Electronic device and method for entity alignment in vertical federated learning

Abstract

The present disclosure relates to an electronic device and method for entity alignment in vertical federated learning. An electronic device for a base station side is described, and comprises a processing circuit. The processing circuit is configured to: acquire, from a terminal device, a user requirement for an entity alignment task based on vertical federated learning; determine, on the basis of the user requirement from the terminal device and a network state between a base station and the terminal device, an entity alignment scheme for performing the entity alignment task; and indicate the determined entity alignment scheme to the terminal device.

Description

Electronic devices and methods for entity alignment in vertical federated learning

[0001] Priority Statement

[0002] This application claims priority to Chinese patent application filed on December 26, 2024, application number 202411943160.X, entitled “Electronic device and method for entity alignment for vertical federated learning”, the entire contents of which are incorporated herein by reference. Technical Field

[0003] This disclosure relates to the field of machine learning, and more specifically, to electronic devices and methods for entity alignment in vertical federated learning. Background Technology

[0004] In recent years, federated learning (FL) technology has been increasingly widely used in wireless communication networks. Federated learning is a distributed machine learning technique that prioritizes privacy and data security. It allows multiple participants to collaboratively train a global model without directly exchanging data. Based on its data distribution, federated learning can be categorized into, for example, horizontal federated learning, vertical federated learning, and federated transfer learning.

[0005] In Vertical Federated Learning (VFL), there is little overlap in features but a lot in samples among the data from different participants. Therefore, entity alignment is needed among the participants to identify overlapping samples. To ensure data security, entity alignment (also known as sample alignment, entity resolution, entity matching, record joining, etc.) requires each participant to align (match) features belonging to the same entity (sample) without exposing its own information.

[0006] There is a need for more efficient electronic devices and methods for entity alignment. Summary of the Invention

[0007] A brief overview of this disclosure is given below to provide a basic understanding of some aspects of it. However, it should be understood that this overview is not an exhaustive summary of this disclosure. It is not intended to identify key or essential parts of this disclosure, nor is it intended to limit the scope of this disclosure. Its purpose is merely to present certain concepts of this disclosure in a simplified form as a prelude to the more detailed description that follows.

[0008] Among entity alignment techniques known to the inventors, the Private Set Intersection (PSI) technique is typically used to obtain the intersection between the data samples of each participant. PSI implementations can be based on public-key encryption, key exchange, unintentional transmission protocols, and homomorphic encryption, among others.

[0009] However, known entity-aligned encryption schemes typically employ a "one-size-fits-all" approach (i.e., always using the same scheme regardless of the scenario), without being customized to the requirements of different participants or the state of the network. In practical applications, different participants may have varying requirements regarding data characteristics, computing resources, and privacy protection. A one-size-fits-all encryption scheme cannot flexibly address these diverse needs, potentially leading to inefficiency or insufficient privacy protection. Furthermore, in wireless network environments, dynamic changes in network conditions (e.g., bandwidth fluctuations, latency, and packet loss rates) further exacerbate this problem. Due to the lack of targeted adjustments to the encryption process, known entity-aligned encryption schemes may perform poorly in response to these dynamic changes, resulting in decreased communication efficiency and system performance instability.

[0010] Therefore, to address one or more of the aforementioned issues, it is advisable to further gather user requirements for the entity alignment task from all participating parties and monitor network status. This allows for a comprehensive consideration of user requirements and network conditions, enabling the flexible selection of an appropriate entity alignment scheme. Consequently, the potential negative impact of network conditions on entity alignment can be reduced, ensuring that user needs are met.

[0011] According to one aspect of this disclosure, an electronic device for a base station side is provided. The electronic device may include processing circuitry. The processing circuitry may be configured to: acquire user requirements from a terminal device for an entity alignment task based on vertical federated learning; determine an entity alignment scheme for performing the entity alignment task based on the user requirements of the terminal device and the network state between the base station and the terminal device; and indicate the determined entity alignment scheme to the terminal device.

[0012] According to another aspect of this disclosure, an electronic device for a terminal device is provided. The electronic device may include processing circuitry. This processing circuitry may be configured to: send a user request to a base station for an entity alignment task based on vertical federated learning; and receive an entity alignment scheme from the base station, the entity alignment scheme being determined by the base station based on the user request and the network state between the base station and the terminal device.

[0013] According to another aspect of this disclosure, a method for entity alignment in vertical federated learning is provided. The method may include: obtaining user requirements from a terminal device for an entity alignment task based on vertical federated learning; determining an entity alignment scheme for performing the entity alignment task based on the user requirements of the terminal device and the network state between the base station and the terminal device; and instructing the determined entity alignment scheme to the terminal device.

[0014] According to another aspect of this disclosure, a method for entity alignment in vertical federated learning is provided. The method may include: sending a user request to a base station for an entity alignment task based on vertical federated learning; and receiving an entity alignment scheme from the base station, the entity alignment scheme being determined by the base station based on the user request and the network state between the base station and the terminal device.

[0015] According to another aspect of this disclosure, a computer-readable storage medium is provided, including executable instructions that, when executed by an information processing apparatus, cause the information processing apparatus to perform a method for entity alignment for vertical federated learning according to this disclosure.

[0016] According to another aspect of this disclosure, a computer program product is provided, including a computer program that, when run by a processor, causes the processor to perform a method for entity alignment for vertical federated learning according to this disclosure. Attached Figure Description

[0017] The accompanying drawings, which form part of this specification, illustrate embodiments of this disclosure and, together with the specification, serve to explain the principles of this disclosure.

[0018] This disclosure will be more clearly understood with reference to the accompanying drawings and the following detailed description, wherein:

[0019] Figure 1 is a schematic diagram illustrating an example environment for vertical federated learning according to an embodiment of the present disclosure;

[0020] Figure 2 is an exemplary configuration block diagram of a base station-side electronic device for entity alignment in vertical federated learning according to an embodiment of the present disclosure.

[0021] Figure 3 is an exemplary flowchart illustrating a method for entity alignment in vertical federated learning according to an embodiment of the present disclosure;

[0022] Figure 4 is an exemplary configuration block diagram of an electronic device on the terminal device side for entity alignment in vertical federated learning according to an embodiment of the present disclosure.

[0023] Figure 5 is an exemplary flowchart illustrating a method for entity alignment in vertical federated learning according to an embodiment of the present disclosure;

[0024] Figure 6 is a schematic diagram illustrating an example of a user request according to an embodiment of the present disclosure and an example of a first criterion for classifying the user request.

[0025] Figure 7 is a schematic diagram illustrating an example of a network state according to an embodiment of the present disclosure and an example of a second criterion for classifying the network state.

[0026] Figure 8 is a schematic diagram illustrating the characteristics and levels of an example entity alignment encryption scheme according to an embodiment of the present disclosure;

[0027] Figure 9A is an example flowchart illustrating an entity alignment process according to an embodiment of the present disclosure;

[0028] Figure 9B is a schematic diagram illustrating information interaction between terminal devices during entity alignment according to an embodiment of the present disclosure;

[0029] Figure 10 is a sequence diagram illustrating the signaling interactions between a terminal device and a base station-side device for entity alignment in vertical federated learning according to an embodiment of the present disclosure.

[0030] Figure 11 illustrates an exemplary configuration of a computing device that can implement an embodiment of the present invention;

[0031] Figure 12 is a block diagram illustrating a first example of a schematic configuration of a gNB according to an embodiment of the present disclosure;

[0032] Figure 13 is a block diagram illustrating a second example of a schematic configuration of a gNB according to an embodiment of the present disclosure;

[0033] Figure 14 is a block diagram illustrating an example of a schematic configuration of a smartphone according to an embodiment of the present disclosure; and

[0034] Figure 15 is a block diagram illustrating an example of a schematic configuration of a car navigation device according to an embodiment of the present disclosure. Detailed Implementation

[0035] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present disclosure.

[0036] At the same time, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn according to actual scale.

[0037] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this disclosure or its application or use.

[0038] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0039] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0040] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0041] Figure 1 is a schematic diagram illustrating an example environment 100 for vertical federated learning according to an embodiment of the present disclosure. As shown in Figure 1, environment 100 may include a base station-side device 110 and a plurality of (two illustrated in Figure 1) terminal devices 120a, 120b. The base station-side device 110 may act as a coordinator in the federated learning task, while the plurality of terminal devices 120a, 120b may act as participants in the federated learning task. The base station-side device 110 and the plurality of terminal devices 120a, 120b may communicate with each other via network 130.

[0042] Base station-side device 110 is an electronic device located on the side of a base station (not shown in Figure 1). For example, base station-side device 110 may be mounted in the base station, be part of the base station, or be communicatively coupled to the base station. As a coordinator, base station-side device 110 can provide support or assistance to the activities of each participant. For example, base station-side device 110 can build a basic model for a federated learning task and send the basic structure and parameters of the model to each participant. Base station-side device 110 can also receive locally trained models from each participant and build a final global model based on these models. In some embodiments, base station-side device 110 may also have the functions of a participant or a part thereof. Note that in this document, in the context of describing federated learning or entity alignment, the terms "base station-side device" and "base station" are used interchangeably.

[0043] Additionally, base station-side device 110 can also provide support or assistance for entity alignment tasks in vertical federated learning. As described in more detail below, base station-side device 110 can indicate an entity alignment scheme for the entity alignment task to each participant and receive encrypted data as the result of entity alignment from each participant for subsequent federated learning. More specifically, base station-side device 110 can (e.g., via network 130) obtain user requests for the entity alignment task from each of a plurality of terminal devices 120a, 120b. Base station-side device 110 can determine an entity alignment scheme for performing the entity alignment task based on the user requests of the terminal devices and the network status of the network between the base station and the terminal devices (i.e., network 130). Then, base station-side device 110 can indicate the determined entity alignment scheme to the terminal devices. The processing of base station-side device 110 will be described in more detail below with reference to Figures 2, 3, 6 to 8, etc.

[0044] In some embodiments, the base station-side device 110 may include or be coupled to a component for monitoring the network status of the network 130. Thus, the base station-side device 110 can acquire network status information between the base station and terminal devices, such as bandwidth, latency, packet loss rate, throughput, jitter, etc., to determine an appropriate entity alignment scheme. In some embodiments, the base station-side device 110 can dynamically adjust the frequency of acquiring network status information based on real-time changes in network status. For example, in an embodiment where the base station-side device 110 is coupled to a separate network status monitoring component, when the base station-side device 110 senses that changes in network status are infrequent (e.g., the amount of change in network status does not exceed a predetermined threshold within a predetermined time interval), the base station-side device 110 can control the network status monitoring component to reduce the frequency at which it sends network status information. Therefore, unnecessary communication overhead can be reduced and the use of network resources can be optimized.

[0045] As described above, multiple terminal devices 120a and 120b can act as participants in the federated learning task. For example, terminal devices 120a and 120b can obtain the structure and parameters of the basic model for the federated learning task from the base station-side device 110, use their local data to train a local model, and upload the locally trained local model to the base station-side device 110. In the vertical federated learning environment 100, multiple terminal devices 120a and 120b each have feature data with different dimensions. These data do not overlap (or overlap only slightly) in the feature space, but may overlap (overlap relatively significantly) in the sample space. Therefore, it is necessary to first perform entity alignment among the multiple terminal devices 120a and 120b, for example, to identify and match samples belonging to the same entity, so as to utilize this part for subsequent federated learning tasks.

[0046] More specifically, in some embodiments, terminal devices 120a and 120b may have an interface such as a graphical user interface (GUI) to obtain user requirements for the entity alignment task, such as security and real-time requirements. Multiple terminal devices 120a and 120b may (e.g., via network 130) send user requirements for the entity alignment task to a base station. Each of the multiple terminal devices 120a and 120b may then receive an entity alignment scheme determined by the base station, perform the entity alignment task with other terminal devices according to the scheme, and send encrypted data as the result of the entity alignment task to the base station for subsequent vertical federated learning. The processing of the multiple terminal devices 120a and 120b will be described in more detail later with reference to Figures 4, 5, 9A, and 9B.

