. Unless otherwise specified, the embodiments of this disclosure do not limit the order in which the first device / first terminal / second device / fourth device performs the following steps. Furthermore, the methods used to process data in different embodiments may be the same or different methods.

[0262] In the several embodiments provided in this disclosure, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components may be combined, or integrated into another system, or some features may be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed may be through some interfaces, and the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0263] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0264] In addition, each functional unit in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0265] The methods disclosed in the several method embodiments provided in this disclosure can be arbitrarily combined to obtain new method embodiments without conflict. The features disclosed in the several product embodiments provided in this disclosure can be arbitrarily combined to obtain new product embodiments without conflict. The features disclosed in the several method or device embodiments provided in this disclosure can be arbitrarily combined to obtain new method embodiments or device embodiments without conflict.

[0266] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0267] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer storage medium. Based on this understanding, the technical solutions of the embodiments of this disclosure, or the parts that contribute to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0268] The singular forms “a,” “the,” and “the” used in this disclosure and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise.

[0269] It should be noted that in the various embodiments involved in this disclosure, all steps or some steps may be performed, as long as a complete technical solution can be formed.

[0270] The above description is merely an embodiment of this disclosure, but the scope of protection of this disclosure is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this disclosure should be included within the scope of protection of this disclosure. Therefore, the scope of protection of this disclosure should be determined by the scope of the claims.

Claims

1. A data transmission method applied to a first device, the method comprising: Send a first message to a first terminal; wherein the first message is used to establish one or more first sessions; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier in the first device side for training or inference through the wireless bearer corresponding to the first identifier; Receive a second message sent by the first terminal; wherein the second message is used to indicate that the first session has been established.

2. The method according to claim 1, wherein, The method further includes: A first session for artificial intelligence or machine learning lifecycle management is established for the first terminal; wherein, the first session includes first information; the first information includes quality requirements for artificial intelligence or machine learning training / inference, quality of service requirements for wireless connectivity, artificial intelligence or machine learning model information, base station identifier where the artificial intelligence or machine learning model is located, and wireless bearer information for artificial intelligence or machine learning data; Record the first information into the context; wherein, the context includes the context of the first terminal on the first device side and the first session context established for the first session; Based on the first information, allocate data wireless bearer resources.

3. The method according to claim 2, wherein, The method further includes: The context is sent to the second device so that the second device can establish an artificial intelligence or machine learning data transmission session for the first terminal based on the context.

4. The method according to claim 2, wherein, The method further includes: Based on the aforementioned context, a data transmission channel for artificial intelligence or machine learning is established between the first device and the second device.

5. The method according to claim 1, wherein, The method further includes: The sensing terminal has artificial intelligence or machine learning capabilities; wherein, the terminal includes a first terminal; Establish transmission sessions and wireless bearers for AI or machine learning data for second terminals with AI or machine learning capabilities.

6. The method according to claim 1, wherein, The first session includes one or more of the following: Session identifier; The identifier of the first terminal; Identifiers for artificial intelligence or machine learning models; The model processing method; Does the artificial intelligence or machine learning data to be transmitted require integrity protection or encryption? Integrity protection or encryption algorithms that can be used for artificial intelligence or machine learning data that requires integrity protection or encryption; Does the artificial intelligence or machine learning data to be transmitted need to be compressed or encoded? Encoding or compression methods that can be used for artificial intelligence or machine learning data that needs to be compressed or encoded; Wireless bearer information; Artificial intelligence or machine learning data.

7. The method according to claim 1, wherein, The method further includes: If the first session satisfies the first condition, the first session is updated; wherein, the first session satisfying the first condition includes one or more of the following: The session identifier has been adjusted; The identifier of the first terminal has been adjusted; The labels for artificial intelligence or machine learning models have been adjusted; Adjust AI or machine learning data that requires integrity protection or encryption to be free of integrity protection or encryption; Adjust artificial intelligence or machine learning data that does not require integrity protection or encryption to require integrity protection or encryption. Adjust artificial intelligence or machine learning data that requires compression or encoding to be uncompressible or unencoded. Adjust artificial intelligence or machine learning data that does not require compression or encoding to require compression or encoding. The integrity protection or encryption algorithms used for artificial intelligence or machine learning data that require integrity protection or encryption have been adjusted. The encoding or compression method used for artificial intelligence or machine learning data that needs to be compressed or encoded has been adjusted; Adjustment of wireless bearer resources.