[0047] Network 130 is a communication network responsible for data transmission between base station-side equipment 110 and multiple terminal devices 120a, 120b. Network 130 can be a wireless network and / or a wired network. Examples of network 130 may include, but are not limited to, local area networks (LANs), wide area networks (WANs), metropolitan area networks (MANs), personal area networks (PANs), Wi-Fi networks, mobile communication networks (e.g., 4G, 5G, etc.), Internet of Things (IoT), and vehicle-to-everything (V2X) networks. Network 130 can be configured to ensure certain data security and efficiency during transmission.

[0048] Note that although Figure 1 shows communication between base station device 110 and multiple terminal devices via a network 130, this disclosure is not limited thereto. In an alternative embodiment, base station device 110 may communicate with each terminal device via a separate network. In this case, base station device 110 may obtain the network status of its network with each terminal device and may consider the network status with each terminal device to select an appropriate entity alignment scheme.

[0049] Next, with reference to Figures 2 and 3, the base station-side electronic device and method for entity alignment in vertical federated learning according to this disclosure will be further described.

[0050] Figure 2 illustrates an exemplary configuration block diagram of a base station-side electronic device 200 for entity alignment in vertical federated learning according to an embodiment of the present disclosure. The electronic device 200 may, for example, be used to implement the base station-side device 110 shown in Figure 1.

[0051] In some embodiments, the electronic device 200 may include processing circuitry 210. The processing circuitry 210 of the electronic device 200 provides various functions of the electronic device 200.

[0052] Processing circuitry 210 can refer to various implementations of digital circuitry, analog circuitry, or mixed-signal (analog and digital combination) circuitry that perform functions in a computing system. Processing circuitry can include, for example, circuitry such as integrated circuits (ICs), application-specific integrated circuits (ASICs), portions or circuitry of a single processor core, an entire processor core, a single processor, programmable hardware devices such as field-programmable gate arrays (FPGAs), and / or systems comprising multiple processors.

[0053] In some embodiments, the processing circuit 210 may include an information acquisition unit 212, a scheme determination unit 214, and a scheme indication unit 216, configured to perform the corresponding steps in the entity alignment method 300 for vertical federated learning shown in FIG3 below.

[0054] In some embodiments, the electronic device 200 may further include a memory (not shown). The memory of the electronic device 200 may store information generated by the processing circuitry 210, as well as programs and data used for the operation of the processing circuitry 210. The memory may be volatile memory and / or non-volatile memory. For example, the memory may include, but is not limited to, random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), read-only memory (ROM), and flash memory. Furthermore, the electronic device 200 may be implemented at the chip level, or it may be implemented at the device level by including other external components.

[0055] Figure 3 illustrates an exemplary flowchart of a method 300 for entity alignment in vertical federated learning according to an embodiment of the present disclosure. This method 300 can, for example, be used in an electronic device 200 as shown in Figure 2.

[0056] As shown in Figure 3, in step S310, the information acquisition unit 212 can acquire user requirements from the terminal devices for the entity alignment task based on vertical federated learning. The terminal devices can correspond to each of the multiple terminal devices 120a and 120b shown in Figure 1. User requirements may include the user's security and real-time requirements for the entity alignment task. These user requirements can be categorized according to a first standard.

[0057] Next, in step S320, the scheme determination unit 214 can determine an entity alignment scheme for performing the entity alignment task based on the user requirements of the terminal device and the network status between the base station and the terminal device. The network status between the base station and the terminal device can correspond to the network status of the network 130 connecting the base station-side device 110 and multiple terminal devices 120a, 120b shown in FIG. 1. In some embodiments, the network status may include bandwidth level, latency level, and packet loss rate level obtained by classifying each of bandwidth, latency, and packet loss rate of the network 130 according to a second standard. As described above, the processing circuit 210 can be configured to dynamically adjust the frequency of acquiring the network status according to the real-time changes in the network status, so as to reduce unnecessary communication overhead and optimize the use of network resources.

[0058] In some embodiments, to determine an entity alignment scheme for performing an entity alignment task, the scheme determination unit 214 may select an entity alignment scheme that matches the user requirements and network conditions from a plurality of pre-determined entity alignment schemes. For example, the plurality of pre-determined entity alignment schemes may include, but are not limited to: key exchange-based encryption schemes, RSA (Rivest-Shamir-Adleman) encryption schemes combined with blind signatures (B-RSA), overt transmission encryption schemes (OT), multi-party secure computation encryption schemes (MPC), homomorphic encryption schemes (HE), etc. Note that in the context of this disclosure, the term "entity alignment scheme" primarily refers to a scheme for encrypting entity alignment data, such as an encryption scheme used in entity alignment using privacy set intersection (PSI).

[0059] For example, processing circuit 210 can be configured to, for each of a plurality of entity alignment schemes, classify the entity alignment scheme into levels based on its characteristics and a first criterion for each of security and real-time performance, to obtain the security level and real-time performance level of the entity alignment scheme. Scheme determination unit 214 can select from the plurality of entity alignment schemes an entity alignment scheme whose security level and real-time performance level respectively meet the user's security requirements and real-time requirements (e.g., those of terminal devices 120a and 120b). Additionally or alternatively, processing circuit 210 can be configured to, for each of the plurality of entity alignment schemes, obtain, based on the characteristics of the entity alignment scheme and a second criterion, classify the bandwidth requirements, latency requirements, and packet loss rate requirements of the entity alignment scheme for the network, according to the classification. Scheme determination unit 214 can select from the plurality of entity alignment schemes an entity alignment scheme whose bandwidth requirements, latency requirements, and packet loss rate requirements are respectively met by the bandwidth level, latency level, and packet loss rate level of the network (e.g., 130 in FIG. 1). The determination of the entity alignment scheme will be described in more detail later with reference to FIGS. 6 to 8, etc.

[0060] Then, in step S330, the scheme indication unit 216 can indicate the determined entity alignment scheme to the terminal device. For example, the scheme indication unit 216 can issue a code for the determined entity alignment scheme via network 130. Afterward, the processing circuit 210 can obtain encrypted data (e.g., aligned dataset) from the terminal device as the result of the entity alignment task performed by the terminal device according to the determined entity alignment scheme. The processing circuit 210 can then perform subsequent vertical federated learning with the terminal device based on this encrypted data.

[0061] This disclosure comprehensively considers the user's requirements for the entity alignment task and the network's state to appropriately select an entity alignment scheme. This optimizes the entity alignment process and allows for flexible adjustment of the entity alignment strategy according to specific scenarios to ensure that various needs are met. This flexibility makes the system applicable to various complex and changing application environments, thereby improving the applicability of vertical federated learning in practical applications.

[0062] Furthermore, by selecting the optimal encryption mode under different network conditions, this disclosure effectively reduces computational and communication overhead during transmission, improving bandwidth utilization and communication efficiency. For example, in bandwidth-constrained scenarios, a lightweight encryption scheme can be selected, thereby reducing data load. In latency-sensitive scenarios, an appropriate scheme is selected to optimize data transmission paths and encryption computation, thereby achieving rapid response. Additionally, the system according to this disclosure can maintain stable alignment performance when network conditions fluctuate (such as changes in bandwidth, latency, and packet loss rate), reducing the impact of network condition changes on system performance and thus improving the system's adaptability in complex environments.

[0063] Next, with reference to Figures 4 and 5, the electronic device and method for entity alignment in vertical federated learning on the terminal device side according to the present disclosure will be further described.

[0064] Figure 4 illustrates an exemplary configuration block diagram of an electronic device 400 on the terminal device side for entity alignment in vertical federated learning according to an embodiment of the present disclosure. The electronic device 400 may, for example, be used to implement the plurality of terminal devices 120a, 120b shown in Figure 1.

[0065] In some embodiments, the electronic device 400 may include processing circuitry 410. The processing circuitry 410 of the electronic device 400 provides various functions of the electronic device 400.

[0066] Processing circuitry 410 can refer to various implementations of digital circuitry systems, analog circuitry systems, or mixed-signal (analog and digital combination) circuitry systems that perform functions in a computing system. Processing circuitry can include, for example, circuitry such as integrated circuits (ICs), application-specific integrated circuits (ASICs), portions or circuitry of a single processor core, an entire processor core, a single processor, programmable hardware devices such as field-programmable gate arrays (FPGAs), and / or systems comprising multiple processors.

[0067] In some embodiments, the processing circuit 410 may include a user request sending unit 412 and a scheme receiving unit 414, configured to perform corresponding steps in the entity alignment method 500 for vertical federated learning shown in FIG5, described later.

[0068] In some embodiments, the electronic device 400 may further include a memory (not shown). The memory of the electronic device 400 may store information generated by the processing circuitry 410, as well as programs and data used for the operation of the processing circuitry 410. The memory may be volatile memory and / or non-volatile memory. For example, the memory may include, but is not limited to, random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), read-only memory (ROM), and flash memory. Furthermore, the electronic device 400 may be implemented at the chip level, or it may be implemented at the device level by including other external components.

[0069] Figure 5 illustrates an exemplary flowchart of a method 500 for entity alignment in vertical federated learning according to an embodiment of the present disclosure. This method 500 can, for example, be used in an electronic device 400 as shown in Figure 4.

[0070] As shown in Figure 5, in step S510, the user request sending unit 412 can send user requests for the entity alignment task based on vertical federated learning to the base station. For example, the processing circuit 410 can be configured (e.g., via an interface such as a graphical user interface) to obtain the user's security and real-time requirements for the entity alignment task, and the user request sending unit 412 can (e.g., via network 130) send these user requests to the base station. These user requests can be classified according to a first criterion.

[0071] Next, in step S520, the scheme receiving unit 414 may receive an entity alignment scheme from the base station. As described above, the entity alignment scheme may be determined by the base station based on user requirements and the network status between the base station and the terminal device. The received entity alignment scheme may be one or more of the following: a key exchange-based encryption scheme, a blind signature-based RSA encryption scheme (B-RSA), an overtly transmitted encryption scheme (OT), a multi-party secure computation encryption scheme (MPC), and a homomorphic encryption scheme (HE).

[0072] As described above, in some embodiments, the received entity alignment scheme may be an entity alignment scheme selected by the base station from a predetermined plurality of entity alignment schemes. Based on the characteristics of the entity alignment scheme and the security level and real-time level determined by a first standard, the security level and real-time level of the entity alignment scheme respectively meet the user's security requirements and real-time requirements. Additionally or alternatively, in some embodiments, the network state may include bandwidth level, latency level, and packet loss rate level obtained by classifying each of the bandwidth, latency, and packet loss rate of the network between the base station and the terminal device according to a second standard. In such embodiments, the received entity alignment scheme is an entity alignment scheme selected by the base station from a predetermined plurality of entity alignment schemes. Based on the characteristics of the entity alignment scheme and the second standard, the bandwidth requirements, latency requirements, and packet loss rate requirements of the entity alignment scheme are respectively met by the bandwidth level, latency level, and packet loss rate level of the network between the base station and the terminal device.

[0073] Upon receiving an entity alignment scheme from the base station, processing circuitry 410 can be configured to perform entity alignment tasks with other terminal devices according to the received scheme. The execution of the entity alignment task will be described in more detail later with reference to Figures 9A and 9B. Processing circuitry 410 can then send encrypted data (e.g., aligned dataset) as a result of the entity alignment task to the base station for subsequent vertical federated learning. For example, with the assistance or coordination of the base station, processing circuitry 410 can begin training a local model using the aligned dataset and upload the locally trained model to the base station.