8. The method according to claim 7, wherein, The update of the first session includes: Send a first instruction message to the first terminal; wherein the first instruction message is used to instruct the first device to update the first bearer resource for artificial intelligence or machine learning data; Receive a third message sent by the first terminal; wherein the third message is used to notify the first bearer resource update is complete.

9. The method according to claim 1, wherein, The method further includes: Send a second instruction message to the first terminal; wherein the second instruction message is used to instruct the release of the first session and the release of a second bearer resource for artificial intelligence or machine learning data; The system receives a fourth message sent by the first terminal; wherein the fourth message is used to notify the first session and the second bearer resource release is complete.

10. The method according to claim 2, wherein, The first session context includes one or more of the following: Identifier for artificial intelligence or machine learning models; Provider of artificial intelligence or machine learning models; Version of the artificial intelligence or machine learning model; Types of artificial intelligence or machine learning models; Size of artificial intelligence or machine learning models; The complexity of artificial intelligence or machine learning model algorithms; Computational resource requirements for artificial intelligence or machine learning models; The amount of training data for artificial intelligence or machine learning models; Input parameters for training artificial intelligence or machine learning models; The amount of data used for inference by artificial intelligence or machine learning models; Inference input parameter information for artificial intelligence or machine learning models; The terminal information associated with the first session; The base station information associated with the first session.

11. The method according to claim 1, wherein, The method further includes: Initiate the lifecycle management process for training artificial intelligence or machine learning models, and perform artificial intelligence or machine learning model selection and deployment. Collect training data for the selected or to be deployed artificial intelligence or machine learning model; The receiving terminal sends artificial intelligence or machine learning data directly through a session or through a third device; the session includes the first session; the terminal includes the first terminal; Training artificial intelligence or machine learning models based on artificial intelligence or machine learning data.

12. The method according to claim 1, wherein, The method further includes: Receive a third instruction message sent by a fourth device; wherein the third instruction message is used to instruct the establishment of an artificial intelligence or machine learning model lifecycle task; the third instruction information includes model identifier, model size, model training or inference, performance requirements of the artificial intelligence or machine learning model lifecycle task; model data acquisition configuration information; and acquisition terminal information; Send a fifth message to the fourth device; wherein the fifth message is used to notify that the lifecycle task of the artificial intelligence or machine learning model has been established. The receiving terminal sends artificial intelligence or machine learning data directly through a session or through a third device; the session includes the first session; the terminal includes the first terminal; The collected artificial intelligence or machine learning data is sent to the fourth device so that the fourth device can train an artificial intelligence or machine learning model based on the collected artificial intelligence or machine learning data.

13. A data transmission method applied to a first terminal, the method comprising: The first terminal receives a first message sent by a first device; wherein the first message is used to establish one or more first sessions; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier in the first device side for training or inference through the wireless bearer corresponding to the first identifier. A second message sent to the first device; wherein the second message is used to indicate that the first session has been established.

14. The method according to claim 13, wherein, The method further includes: Receive a first instruction message sent by the first device; wherein the first instruction message is used to instruct the first device to update the first bearer resource for artificial intelligence or machine learning data; A third message is sent to the first terminal; wherein the third message is used to notify the first bearer resource update is complete.

15. The method according to claim 13, wherein, The method further includes: Receive a second indication message sent by the first device; wherein the second indication message is used to indicate the release of the first session and the release of a second bearer resource for artificial intelligence or machine learning data; A fourth message is sent to the first device; wherein the fourth message is used to notify the first session and the second bearer resource release is complete.

16. A data transmission method applied to a second device, the method comprising: Receive the context sent by the first device; The context includes the context of the first terminal on the first device side and the first session context established for the first session between the first device and the first terminal; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier on the first device side for training or inference through the wireless bearer corresponding to the first identifier. Based on the aforementioned context, an artificial intelligence or machine learning data transmission session is established for the first terminal, a transmission channel for artificial intelligence or machine learning data is established between the first device and the second device, and an artificial intelligence or machine learning model used by the first device is deployed / activated.