[0074] In the above embodiments, the entity alignment scheme to be executed comprehensively considers the user's requirements for the entity alignment task and the network status. This allows for flexible adjustment of the entity alignment strategy according to specific scenarios to ensure that various needs are met.

[0075] Entity alignment scheme determination processing

[0076] Next, the specific process and examples of the entity alignment scheme determination process performed by the base station-side device 110 (or electronic device 200 or its processing circuit 210) will be described with reference to Figures 6 to 8. This can roughly correspond to steps S310 and S320 in method 300 of Figure 3.

[0077] Figure 6 illustrates a schematic diagram of an example of a user request according to an embodiment of the present disclosure and an example of a first criterion for classifying the user request. As described above, the base station-side device 110 can (e.g., via network 130) obtain user requests from each of a plurality of terminal devices 120a, 120b for an entity alignment task. As shown in Figure 6, the user requests may include, but are not limited to, user requirements for the security of the entity alignment task and requirements for the real-time performance of the entity alignment task. The security requirements and real-time performance requirements can be classified according to a first criterion. The classification can be quantitative (i.e., expressed as a score) or non-quantitative.

[0078] For example, in the example in Figure 6, according to the first standard, security requirements are divided into five quantified levels. A security requirement level of "very low," or quantified level "1," indicates that the encryption scheme for the entity alignment task is vulnerable to attack. A security requirement level of "low," or quantified level "2," indicates that the encryption scheme for the entity alignment task provides basic protection but is still somewhat vulnerable to attack. A security requirement level of "medium," or quantified level "3," indicates that the encryption scheme for the entity alignment task provides general protection but may have security vulnerabilities. A security requirement level of "high," or quantified level "4," indicates that the encryption scheme for the entity alignment task provides effective protection, making most attacks difficult to succeed. A security requirement level of "very high," or quantified level "5," indicates that the encryption scheme for the entity alignment task is highly secure, making it virtually impossible to attack.

[0079] Furthermore, in the example of Figure 6, real-time requirements are also categorized into five quantified levels according to the first criterion. A real-time requirement level of "Very Poor," or quantified level "1," indicates that the encryption scheme for the entity alignment task has a long response time, making the task difficult to complete. A real-time requirement level of "Poor," or quantified level "2," indicates that the encryption scheme for the entity alignment task has a long response time, resulting in a poor user experience. A real-time requirement level of "Average," or quantified level "3," indicates that the encryption scheme for the entity alignment task can complete the task, but the response time is unsatisfactory. A real-time requirement level of "Good," or quantified level "4," indicates that the encryption scheme for the entity alignment task can complete the task quickly, resulting in relatively high user satisfaction. A real-time requirement level of "Excellent," or quantified level "5," indicates that the encryption scheme for the entity alignment task is efficient and fast, and can meet all user needs. Note that "requirement" here can, for example, refer to the user's expectation that the corresponding indicators (security, real-time performance, etc.) for the entity alignment to be performed are at least at or better than the required level.

[0080] Note that in Figure 6, for ease of explanation and simplicity, the security and real-time requirements included in the user requirements are divided into five quantified levels according to the first standard, but those skilled in the art should understand that this disclosure is not limited thereto. For example, security and real-time performance can be divided into non-quantified levels that do not indicate specific values, or one can be divided into a quantified level and the other into a non-quantified level. Furthermore, the number of levels is not limited to five; for example, it can be divided into three levels: "high," "medium," and "low," or into more than five levels. The number of levels for each of the security and real-time performance requirements can also be different; for example, security can be divided into three levels, while real-time performance can be divided into five levels.

[0081] Furthermore, those skilled in the art should understand that an appropriate first criterion (e.g., a threshold or benchmark for classifying levels) can be selected based on the needs of the participating party (i.e., the terminal device) and / or the coordinating party (i.e., the base station). This first criterion may be known beforehand between the terminal device and the base station. The classification of user requests can be performed on the terminal device side or on the base station side. For example, terminal devices 120a and 120b may prompt the user to select a pre-classified user request level, or convert the user request into a level after the user inputs it. Alternatively, the terminal device may directly send unclassified user requests to base station-side device 110, where base station-side device 110 converts them into levels. It should also be understood that specific examples of user requests are not limited to security (requirements) and real-time performance (requirements). For example, user requests may also include requirements for Quality of Service (QoL), etc.

[0082] Figure 7 illustrates an example of a network state according to an embodiment of the present disclosure and an example of a second criterion for classifying the network state. As described above, the base station-side device 110 can acquire (e.g., monitor) the network state of the network 130 between the base station and multiple terminal devices 120a, 120b. As shown in Figure 7, the network state can include bandwidth, latency, and packet loss rate of the network (e.g., 130 in Figure 1). The base station-side device 110 can classify each of the network's bandwidth, latency, and packet loss rate according to the second criterion to obtain a specific bandwidth level, latency level, and packet loss rate level as the network state. These levels can be quantified (i.e., expressed as fractions) or non-quantified.

[0083] For example, in the example in Figure 7, according to the second standard, bandwidth (BW) is divided into 5 quantization levels. A bandwidth level of "Very Low" or quantization level "1" indicates a network bandwidth of less than 1 Mbps. A bandwidth level of "Low" or quantization level "2" indicates a network bandwidth between 1 Mbps and 5 Mbps. A bandwidth level of "Medium" or quantization level "3" indicates a network bandwidth between 5 Mbps and 20 Mbps. A bandwidth level of "High" or quantization level "4" indicates a network bandwidth between 20 Mbps and 100 Mbps. A bandwidth level of "Very High" or quantization level "5" indicates a network bandwidth greater than 100 Mbps.

[0084] In the example in Figure 7, latency is also divided into five quantized levels according to the second criterion. A latency level of "Very High" or quantization level "1" indicates a network latency greater than 200 ms. A latency level of "High" or quantization level "2" indicates a network latency between 100 ms and 200 ms. A latency level of "Medium" or quantization level "3" indicates a network latency between 50 ms and 100 ms. A latency level of "Low" or quantization level "4" indicates a network latency between 20 ms and 50 ms. A latency level of "Very Low" or quantization level "5" indicates a network latency less than 20 ms. Note that for latency, the literal high or low of the non-quantized latency levels is the opposite of the magnitude of the quantized score levels.

[0085] Furthermore, in the example in Figure 7, the Packet Loss Rate (PLR) is also divided into five quantized levels according to the second criterion. A latency level of "Very High" or quantization level "1" indicates a packet loss rate greater than 10%. A latency level of "High" or quantization level "2" indicates a packet loss rate between 5% and 10%. A latency level of "Medium" or quantization level "3" indicates a packet loss rate between 1% and 5%. A latency level of "Low" or quantization level "4" indicates a packet loss rate less than 1%. A latency level of "Very Low" or quantization level "5" indicates almost no packet loss, for example, a packet loss rate less than a very small threshold (e.g., 0.01%). Note that for packet loss rate, the literal high and low of the unquantized packet loss rate levels are the opposite of the magnitude of the quantized score levels.

[0086] Note that in Figure 7, for ease of explanation and simplicity, the bandwidth, latency, and packet loss rate included in the network state are divided into five quantized levels according to the second standard, but those skilled in the art should understand that this disclosure is not limited thereto. For example, bandwidth, latency, and packet loss rate can be divided into non-quantized levels that do not indicate specific values, or some of them can be divided into quantized levels while others are divided into non-quantized levels. Furthermore, the number of levels is not limited to five; for example, it can be divided into three levels: "high," "medium," and "low," or it can be divided into more than five levels. The number of levels for each of the bandwidth, latency, and packet loss rate can also be different.

[0087] Furthermore, those skilled in the art should understand that an appropriate second criterion (e.g., a threshold or benchmark for classifying levels) can be selected based on the needs of the participating parties (i.e., terminal devices) and / or the coordinating party (i.e., base stations). This second criterion may be known beforehand between the terminal devices and the base stations. It should also be understood that specific examples of network status are not limited to bandwidth (level), latency (level), and packet loss rate (level). For example, network status may also include network jitter, throughput, etc. (levels).

[0088] After obtaining the user's requirements for the entity alignment task and the network status between the base station and the terminal device, as described above, the base station-side device 110 can determine an entity alignment scheme for performing the entity alignment task based on the user's requirements and the network status. In some embodiments, determining the entity alignment scheme may include selecting an entity alignment scheme that matches the user's requirements and the network status from a plurality of predetermined entity alignment schemes (i.e., alternative entity alignment schemes).

[0089] Figure 8 illustrates the characteristics and levels of an example entity alignment encryption scheme according to embodiments of the present disclosure. As shown in Figure 8, the predetermined plurality of entity alignment schemes may include, but are not limited to: key exchange-based encryption schemes, RSA encryption schemes combined with blind signatures (B-RSA), overt transmission encryption schemes (OT), multi-party secure computation encryption schemes (MPC), and homomorphic encryption schemes (HE). Each encryption scheme may have different characteristics, such as different security and real-time performance metrics, and different requirements for bandwidth, latency, and packet loss rate. Note that "requirements" here may, for example, refer to the encryption scheme requiring the corresponding network state metrics to reach at least a required level for normal operation.

[0090] Key-exchange-based encryption schemes can achieve secure key exchange based on the discrete logarithm problem, where each participant can use the other's public key for encryption to ensure the privacy of the data's hash value during transmission. Key-exchange-based encryption schemes can achieve dynamic key generation, thus adapting to changing network environments. Typical key-exchange-based encryption schemes can include those based on Diffie-Hellman (DH) key exchange. Therefore, in this disclosure, the scheme designation for key-exchange-based encryption schemes is written as "DH". However, it should be understood that other key-exchange-based encryption schemes can also be chosen, such as those based on PSK key exchange, SM2 key exchange, etc. Key-exchange-based encryption schemes can guarantee high security, but have poor real-time performance and high network bandwidth requirements, while having low requirements for latency and packet loss rate. Here, "low" or "not high" requirements for latency or packet loss rate can mean that the scheme can still function normally even with high latency or packet loss rate. Conversely, requiring "high" or "not low" latency or packet loss rate means that the solution can only function normally when the latency or packet loss rate is low, and so on.

[0091] Blind signature-based RSA encryption (B-RSA) provides asymmetric encryption using public and private keys based on the large number factorization problem. It can be combined with blind signature techniques (e.g., blinding and signing) to enhance security and guarantee privacy. B-RSA offers very high security but has moderate real-time performance, high network bandwidth requirements, moderate latency requirements, and moderate packet loss requirements.

[0092] Unintentional transmission (OT) encryption schemes allow one party to securely receive information without knowing the other party's choices. Unintentional transmission emphasizes correctness (the receiver knows the information it selects), receiver privacy (the sender cannot know what information the receiver has selected), and sender privacy (the receiver cannot know information other than its selected information). OT encryption schemes offer extremely high security, good real-time performance, moderate network bandwidth requirements, low latency requirements, and low packet loss requirements.

[0093] Multi-party secure computation (MPC) encryption schemes allow multiple parties to collaboratively compute results without revealing their inputs. This enhances both data privacy and computational efficiency. While MPC offers high security, its real-time performance is generally average, and it has moderate requirements for network bandwidth, latency, and packet loss rate.

[0094] Homomorphic encryption (HE) schemes allow computations to be performed on encrypted data, and the results remain encrypted. It enables computation while ensuring data privacy. HE schemes offer high security but suffer from poor real-time performance, high network bandwidth requirements, high latency requirements, and moderate packet loss rate requirements.

[0095] It should be noted that the predetermined multiple entity alignment schemes (alternative entity alignment schemes) are not limited to the schemes described above. Alternative entity alignment schemes may also include other entity alignment encryption schemes that appeared before or after the filing date of this application, whether or not they are described in this disclosure. By way of example and not limitation, alternative entity alignment schemes may also include encryption schemes based on differential privacy.

[0096] Next, we will use the user requirements and first standard in Figure 6, the network status and second standard in Figure 7, and the entity alignment encryption scheme in Figure 8 as examples to illustrate the specific mode of selecting an entity alignment scheme from a number of predetermined entity alignment schemes.

[0097] In the first mode, the base station side device 110 can classify each entity alignment scheme among a plurality of predetermined entity alignment schemes (alternate entity alignment schemes) based on the characteristics of the entity alignment scheme and a first standard for each of security and real-time performance, so as to obtain the security level and real-time performance level of the entity alignment scheme. Specifically, based on the characteristics of the entity alignment encryption scheme shown in Figure 8 and the first standard in Figure 6, the base station-side device 110 can rate the key exchange-based encryption scheme (scheme code: DH) as security level "4" and real-time level "2"; the RSA encryption scheme combined with blind signature (scheme code: B-RSA) as security level "5" and real-time level "3"; the encryption scheme based on unintentional transmission (scheme code: OT) as security level "5" and real-time level "4"; the multi-party secure computation encryption scheme (scheme code: MPC) as security level "4" and real-time level "3"; and the homomorphic encryption scheme (scheme code: HE) as security level "4" and real-time level "2".

[0098] Then, the base station device 110 can select an entity alignment scheme from the candidate entity alignment schemes that satisfies the user's security requirements and real-time requirements (already expressed in levels according to the first standard), respectively. As an example and not a limitation, assuming the user's security requirements are classified as level "5" and the real-time requirements as level "4", then among the multiple candidate entity alignment encryption schemes, only the encryption scheme based on unintentional transmission (OT) satisfies (in this scenario, "satisfies" means that the scheme's security level and real-time level are not less than the user's security requirements and real-time requirements, respectively) the user's requirements. Therefore, the base station device 110 can select the encryption scheme based on unintentional transmission (OT) as the determined entity alignment scheme.

[0099] In some embodiments, when multiple entity alignment schemes that meet the user's requirements exist among the alternative entity alignment schemes, only one scheme may be selected. For example, one entity alignment scheme may be randomly selected, or the entity alignment scheme among these alternative entity alignment schemes whose security level and real-time performance level are closest to or least close to the user's security and real-time requirements (e.g., based on Euclidean distance) may be selected, or an entity alignment scheme may be selected in any appropriate manner as needed. In alternative embodiments, when multiple entity alignment schemes that meet the user's requirements exist, a hybrid entity alignment scheme based on two or more of the multiple entity alignment schemes that meet the user's requirements may also be selected, thereby combining the advantages of each scheme.

[0100] In some embodiments, when no entity alignment scheme among the candidate entity alignment schemes meets the user's requirements, the terminal device (the user) can be notified that no matching entity alignment scheme exists, and the user can be prompted to re-enter their requirements. For example, an error message can be sent to indicate that the entity alignment task cannot continue under the current user requirements. In an alternative embodiment, when no entity alignment scheme meets the user's requirements, the entity alignment scheme closest to meeting the user's requirements can be selected from the candidate entity alignment schemes (e.g., the entity alignment scheme with the smallest Euclidean distance between the security level and real-time level of the scheme and the user's security and real-time requirements). As an example and not a limitation, suppose the user's security requirements are classified as level "5" and the real-time requirements are classified as level "5". In this case, no entity alignment scheme among the entity alignment schemes shown in Figure 8 meets the user's requirements. In this case, the OT that is closest to meeting the user's requirements can be selected as the entity alignment scheme to be returned to the participants. In this case, an appropriate threshold can be set to ensure that the selected scheme does not deviate too far from the user's requirements.

[0101] An entity alignment scheme that satisfies the user requirements of all terminal devices 120a and 120b can be determined as described above. When no entity alignment scheme satisfies the user requirements of all terminal devices, an error can be returned or the entity alignment scheme that most closely satisfies the user requirements of all terminal devices can be selected, as described above. As an example, and not a limitation, suppose that the security requirements of terminal device 120a are classified as level "5" and the real-time requirements as level "4," while the security requirements of terminal device 120b are classified as level "5" and the real-time requirements as level "5." In this case, no entity alignment scheme simultaneously satisfies the user requirements of all terminal devices. An error can be returned, or the entity alignment scheme that most closely satisfies the user requirements of all terminal devices can still be selected.

[0102] Additionally or alternatively, in the second mode, the base station-side device 110 can, for each entity alignment scheme among the alternative entity alignment schemes, obtain, based on the characteristics of that entity alignment scheme and a second standard, a tiered classification of the network bandwidth requirements, latency requirements, and packet loss rate requirements for that entity alignment scheme. Specifically, based on the characteristics of the entity alignment encryption scheme shown in Figure 8 and the second standard in Figure 7, the base station-side device 110 can classify the bandwidth requirements, latency requirements, and packet loss rate requirements of the key exchange-based encryption scheme (scheme code: DH) into levels "4", "2", and "2", respectively; classify the bandwidth requirements, latency requirements, and packet loss rate requirements of the RSA encryption scheme combined with blind signatures (scheme code: B-RSA) into levels "4", "3", and "3", respectively; and classify the requirements based on the characteristics of the entity alignment encryption scheme shown in Figure 8 and the second standard in Figure 7, respectively. The bandwidth requirements, latency requirements, and packet loss rate requirements of the encryption scheme for intentional transmission (scheme code: OT) are divided into levels "3", "2", and "2", respectively; the bandwidth requirements, latency requirements, and packet loss rate requirements of the multi-party secure computation encryption scheme (scheme code: MPC) are divided into levels "3", "3", and "3", respectively; and the bandwidth requirements, latency requirements, and packet loss rate requirements of the homomorphic encryption scheme (scheme code: HE) are divided into levels "4", "4", and "3", respectively.

[0103] Then, the base station device 110 can select from the candidate entity alignment schemes an entity alignment scheme in which the bandwidth requirement, latency requirement, and packet loss rate requirement are satisfied by the network's bandwidth level, latency level, and packet loss rate level, respectively. As an example, and not a limitation, assuming the current network (e.g., 130 in Figure 1) has a bandwidth level of "3", a latency level of "3", and a packet loss rate level of "2", then among the multiple candidate entity alignment encryption schemes, only the encryption scheme based on unintentional transmission (OT) satisfies the network state in terms of bandwidth requirement, latency requirement, and packet loss rate requirement (in this scenario, "satisfies" means that the scheme's bandwidth requirement, latency requirement, and packet loss rate requirement are no greater than the network's bandwidth level, latency level, and packet loss rate level, respectively). Therefore, the base station device 110 can select the encryption scheme based on unintentional transmission (OT) as the determined matching entity alignment scheme.

[0104] In some embodiments, when there are multiple entity alignment schemes among the candidate entity alignment schemes that match the network state, as in the first mode, only one scheme can be selected, or a hybrid entity alignment scheme based on two or more of the multiple matching entity alignment schemes can be selected to combine the advantages of each scheme.

[0105] In some embodiments, when no entity alignment scheme matching the network state is found among the candidate entity alignment schemes, the terminal device (user) can be notified that no matching entity alignment scheme exists. For example, an error message can be sent to indicate that the entity alignment task cannot continue under the current network state. In an alternative embodiment, when no entity alignment scheme matching the network state is found, the entity alignment scheme closest to matching the network state can be selected from the candidate entity alignment schemes (e.g., the entity alignment scheme with the smallest Euclidean distance between its bandwidth requirement, latency requirement, and packet loss rate requirement and the network's bandwidth level, latency level, and packet loss rate level). By way of example and not limitation, suppose the network bandwidth is level "2", latency is level "2", and packet loss rate is level "2". In this case, no entity alignment scheme matching the network state is found among the entity alignment schemes shown in Figure 8. In this case, the OT closest to matching the network state can be selected as the entity alignment scheme to be returned to the participants. In this case, an appropriate threshold can be set to ensure that the selected scheme does not deviate too far from the network state.

[0106] As described above, in the embodiment where the base station-side device 110 communicates with each terminal device through a separate network, the base station-side device 110 can obtain the network status of the network between itself and each terminal device. In this case, the base station-side device 110 can select an entity alignment scheme that matches or is closest to matching the network status between the base station-side device 110 and all terminal devices.

[0107] Both of the above modes can be considered simultaneously to select an appropriate entity alignment scheme. In other words, an entity alignment scheme can be selected from the alternative entity alignment schemes that satisfies the user's security and real-time requirements respectively, and whose bandwidth, latency, and packet loss rate requirements are satisfied by the network's bandwidth, latency, and packet loss rate levels respectively. As an example, and not a limitation, assuming the user's security requirements are classified as level "5", the real-time requirements as level "3", the network's bandwidth level as "4", latency level as "4", and packet loss rate level as "4", then among the multiple alternative entity alignment encryption schemes, the combination of blind signature RSA encryption (B-RSA) and overt oblivious transmission encryption (OT) simultaneously matches both the user requirements and the network state. In this case, as described above, the base station device 110 can select one of the schemes (e.g., B-RSA) or a hybrid scheme combining the two schemes as the determined entity alignment scheme, as needed.

[0108] After determining (selecting) an entity alignment scheme, the base station device 110 can indicate the determined entity alignment scheme to multiple terminal devices 120a and 120b. For example, the base station device 110 can send the code of the determined entity alignment scheme to the multiple terminal devices 120a and 120b.

[0109] By employing the entity alignment scheme determination process described above, the optimal entity alignment strategy can be flexibly selected based on different user needs and network conditions. This flexibility enables the system according to this disclosure to be applicable to various complex and dynamic application scenarios, thereby enhancing the applicability of vertical federated learning in practical applications.

[0110] Furthermore, by selecting the optimal encryption mode under different network conditions, this disclosure effectively reduces computational and communication overhead during transmission, improving bandwidth utilization and communication efficiency. For example, in bandwidth-constrained scenarios, a lightweight encryption scheme can be selected, thereby reducing data load. In latency-sensitive scenarios, an appropriate scheme is selected to optimize data transmission paths and encryption computation, thereby achieving rapid response. Additionally, the system according to this disclosure can maintain stable alignment performance when network conditions fluctuate (such as changes in bandwidth, latency, and packet loss rate), reducing the impact of network condition changes on system performance and thus improving the system's adaptability in complex environments.

[0111] Next, the entity alignment process performed by the terminal device as a participant after receiving the entity alignment scheme from the base station side device will be described with reference to FIG9A.

[0112] Figure 9A is an example flowchart illustrating an entity alignment process 900 according to an embodiment of the present disclosure. This process 900 can be used, for example, with a plurality of terminal devices 120a, 120b as shown in Figure 1, or with an electronic device 400 or its processing circuitry 410 as shown in Figure 4. For example, process 900 can occur after step S520 of Figure 5.

[0113] As shown in Figure 9A, in step S910, terminal devices 120a and 120b can initialize the entity alignment task based on (for example, the entity alignment scheme received in step S520 of Figure 5). For example, terminal device 120a can generate its public-private key pair (pk A ,sk A The terminal device 120b can generate its key pair (pk). B ,sk B When the received entity alignment scheme is a homomorphic encryption scheme, the encryption parameter params can be selected in this initialization step.

[0114] In step S920, terminal devices 120a and 120b can prepare data for the entity alignment task based on the entity alignment scheme. For example, when the received entity alignment scheme is a key-exchange encryption scheme, terminal device 120a can calculate... The terminal device 120b can calculate... Where H is a hash function, A is the dataset of terminal device 120a, and a i B is an item in the dataset of terminal device 120a, while B is a data item in the dataset of terminal device 120b. j This refers to an item in the dataset of terminal device 120b. When the received entity alignment scheme is a blind signature-based RSA encryption scheme (B-RSA), terminal device 120b can blind its dataset and generate a blinded dataset B′={b′j}. When the received entity alignment scheme is based on an unintentional transmission encryption scheme (OT), terminal device 120a can prepare a data copy {D}. k When the received entity alignment scheme is a multi-party secure computation (MPC) encryption scheme, terminal device 120a can convert its dataset into MPC format, represented as MPC(A), and terminal device 120b can also convert its dataset into MPC format, represented as MPC(B). When the received entity alignment scheme is a homomorphic encryption scheme (HE), terminal device 120a can encrypt its dataset to generate an encrypted dataset. Terminal device 120b can also encrypt its dataset to generate an encrypted dataset.