17. A data transmission method applied to a fourth device, the method comprising: A third instruction message is sent to the first device; wherein the third instruction message is used to instruct the establishment of an artificial intelligence or machine learning model lifecycle task; the third instruction information includes model identifier, model size, model training or inference, performance requirements of the artificial intelligence or machine learning model lifecycle task; model data acquisition configuration information; and acquisition terminal information; Receive a fifth message sent by the first device; wherein the fifth message is used to notify that the lifecycle task of the artificial intelligence or machine learning model has been established. Receive artificial intelligence or machine learning data sent by the first device; wherein the artificial intelligence or machine learning data is sent directly by the terminal through a session, or sent through a third device; the session includes the first session; the terminal includes the first terminal; Based on the collected artificial intelligence or machine learning data, artificial intelligence or machine learning models are trained.

18. A first device, the first device comprising: The first sending part is configured to send a first message to a first terminal; wherein the first message is used to establish one or more first sessions; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier in the first device side for training or inference through the wireless bearer corresponding to the first identifier. The first receiving part is configured to receive a second message sent by the first terminal; wherein the second message is used to indicate that the first session has been established.

19. A first terminal, the first terminal comprising: The second receiving part is configured to receive a first message sent by the first device; wherein the first message is used to establish one or more first sessions; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier in the first device side for training or inference through the wireless bearer corresponding to the first identifier. The second sending part is configured to send a second message to the first device; wherein the second message is used to indicate that the first session has been established.

20. A second device, the second device comprising: The third receiving section is configured to receive context sent by the first device; The context includes the context of the first terminal on the first device side and the first session context established for the first session between the first device and the first terminal; the first session includes one or more wireless bearers for transmitting artificial intelligence or machine learning data of arbitrary size; the wireless bearers of the first session are identified by the identifier of the artificial intelligence or machine learning model; the first terminal directly accesses the artificial intelligence or machine learning data corresponding to the first identifier on the first device side for training or inference through the wireless bearer corresponding to the first identifier. The third processing unit is configured to, based on the context, establish an artificial intelligence or machine learning data transmission session for the first terminal, establish a transmission channel for artificial intelligence or machine learning data between the first device and the second device, and deploy / activate an artificial intelligence or machine learning model used by the first device.

21. A fourth device, the fourth device comprising: The fourth sending part is configured to send a third indication message to the first device; wherein, the third indication message is used to indicate the establishment of an artificial intelligence or machine learning model lifecycle task; the third indication information includes model identifier, model size, model training or inference, performance requirements of the artificial intelligence or machine learning model lifecycle task; model data acquisition configuration information; and acquisition terminal information; The fourth receiving part is configured to receive a fifth message sent by the first device; wherein the fifth message is used to notify that the lifecycle task of the artificial intelligence or machine learning model has been established. The fourth transmitting section is configured to receive artificial intelligence or machine learning data transmitted by the first device; wherein the artificial intelligence or machine learning data is transmitted directly by the terminal through a session, or transmitted through a third device; the session includes the first session; the terminal includes the first terminal; The fourth processing section is configured to train artificial intelligence or machine learning models based on the collected artificial intelligence or machine learning data.

22. A communication device, the communication device comprising: Memory, used to store executable instructions; The processor, when executing executable instructions stored in the memory, implements the data transmission method of any one of claims 1 to 12, or the data transmission method of any one of claims 13 to 15, or the data transmission method of claim 16, or the data transmission method of claim 17.

23. A computer-readable storage medium storing one or more programs, said one or more programs being executable by one or more processors to implement the data transmission method of any one of claims 1 to 12, or the data transmission method of any one of claims 13 to 15, or the data transmission method of claim 16, or the data transmission method of claim 17.

24. A computer program product comprising a computer program, which, when executed by a processor, implements the data transmission method of any one of claims 1 to 12, or the data transmission method of any one of claims 13 to 15, or the data transmission method of claim 16, or the data transmission method of claim 17.

Images