[0115] In step S930, information transmission can be performed between the terminal devices 120a and 120b. FIG9B shows a schematic diagram of information interaction between the terminal devices during entity alignment according to an embodiment of the present disclosure. The information interaction shown in FIG9B can occur in step S930, and the two terminal devices can correspond to the terminal devices 120a and 120b in FIG1. ​​The information interaction depicted in FIG9B can be performed, for example, via the network 130 in FIG1.

[0116] Step S930 is described with reference to FIG9B. In step S930, each participating party may send the data prepared in step S920 to the other party. For example, when the entity alignment encryption scheme is a key exchange-based encryption scheme, terminal device 120a may send its public key pk A It is sent to terminal device 120b, and terminal device 120b can then use its public key to PK. BThe entity alignment encryption scheme is an RSA encryption scheme combined with blind signatures, which is sent to terminal device 120a. When the entity alignment encryption scheme is an RSA encryption scheme combined with blind signatures, terminal device 120b can send its blinded dataset B′ to terminal device 120a. When the entity alignment encryption scheme is an encryption scheme based on unintentional transmission, terminal device 120a can send its data copy {D} to terminal device 120a. k The entity alignment encryption scheme is a multi-party computational encryption scheme. Terminal device 120a can send its MPC-formatted dataset MPC(A) to terminal device 120b, while terminal device 120b can send its MPC-formatted dataset MPC(B) to terminal device 120a. When the entity alignment encryption scheme is a homomorphic encryption scheme, terminal device 120a can send its encrypted dataset C... A It is sent to terminal device 120b, and terminal device 120b can send its encrypted dataset C. B Send to terminal device 120a.

[0117] In step S940, data encryption or blinding can be performed based on an entity alignment scheme. For example, when the entity alignment encryption scheme is a key exchange-based encryption scheme, terminal device 120a can utilize the public key pk received from terminal device 120b. B For H A Encrypt to generate And send it to terminal device 120b, which can then use the public key pk received from terminal device 120a. A For H B Encrypt to generate The received dataset B′ is then sent to terminal device 120a. When the entity alignment encryption scheme is an RSA encryption scheme combined with blind signatures, terminal device 120a can sign the received dataset B′ to generate σ and send it to terminal device 120b. When the entity alignment encryption scheme is based on an inadvertent transmission encryption scheme, terminal devices 120a and 120b can execute the OT protocol, and terminal device 120b can selectively receive data D of interest. k When the entity alignment encryption scheme is a multi-party secure computation encryption scheme, terminal devices 120a and 120b can execute the MPC protocol and compute the intersection "MPC(A), MPC(B)" of MPC(A) and MPC(B). When the entity alignment encryption scheme is a homomorphic encryption scheme, terminal devices 120a and 120b can perform operations on dataset C. A and C B Perform an intersection operation to generate an Intersection(C) A C BNote that, as mentioned above, the information interaction depicted in Figure 9B can also occur in step S940, for example, when the entity alignment encryption scheme is a key exchange-based encryption scheme or an RSA encryption scheme combined with blind signatures, etc.

[0118] In step S950, data decryption, deblinding, or intersection can be performed. For example, when the entity alignment encryption scheme is a key-exchange based encryption scheme, terminal device 120a can decrypt... And perform intersection, while terminal device 120b can decrypt. And perform intersection. When the entity alignment encryption scheme is an RSA encryption scheme combined with blind signatures, terminal device 120b can deblind σ and verify the signature, and then find the intersection. When the entity alignment encryption scheme is an encryption scheme based on unintentional transmission, terminal device 120b can receive the data of interest D. k The intersection is compared with its own dataset to find the intersection. When the entity alignment encryption scheme is a multi-party secure computation encryption scheme, Intersection "MPC(A), MPC(B)# is the intersection to be output. When the entity alignment encryption scheme is a homomorphic encryption scheme, terminal device 120a and terminal device 120b can compare Intersection (C) with their own dataset to find the intersection. A C B Decryption is performed.

[0119] In step S960, data verification and output can be performed. For example, terminal devices 120a and 120b can verify the correctness of the intersection obtained in step S950. Then, terminal devices 120a and 120b can use the intersection result to perform subsequent federated learning tasks, and can output the intersection result as encrypted data to base station device 110 for subsequent federated learning tasks.

[0120] Note that, for simplicity, only the processing of two terminal devices has been described above, and Figure 9B also shows only two terminal devices. However, this disclosure is not limited to this, and the process 900 of Figure 9A can be applied to more than two terminal devices, and information exchange can also occur between more than two terminal devices. For example, when there are three or more terminal devices, the process 900 shown in Figure 9A can be executed between any two of these terminal devices. Alternatively, multi-party schemes such as multi-party secure computation encryption schemes can be performed directly between three or more terminal devices. Furthermore, it should be noted that the roles of terminal devices 120a and 120b in the above-described process 900 can be interchanged without departing from the scope of this disclosure.

[0121] Through the process described above (900), the entity alignment workflow can be optimized, and a standardized process can be provided for entity alignment tasks. This can solve the compatibility problems that may arise when traditional entity alignment schemes face different data distributions or inconsistent encryption schemes among different participants, thereby avoiding impacts on alignment efficiency and accuracy.

[0122] Next, the signaling interaction process between the base station side equipment and the terminal equipment in the environment shown in Figure 1 will be described.

[0123] Figure 10 illustrates a sequence diagram of signaling interactions between a terminal device and a base station-side device for entity alignment in vertical federated learning according to an embodiment of the present disclosure. Figure 10 is described in conjunction with the contents of Figures 1 through 9B. The terminal device in Figure 10 may correspond to the plurality of terminal devices 120a, 120b in Figure 1. The base station-side device in Figure 10 may correspond to the base station-side device 110 in Figure 1.

[0124] As shown in Figure 10, in step S1001, the terminal device can obtain its user's requirements for the entity alignment task based on vertical federated learning. For example, terminal devices 120a and 120b can obtain their user's requirements for the entity alignment task, such as security requirements and real-time requirements, through an interface such as a graphical user interface.

[0125] In step S1002, the terminal device can upload the user request to the base station-side device. For example, as described in conjunction with FIG1, terminal devices 120a and 120b can send the user request to the base station-side device 110 via network 130.

[0126] In step S1003, the base station-side device can determine an entity alignment scheme. For example, the base station-side device 110 can determine the entity alignment scheme based on the received user request and the obtained network status between the base station and the terminal device (i.e., the network status of network 130 in FIG. 1, including bandwidth, latency, packet loss rate, etc.). For example, the determination of the entity alignment scheme is described in more detail in FIG. 2, FIG. 3, and FIG. 6 to FIG. 8.

[0127] In step S1004, the base station-side device can issue a determined entity alignment scheme. For example, the base station-side device 110 can send the entity alignment scheme (e.g., its scheme code) determined in step S1003 to the terminal devices 120a and 120b via the network 130.

[0128] In step S1005, the terminal device can perform entity alignment based on the received entity alignment scheme. For example, each of the plurality of terminal devices 120a, 120b can perform entity alignment tasks with other terminal devices according to the received entity alignment scheme. For example, the execution of entity alignment is described in more detail in Figures 9A and 9B.

[0129] In step S1006, the terminal device can upload encrypted data, which is the result of entity alignment, to the base station-side device. For example, terminal devices 120a and 120b can send the intersection result obtained in process 900 of FIG9A as encrypted data to the base station-side device 110.

[0130] In step S1007, subsequent federated learning can be performed. For example, the base station device 110 can build a basic model and send its structure and parameters to the terminal devices 120a and 120b, while the terminal devices 120a and 120b can train the model locally using the entity-aligned dataset and upload the locally trained model to the base station, and so on.

[0131] The signaling interaction described above optimizes the entity alignment process and provides a standardized workflow for entity alignment tasks. This resolves compatibility issues that may arise in traditional entity alignment schemes, thereby avoiding impacts on alignment efficiency and accuracy.

[0132] Figure 11 illustrates an exemplary configuration that enables the implementation of a computing device 1100 according to an embodiment of the present invention.

[0133] Computing device 1100 is an example of a hardware device capable of applying the above aspects of the present invention. For example, computing device 1100 can implement electronic device 200 of FIG. 2 or electronic device 400 of FIG. 4. Computing device 1100 can be any machine configured to perform processing and / or calculations. Computing device 1100 can be, but is not limited to, a workstation, server, desktop computer, laptop computer, tablet computer, personal data assistant (PDA), wireless terminal, drone, portable smart device, vehicle terminal, Internet of Things device, or a combination thereof.

[0134] As shown in Figure 11, computing device 1100 may include one or more components that can be connected to or communicate with bus 1102 via one or more interfaces. Bus 1102 may include, but is not limited to, Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MCA) bus, Enhanced ISA (EISA) bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus. Computing device 1100 may include, for example, one or more processors 1104, one or more input devices 1106, and one or more output devices 1108. The one or more processors 1104 may be any type of processor and may include, but is not limited to, one or more general-purpose processors or dedicated processors (such as dedicated processing chips). Processor 1102 may, for example, correspond to processing circuit 210 in Figure 2 or processing circuit 410 in Figure 4, and is configured to implement the functions of various units of the electronic device on the base station side or terminal device side of the entity alignment for vertical federated learning disclosed herein. Input device 1106 may be any type of input device capable of inputting information to the computing device and may include, but is not limited to, a mouse, keyboard, touch screen, microphone, and / or remote controller. The output device 1108 can be any type of device capable of presenting information, and may include, but is not limited to, a monitor, speaker, video / audio output terminal, vibrator and / or printer.

[0135] The computing device 1100 may also include or be connected to a non-transitory storage device 1114, which may be any non-transitory storage device capable of storing data, and may include, but is not limited to, disk drives, optical storage devices, solid-state storage, floppy disks, flexible disks, hard disks, magnetic tapes or any other magnetic media, compressed disks or any other optical media, cache memory and / or any other storage chip or module, and / or any other medium from which a computer may read data, instructions and / or code. The computing device 1100 may also include random access memory (RAM) 1110 and read-only memory (ROM) 1112. ROM 1112 may store executable programs, utilities, or processes in a non-volatile manner. RAM 1110 provides volatile data storage and stores instructions related to the operation of the computing device 1100. The computing device 1100 may also include a network / bus interface 1116 coupled to a data link 1118. The network / bus interface 1116 can be any kind of device or system capable of enabling communication with external devices and / or networks, and may include, but is not limited to, modems, network cards, infrared communication devices, wireless communication devices and / or chipsets (such as Bluetooth™ devices, IEEE 802.11 devices, WiFi devices, WiMax devices, mobile cellular communication facilities, etc.).

[0136] The following will introduce application examples based on this disclosure.

[0137] The technology disclosed herein can be applied to a variety of products.

[0138] For example, a base station can be implemented as any type of evolved Node B (eNB) or gNode B (gNB) in next-generation radio access technologies, such as macro eNB / gNB and small eNB / gNB. Small eNB / gNBs can be eNB / gNBs covering cells smaller than macro cells, such as pico eNB / gNBs, micro eNB / gNBs, and femtocell eNB / gNBs. Alternatively, a base station can be implemented as any other type of base station, such as one or both of a Base Transceiver Station (BTS) and Base Station Controller (BSC) in a GSM system, one or both of a Radio Network Controller (RNC) and Node B in a WCDMA system, or a corresponding network node in a future communication system. A base station may include: a subject configured to control wireless communication (also called base station equipment); and one or more Remote Radio Headers (RRHs) located in a different location from the subject. Furthermore, the various types of terminals described below can operate as base stations by temporarily or semi-persistently performing base station functions.

[0139] For example, a terminal device (UE) can be implemented as a mobile terminal (such as a smartphone, tablet PC, laptop PC, portable gaming terminal, portable / dongle-type mobile router, and digital camera device) or an in-vehicle terminal (such as a car navigation device). A terminal device can also be implemented as a terminal performing machine-to-machine (M2M) communication (also known as a machine-type communication (MTC) terminal). Furthermore, a terminal device can be a wireless communication module (such as an integrated circuit module comprising a single chip) installed on each of the aforementioned terminals.

[0140] [Application examples of base stations]

[0141] (First application example)

[0142] Figure 12 is a block diagram illustrating a first example of a schematic configuration of a gNB to which the technologies of this disclosure can be applied. The gNB 1200 includes one or more antennas 1210 and a base station device 1220. The base station device 1220 and each antenna 1210 can be connected to each other via RF cables.

[0143] Each of the antennas 1210 includes one or more antenna elements (such as multiple antenna elements included in a multiple-input multiple-output (MIMO) antenna) and is used by the base station device 1220 to transmit and receive wireless signals. As shown in FIG12, the gNB 1200 may include multiple antennas 1210. For example, the multiple antennas 1210 may be compatible with multiple frequency bands used by the gNB 1200. The base station device 1220 includes a controller 1221, a memory 1222, a network interface 1223, and a wireless communication interface 1225.

[0144] The controller 1221 can be, for example, a CPU or a DSP, and operates various higher-level functions of the base station equipment 1220. For example, the controller 1221 generates data packets based on data in signals processed by the wireless communication interface 1225, and transmits the generated packets via the network interface 1223. The controller 1221 can bundle data from multiple baseband processors to generate bundled packets and transmit the generated bundled packets. The controller 1221 may have logical functions that perform controls such as radio resource control, radio bearer control, mobility management, admission control, and scheduling. This control can be performed in conjunction with nearby gNBs, eNBs, or core network nodes (e.g., Access and Mobility Management Functions (AMF)). The memory 1222 includes RAM and ROM, and stores programs executed by the controller 1221 and various types of control data (such as terminal lists, transmission power data, and scheduling data).

[0145] Network interface 1223 is a communication interface for connecting base station equipment 1220 to core network 1224. Controller 1221 can communicate with core network nodes or other gNBs / eNBs via network interface 1223. In this case, gNB 1200 and core network nodes or other gNBs / eNBs can be connected to each other via logical interfaces (such as the N2 interface with AMF and the Xn interface with the gNB). Network interface 1223 can also be a wired communication interface or a wireless communication interface for wireless backhaul. If network interface 1223 is a wireless communication interface, it can use a higher frequency band for wireless communication compared to the frequency band used by wireless communication interface 1225.

[0146] Wireless communication interface 1225 supports any cellular communication scheme (such as LTE, LTE-Advanced, NR (New Radio)) and provides wireless connectivity to terminals located in the cell of gNB 1200 via antenna 1210. Wireless communication interface 1225 typically includes, for example, a baseband (BB) processor 1226 and RF circuitry 1227. BB processor 1226 can perform, for example, encoding / decoding, modulation / demodulation, and multiplexing / demultiplexing, and performs various types of signal processing at layers (e.g., L1, Media Access Control (MAC), Radio Link Control (RLC), and Packet Data Convergence Protocol (PDCP)). Instead of controller 1221, BB processor 1226 may have some or all of the above-described logical functions. BB processor 1226 may be a memory storing communication control programs, or a module including a processor and associated circuitry configured to execute programs. Update programs can change the functionality of BB processor 1226. The module may be a card or blade inserted into a slot in base station equipment 1220. Alternatively, the module may be a chip mounted on a card or blade. Meanwhile, the RF circuit 1227 may include, for example, a mixer, a filter, and an amplifier, and transmits and receives wireless signals via the antenna 1210.

[0147] As shown in Figure 12, the wireless communication interface 1225 may include multiple BB processors 1226. For example, the multiple BB processors 1226 may be compatible with multiple frequency bands used by the gNB 1200. As shown in Figure 12, the wireless communication interface 1225 may include multiple RF circuits 1227. For example, the multiple RF circuits 1227 may be compatible with multiple antenna elements. Although Figure 12 shows an example in which the wireless communication interface 1225 includes multiple BB processors 1226 and multiple RF circuits 1227, the wireless communication interface 1225 may also include a single BB processor 1226 or a single RF circuit 1227.

[0148] (Second application example)

[0149] Figure 13 is a block diagram illustrating a second example of a schematic configuration of a gNB to which the technologies of this disclosure can be applied. The gNB 1330 includes one or more antennas 1340, a base station device 1350, and an RRH 1360. The RRH 1360 and each antenna 1340 can be connected to each other via RF cables. The base station device 1350 and the RRH 1360 can be connected to each other via high-speed lines such as fiber optic cables.

[0150] Each of the antennas 1340 includes one or more antenna elements (such as multiple antenna elements included in a MIMO antenna) and is used by the RRH 1360 to transmit and receive wireless signals. As shown in FIG13, the gNB 1330 may include multiple antennas 1340. For example, the multiple antennas 1340 may be compatible with multiple frequency bands used by the gNB 1330. The base station device 1350 includes a controller 1351, a memory 1352, a network interface 1353, a wireless communication interface 1355, and a connection interface 1357. The controller 1351, memory 1352, and network interface 1353 are the same as the controller 1221, memory 1222, and network interface 1223 described with reference to FIG12.

[0151] Wireless communication interface 1355 supports any cellular communication scheme (such as LTE and LTE-Advanced) and provides wireless communication to terminals located in the sector corresponding to RRH 1360 via RRH 1360 and antenna 1340. Wireless communication interface 1355 may typically include, for example, a BB processor 1356. The BB processor 1356 is identical to the BB processor 1226 described with reference to FIG12, except that it is connected to the RF circuitry 1364 of RRH 1360 via connection interface 1357. As shown in FIG13, wireless communication interface 1355 may include multiple BB processors 1356. For example, multiple BB processors 1356 may be compatible with multiple frequency bands used by gNB 1330. Although FIG13 shows an example in which wireless communication interface 1355 includes multiple BB processors 1356, wireless communication interface 1355 may also include a single BB processor 1356.

[0152] Connection interface 1357 is an interface for connecting base station device 1350 (wireless communication interface 1355) to RRH 1360. Connection interface 1357 may also be a communication module for communication in the aforementioned high-speed line connecting base station device 1350 (wireless communication interface 1355) to RRH 1360.

[0153] The RRH 1360 includes a connectivity interface 1361 and a wireless communication interface 1363.

[0154] Connection interface 1361 is an interface for connecting RRH 1360 (wireless communication interface 1363) to base station equipment 1350. Connection interface 1361 can also be a communication module for communication in the aforementioned high-speed line.

[0155] Wireless communication interface 1363 transmits and receives wireless signals via antenna 1340. Wireless communication interface 1363 typically includes, for example, RF circuitry 1364. RF circuitry 1364 may include, for example, mixers, filters, and amplifiers, and transmits and receives wireless signals via antenna 1340. As shown in FIG13, wireless communication interface 1363 may include multiple RF circuits 1364. For example, multiple RF circuits 1364 may support multiple antenna elements. Although FIG13 shows an example in which wireless communication interface 1363 includes multiple RF circuits 1364, wireless communication interface 1363 may also include a single RF circuit 1364.

[0156] In the gNB 1200 and gNB 1330 shown in Figures 12 and 13, one or more components included in the processing circuit 210 described with reference to Figure 2 may be implemented in the wireless communication interface 1225 or the wireless communication interface 1355. Alternatively, at least some of these components may also be implemented by the controllers 1221 and 1351.

[0157] [Application examples of terminal devices]

[0158] (First application example)

[0159] Figure 14 is a block diagram illustrating an example of a schematic configuration of a smartphone 1400 to which the technologies of this disclosure can be applied. The smartphone 1400 includes a processor 1401, a memory 1402, a storage device 1403, an external connection interface 1404, a camera device 1406, a sensor 1407, a microphone 1408, an input device 1409, a display device 1410, a speaker 1411, a wireless communication interface 1412, one or more antenna switches 1415, one or more antennas 1416, a bus 1417, a battery 1418, and an auxiliary controller 1419.

[0160] Processor 1401 may be, for example, a CPU or a system-on-a-chip (SoC), and controls the application layer and other functions of smartphone 1400. Memory 1402 includes RAM and ROM, and stores data and programs executed by processor 1401. Storage device 1403 may include storage media such as semiconductor memory and hard disk. External connectivity interface 1404 is an interface for connecting external devices, such as memory cards and Universal Serial Bus (USB) devices, to smartphone 1400.

[0161] The camera device 1406 includes an image sensor (such as a charge-coupled device (CCD) and complementary metal-oxide-semiconductor (CMOS)) and generates captured images. The sensor 1407 may include a set of sensors, such as a measurement sensor, a gyroscope sensor, a magnetometer sensor, and an accelerometer sensor. The microphone 1408 converts sound input to the smartphone 1400 into an audio signal. The input device 1409 includes, for example, a touch sensor, keypad, keyboard, buttons, or switches configured to detect touches on the screen of the display device 1410 and receives operations or information input from the user. The display device 1410 includes a screen (such as a liquid crystal display (LCD) and an organic light-emitting diode (OLED) display) and displays the output image of the smartphone 1400. The speaker 1411 converts the audio signal output from the smartphone 1400 into sound.

[0162] The wireless communication interface 1412 supports any cellular communication scheme (such as LTE and LTE-Advanced) and performs wireless communication. The wireless communication interface 1412 typically includes, for example, a BB processor 1413 and RF circuitry 1414. The BB processor 1413 can perform, for example, encoding / decoding, modulation / demodulation, and multiplexing / demultiplexing, and performs various types of signal processing for wireless communication. Meanwhile, the RF circuitry 1414 can include, for example, mixers, filters, and amplifiers, and transmits and receives wireless signals via antenna 1416. The wireless communication interface 1412 can be a single chip module on which the BB processor 1413 and RF circuitry 1414 are integrated. As shown in Figure 14, the wireless communication interface 1412 can include multiple BB processors 1413 and multiple RF circuits 1414. Although Figure 14 shows an example where the wireless communication interface 1412 includes multiple BB processors 1413 and multiple RF circuits 1414, the wireless communication interface 1412 can also include a single BB processor 1413 or a single RF circuitry 1414.

[0163] In addition to cellular communication schemes, wireless communication interface 1412 can support other types of wireless communication schemes, such as short-range wireless communication schemes, near-field communication schemes, and wireless local area network (LAN) schemes. In this case, wireless communication interface 1412 may include a BB processor 1413 and RF circuitry 1414 for each wireless communication scheme.

[0164] Each of the antenna switches 1415 switches the connection destination of the antenna 1416 among multiple circuits (e.g., circuits for different wireless communication schemes) included in the wireless communication interface 1412.

[0165] Each of the antennas 1416 includes one or more antenna elements (such as multiple antenna elements included in a MIMO antenna) and is used for transmitting and receiving wireless signals through the wireless communication interface 1412. As shown in Figure 14, the smartphone 1400 may include multiple antennas 1416. Although Figure 14 shows an example in which the smartphone 1400 includes multiple antennas 1416, the smartphone 1400 may also include a single antenna 1416.

[0166] Furthermore, the smartphone 1400 may include an antenna 1416 for each wireless communication scheme. In this case, the antenna switch 1415 can be omitted from the configuration of the smartphone 1400.

[0167] Bus 1417 connects the processor 1401, memory 1402, storage device 1403, external connection interface 1404, camera device 1406, sensor 1407, microphone 1408, input device 1409, display device 1410, speaker 1411, wireless communication interface 1412, and auxiliary controller 1419 to each other. Battery 1418 supplies power to the various blocks of smartphone 1400 shown in FIG. 14 via feeders, which are partially shown as dashed lines in the figure. Auxiliary controller 1419 operates the minimum necessary functions of smartphone 1400, for example, in sleep mode.

[0168] In the smartphone 1400 shown in FIG14, one or more components included in the processing circuitry 410 described with reference to FIG4 may be implemented in the wireless communication interface 1412. Alternatively, at least some of these components may also be implemented by the processor 1401 or the auxiliary controller 1419.

[0169] (Second application example)

[0170] Figure 15 is a block diagram illustrating an example of a schematic configuration of a car navigation device 1520 to which the technology of this disclosure can be applied. The car navigation device 1520 includes a processor 1521, a memory 1522, a Global Positioning System (GPS) module 1524, a sensor 1525, a data interface 1526, a content player 1527, a storage medium interface 1528, an input device 1529, a display device 1530, a speaker 1531, a wireless communication interface 1533, one or more antenna switches 1536, one or more antennas 1537, and a battery 1538.

[0171] The processor 1521 can be, for example, a CPU or a SoC, and controls the navigation functions and other functions of the car navigation device 1520. The memory 1522 includes RAM and ROM, and stores data and programs executed by the processor 1521.

[0172] GPS module 1524 uses GPS signals received from GPS satellites to measure the location (such as latitude, longitude, and altitude) of car navigation device 1520. Sensor 1525 may include a set of sensors, such as a gyroscope sensor, a geomagnetic sensor, and an air pressure sensor. Data interface 1526 is connected to, for example, an in-vehicle network 1541 via a terminal not shown, and acquires data generated by the vehicle (such as vehicle speed data).

[0173] Content player 1527 reproduces content stored on storage media (such as CDs and DVDs), which is inserted into storage media interface 1528. Input device 1529 includes, for example, a touch sensor, button, or switch configured to detect touch on the screen of display device 1530, and receives operations or information input from the user. Display device 1530 includes a screen such as an LCD or OLED display and displays images or reproduced content for navigation functions. Speaker 1531 outputs sound for navigation functions or reproduced content.

[0174] The wireless communication interface 1533 supports any cellular communication scheme (such as LTE and LTE-Advanced) and performs wireless communication. The wireless communication interface 1533 typically includes, for example, a BB processor 1534 and RF circuitry 1535. The BB processor 1534 can perform, for example, encoding / decoding, modulation / demodulation, and multiplexing / demultiplexing, and performs various types of signal processing for wireless communication. Meanwhile, the RF circuitry 1535 can include, for example, mixers, filters, and amplifiers, and transmits and receives wireless signals via antenna 1537. The wireless communication interface 1533 can also be a chip module on which the BB processor 1534 and RF circuitry 1535 are integrated. As shown in Figure 15, the wireless communication interface 1533 can include multiple BB processors 1534 and multiple RF circuits 1535. Although Figure 15 shows an example where the wireless communication interface 1533 includes multiple BB processors 1534 and multiple RF circuits 1535, the wireless communication interface 1533 can also include a single BB processor 1534 or a single RF circuitry 1535.

[0175] In addition to cellular communication schemes, the wireless communication interface 1533 can support other types of wireless communication schemes, such as short-range wireless communication schemes, near-field communication schemes, and wireless LAN schemes. In this case, for each wireless communication scheme, the wireless communication interface 1533 may include a BB processor 1534 and an RF circuit 1535.

[0176] Each of the antenna switches 1536 switches the connection destination of the antenna 1537 among multiple circuits (such as circuits for different wireless communication schemes) included in the wireless communication interface 1533.

[0177] Each of the antennas 1537 includes one or more antenna elements (such as multiple antenna elements included in a MIMO antenna) and is used for transmitting and receiving wireless signals through the wireless communication interface 1533. As shown in Figure 15, the car navigation device 1520 may include multiple antennas 1537. Although Figure 15 shows an example in which the car navigation device 1520 includes multiple antennas 1537, the car navigation device 1520 may also include a single antenna 1537.

[0178] Furthermore, the car navigation device 1520 may include an antenna 1537 for each wireless communication scheme. In this case, the antenna switch 1536 can be omitted from the configuration of the car navigation device 1520.

[0179] Battery 1538 supplies power to the various blocks of the car navigation device 1520 shown in Figure 15 via feeders, which are partially shown as dashed lines in the figure. Battery 1538 accumulates the power supplied from the vehicle.

[0180] In the car navigation device 1520 shown in FIG15, one or more components included in the processing circuit 410 described with reference to FIG4 may be implemented in the wireless communication interface 1512. Alternatively, at least some of these components may also be implemented by the processor 1521.

[0181] The technology disclosed herein can also be implemented as an in-vehicle system (or vehicle) 1540 including one or more blocks of a car navigation device 1520, an in-vehicle network 1541, and a vehicle module 1542. The vehicle module 1542 generates vehicle data (such as vehicle speed, engine speed, and fault information) and outputs the generated data to the in-vehicle network 1541.

[0182] It should be understood that the reference to "embodiment" or similar expressions in this specification means that a specific feature, structure, or characteristic described in connection with that embodiment is included in at least one specific embodiment of this disclosure. Therefore, the appearance of the terms "in embodiments of this disclosure" and similar expressions in this specification does not necessarily refer to the same embodiment.

[0183] Those skilled in the art will understand that this disclosure can be implemented as a system, apparatus, method, or as a computer-readable storage medium (e.g., a non-transient storage medium) as a computer program product. Therefore, this disclosure can be implemented in various forms, such as a completely hardware embodiment, a completely software embodiment (including firmware, resident software, microprogram code, etc.), or a software and hardware embodiment, hereinafter referred to as a "circuit," "module," or "system." Furthermore, this disclosure can also be implemented as a computer program product in any tangible media form, having computer-usable program code stored thereon.

[0184] The description herein is based on flowcharts and / or block diagrams of systems, apparatuses, methods, and computer program products according to specific embodiments of this disclosure. It will be understood that each block in each flowchart and / or block diagram, and any combination of blocks in the flowcharts and / or block diagrams, can be implemented using computer program instructions. These computer program instructions are executable by a machine comprising a processor of a general-purpose computer or a special-purpose computer, or other programmable data processing means, and are processed by the computer or other programmable data processing means to perform the functions or operations described in the flowcharts and / or block diagrams.

[0185] The accompanying drawings illustrate flowcharts and block diagrams showing the architecture, functionality, and operation of systems, apparatuses, methods, and computer program products achievable according to various embodiments of the present disclosure. Thus, each block in a flowchart or block diagram may represent a module, segment, or portion of program code, including one or more executable instructions to implement a specified logical function. It should also be noted that in some other embodiments, the functions described in a block may not be performed in the order shown in the figures. For example, two blocks illustrated as connected may actually be executed simultaneously, or in some cases, depending on the functions involved, they may be executed in the reverse order shown in the figures. Furthermore, it should be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, may be implemented by a system based on dedicated hardware, or by a combination of dedicated hardware and computer instructions, to perform specific functions or operations.

[0186] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical applications, or technical improvements to market technology of the embodiments, or to enable others skilled in the art to understand the embodiments disclosed herein.

[0187] Note that the technology disclosed in this specification may have the following configurations.

[0188] (1) An electronic device for use on a base station side, the electronic device comprising:

[0189] Processing circuit, the processing circuit being configured to:

[0190] Obtain user requirements from the terminal device for the entity alignment task based on vertical federated learning;

[0191] Based on the user requirements of the terminal device and the network status between the base station and the terminal device, an entity alignment scheme is determined for performing the entity alignment task; and

[0192] Instruct the terminal device on the determined entity alignment scheme.

[0193] (2) The electronic device according to (1), wherein determining the entity alignment scheme includes:

[0194] Select an entity alignment scheme that matches the user requirements and the network status from a plurality of pre-determined entity alignment schemes.

[0195] (3) The electronic device according to (2),

[0196] The user requirements include the user's security requirements and real-time requirements for the entity alignment task, and the security requirements and real-time requirements are classified into levels according to a first standard.

[0197] Selecting an entity alignment scheme from the plurality of entity alignment schemes includes:

[0198] For each of the plurality of entity alignment schemes, based on the characteristics of the entity alignment scheme and a first criterion, the entity alignment scheme is classified into a level for each of security and real-time performance, thereby obtaining the security level and real-time performance level of the entity alignment scheme; and

[0199] Select an entity alignment scheme from the plurality of entity alignment schemes that satisfies the security requirements and the real-time requirements, respectively.

[0200] (4) The electronic device according to (2) or (3),

[0201] The network status includes bandwidth level, latency level, and packet loss rate level obtained by classifying each of the bandwidth, latency, and packet loss rate of the network between the base station and the terminal device according to the second standard.

[0202] Selecting an entity alignment scheme from the plurality of entity alignment schemes includes:

[0203] For each of the plurality of entity alignment schemes, based on the characteristics of the entity alignment scheme and a second criterion, the bandwidth requirements, latency requirements, and packet loss rate requirements for the network for that entity alignment scheme are obtained in a graded manner; and

[0204] Select an entity alignment scheme from the plurality of entity alignment schemes in which the bandwidth requirement, latency requirement, and packet loss rate requirement are satisfied by the bandwidth level, latency level, and packet loss rate level, respectively.

[0205] (5) The electronic device according to (2), wherein the plurality of entity alignment schemes include: a key exchange-based encryption scheme, a blind signature-based RSA encryption scheme (B-RSA), an overt transmission-based encryption scheme (OT), a multi-party secure computation encryption scheme (MPC), and a homomorphic encryption scheme (HE).

[0206] (6) The electronic device according to (1), wherein the processing circuit is configured to:

[0207] Obtain encrypted data from the terminal device as the result of the entity alignment task performed by the terminal device according to the determined entity alignment scheme; and

[0208] Vertical federated learning is performed based on the encrypted data.

[0209] (7) The electronic device according to (1), wherein the processing circuit is configured to:

[0210] The frequency of obtaining network status is dynamically adjusted based on real-time changes in network status.

[0211] (8) An electronic device for use on a terminal device side, the electronic device comprising:

[0212] Processing circuit, the processing circuit being configured to:

[0213] Send user requests for entity alignment tasks based on vertical federated learning to the base station; and

[0214] The system receives an entity alignment scheme from the base station, which is determined by the base station based on the user's request and the network status between the base station and the terminal device.

[0215] (9) The electronic device according to (8), wherein the processing circuit is further configured to:

[0216] The user's security and real-time requirements for the entity alignment task are obtained as the user requirements, wherein the security and real-time requirements are classified into levels according to a first standard;

[0217] The entity alignment scheme is selected from a plurality of predetermined entity alignment schemes. Based on the characteristics of the entity alignment scheme and the first standard, the security level and real-time level of the entity alignment scheme meet the security requirements and real-time requirements, respectively.

[0218] (10) The electronic device according to (8) or (9),

[0219] The network status includes bandwidth level, latency level, and packet loss rate level obtained by classifying each of the bandwidth, latency, and packet loss rate of the network between the base station and the terminal device according to a second standard.

[0220] The entity alignment scheme is selected from a plurality of predetermined entity alignment schemes. Based on the characteristics of the entity alignment scheme and the second standard, the bandwidth requirements, latency requirements, and packet loss rate requirements of the entity alignment scheme are respectively satisfied by the bandwidth level, latency level, and packet loss rate level of the network between the base station and the terminal device.

[0221] (11) The electronic device according to (8), wherein the entity alignment scheme is one or more of the following:

[0222] Key exchange-based encryption schemes, B-RSA encryption schemes combined with blind signatures, OT encryption schemes based on unintentional transmission, MPC encryption schemes, and homomorphic encryption schemes (HE).

[0223] (12) The electronic device according to (8), wherein the processing circuit is further configured to:

[0224] Perform the entity alignment task with other terminal devices according to the described entity alignment scheme; and

[0225] Encrypted data, as a result of the entity alignment task, is sent to the base station for vertical federated learning.

[0226] (13) A method for entity alignment in vertical federated learning, comprising:

[0227] Obtain user requirements from the terminal device for the entity alignment task based on vertical federated learning;

[0228] Based on the user requirements of the terminal device and the network status between the base station and the terminal device, an entity alignment scheme is determined for performing the entity alignment task; and

[0229] Instruct the terminal device on the determined entity alignment scheme.

[0230] (14) According to the method of (13), wherein determining the entity alignment scheme includes:

[0231] Select an entity alignment scheme that matches the user requirements and the network status from a plurality of pre-determined entity alignment schemes.

[0232] (15) According to the method described in (14),

[0233] The user requirements include the user's security requirements and real-time requirements for the entity alignment task, and the security requirements and real-time requirements are classified into levels according to a first standard.

[0234] Selecting an entity alignment scheme from the plurality of entity alignment schemes includes:

[0235] For each of the plurality of entity alignment schemes, based on the characteristics of the entity alignment scheme and a first criterion, the entity alignment scheme is classified into a level for each of security and real-time performance, thereby obtaining the security level and real-time performance level of the entity alignment scheme; and

[0236] Select an entity alignment scheme from the plurality of entity alignment schemes that satisfies the security requirements and the real-time requirements, respectively.

[0237] (16) According to the method described in (14) or (15),

[0238] The network status includes bandwidth level, latency level, and packet loss rate level obtained by classifying each of the bandwidth, latency, and packet loss rate of the network between the base station and the terminal device according to the second standard.

[0239] Selecting an entity alignment scheme from the plurality of entity alignment schemes includes:

[0240] For each of the plurality of entity alignment schemes, based on the characteristics of the entity alignment scheme and a second criterion, the bandwidth requirements, latency requirements, and packet loss rate requirements for the network for that entity alignment scheme are obtained in a graded manner; and

[0241] Select an entity alignment scheme from the plurality of entity alignment schemes in which the bandwidth requirement, latency requirement, and packet loss rate requirement are satisfied by the bandwidth level, latency level, and packet loss rate level, respectively.

[0242] (17) According to the method of (14), wherein the plurality of entity alignment schemes include: a key exchange-based encryption scheme, a blind signature-based RSA encryption scheme (B-RSA), an accidental transmission-based encryption scheme (OT), a multi-party secure computation encryption scheme (MPC), and a homomorphic encryption scheme (HE).

[0243] (18) The method according to (13) further includes:

[0244] Obtain encrypted data from the terminal device as the result of the entity alignment task performed by the terminal device according to the determined entity alignment scheme; and

[0245] Vertical federated learning is performed based on the encrypted data.

[0246] (19) The method according to (13) further includes:

[0247] The frequency of obtaining network status is dynamically adjusted based on real-time changes in network status.

[0248] (20) A method for entity alignment in vertical federated learning, comprising:

[0249] Send user requests for entity alignment tasks based on vertical federated learning to the base station; and

[0250] The system receives an entity alignment scheme from the base station, which is determined by the base station based on the user's request and the network status between the base station and the terminal device.

[0251] (21) The method according to (20) further includes:

[0252] The user's security and real-time requirements for the entity alignment task are obtained as the user requirements, wherein the security and real-time requirements are classified into levels according to a first standard;

[0253] The entity alignment scheme is selected from a plurality of predetermined entity alignment schemes. Based on the characteristics of the entity alignment scheme and the first standard, the security level and real-time level of the entity alignment scheme meet the security requirements and real-time requirements, respectively.

[0254] (22) According to the method described in (20) or (21),

[0255] The network status includes bandwidth level, latency level, and packet loss rate level obtained by classifying each of the bandwidth, latency, and packet loss rate of the network between the base station and the terminal device according to a second standard.

[0256] The entity alignment scheme is selected from a plurality of predetermined entity alignment schemes. Based on the characteristics of the entity alignment scheme and the second standard, the bandwidth requirements, latency requirements, and packet loss rate requirements of the entity alignment scheme are respectively satisfied by the bandwidth level, latency level, and packet loss rate level of the network between the base station and the terminal device.

[0257] (23) The method according to (20), wherein the entity alignment scheme is one or more of the following:

[0258] Key exchange-based encryption schemes, B-RSA encryption schemes combined with blind signatures, OT encryption schemes based on unintentional transmission, MPC encryption schemes, and homomorphic encryption schemes (HE).

[0259] (24) The method according to (20) further includes:

[0260] Perform the entity alignment task with other terminal devices according to the described entity alignment scheme; and

[0261] Encrypted data, as a result of the entity alignment task, is sent to the base station for vertical federated learning.

[0262] (25) A computer-readable storage medium comprising executable instructions that, when executed by an information processing apparatus, cause the information processing apparatus to perform the method according to any one of (13) to (24).

[0263] (26) A computer program product comprising a computer program that, when executed by a processor, causes the processor to perform the method according to any one of (13) to (24).

Claims

1. An electronic device for use on a base station side, the electronic device comprising: Processing circuit, the processing circuit being configured to: Obtain user requirements from the terminal device for the entity alignment task based on vertical federated learning; Based on the user requirements of the terminal device and the network status between the base station and the terminal device, an entity alignment scheme is determined for performing the entity alignment task; as well as Instruct the terminal device on the determined entity alignment scheme.

2. The electronic device according to claim 1, wherein, Determining the entity alignment scheme includes: Select an entity alignment scheme that matches the user requirements and the network status from a plurality of pre-determined entity alignment schemes.

3. The electronic device according to claim 2, in, The user requirements include the user's security and real-time requirements for the entity alignment task, and the security and real-time requirements are classified into levels according to a first standard. Selecting an entity alignment scheme from the plurality of entity alignment schemes includes: For each of the plurality of entity alignment schemes, based on the characteristics of the entity alignment scheme and a first criterion, the entity alignment scheme is classified into a level for each of security and real-time performance, thereby obtaining the security level and real-time performance level of the entity alignment scheme; and Select an entity alignment scheme from the plurality of entity alignment schemes that satisfies the security requirements and the real-time requirements, respectively.

4. The electronic device according to claim 2 or 3, in, The network status includes bandwidth level, latency level, and packet loss rate level obtained by classifying each of the bandwidth, latency, and packet loss rate of the network between the base station and the terminal device according to the second standard; Selecting an entity alignment scheme from the plurality of entity alignment schemes includes: For each of the plurality of entity alignment schemes, based on the characteristics of the entity alignment scheme and a second criterion, the bandwidth requirements, latency requirements, and packet loss rate requirements for the network for that entity alignment scheme are obtained in a graded manner; and Select an entity alignment scheme from the plurality of entity alignment schemes in which the bandwidth requirement, latency requirement, and packet loss rate requirement are satisfied by the bandwidth level, latency level, and packet loss rate level, respectively.

5. The electronic device according to claim 2, wherein, The multiple entity alignment schemes include: key exchange-based encryption schemes, RSA encryption schemes combined with blind signatures (B-RSA), encryption schemes based on unintentional transmission (OT), multi-party secure computation encryption schemes (MPC), and homomorphic encryption schemes (HE).

6. The electronic device according to claim 1, wherein, The processing circuit is configured as follows: Obtain encrypted data from the terminal device as the result of the entity alignment task performed by the terminal device according to the determined entity alignment scheme; as well as Vertical federated learning is performed based on the encrypted data.

7. The electronic device according to claim 1, wherein, The processing circuit is configured as follows: The frequency of obtaining network status is dynamically adjusted based on real-time changes in network status.

8. An electronic device for use on a terminal device side, the electronic device comprising: Processing circuit, the processing circuit being configured to: Send user requests for entity alignment tasks based on vertical federated learning to the base station; as well as The system receives an entity alignment scheme from the base station, which is determined by the base station based on the user's request and the network status between the base station and the terminal device.

9. The electronic device according to claim 8, wherein, The processing circuit is further configured to: The user's security and real-time requirements for the entity alignment task are obtained as the user requirements, wherein the security and real-time requirements are classified into levels according to a first standard; The entity alignment scheme is selected from a plurality of predetermined entity alignment schemes. Based on the characteristics of the entity alignment scheme and the first standard, the security level and real-time level of the entity alignment scheme meet the security requirements and real-time requirements, respectively.

10. The electronic device according to claim 8 or 9, in, The network status includes bandwidth level, latency level, and packet loss rate level obtained by classifying each of the bandwidth, latency, and packet loss rate of the network between the base station and the terminal device according to a second standard. The entity alignment scheme is selected from a plurality of predetermined entity alignment schemes. Based on the characteristics of the entity alignment scheme and the second standard, the bandwidth requirements, latency requirements, and packet loss rate requirements of the entity alignment scheme are respectively satisfied by the bandwidth level, latency level, and packet loss rate level of the network between the base station and the terminal device.

11. The electronic device according to claim 8, wherein, The entity alignment scheme is one or more of the following: Key exchange-based encryption schemes, B-RSA encryption schemes combined with blind signatures, OT encryption schemes based on unintentional transmission, MPC encryption schemes, and homomorphic encryption schemes (HE).

12. The electronic device according to claim 8, wherein, The processing circuit is further configured to: Perform the entity alignment task with other terminal devices according to the described entity alignment scheme; and Encrypted data, as a result of the entity alignment task, is sent to the base station for vertical federated learning.

13. A method for entity alignment in vertical federated learning, comprising: Obtain user requirements from the terminal device for the entity alignment task based on vertical federated learning; Based on the user requirements of the terminal device and the network status between the base station and the terminal device, an entity alignment scheme is determined for performing the entity alignment task; as well as Instruct the terminal device on the determined entity alignment scheme.

14. The method according to claim 13, wherein, Determining the entity alignment scheme includes: Select an entity alignment scheme that matches the user requirements and the network status from a plurality of pre-determined entity alignment schemes.

15. The method according to claim 14, in, The user requirements include the user's security and real-time requirements for the entity alignment task, and the security and real-time requirements are classified into levels according to a first standard. Selecting an entity alignment scheme from the plurality of entity alignment schemes includes: For each of the plurality of entity alignment schemes, based on the characteristics of the entity alignment scheme and a first criterion, the entity alignment scheme is classified into a level for each of security and real-time performance, thereby obtaining the security level and real-time performance level of the entity alignment scheme; and Select an entity alignment scheme from the plurality of entity alignment schemes that satisfies the security requirements and the real-time requirements, respectively.

16. The method according to claim 14 or 15, in, The network status includes bandwidth level, latency level, and packet loss rate level obtained by classifying each of the bandwidth, latency, and packet loss rate of the network between the base station and the terminal device according to the second standard; Selecting an entity alignment scheme from the plurality of entity alignment schemes includes: For each of the plurality of entity alignment schemes, based on the characteristics of the entity alignment scheme and a second criterion, the bandwidth requirements, latency requirements, and packet loss rate requirements for the network for that entity alignment scheme are obtained in a graded manner; and Select an entity alignment scheme from the plurality of entity alignment schemes in which the bandwidth requirement, latency requirement, and packet loss rate requirement are satisfied by the bandwidth level, latency level, and packet loss rate level, respectively.

17. The method of claim 14, wherein, The multiple entity alignment schemes include: key exchange-based encryption schemes, RSA encryption schemes combined with blind signatures (B-RSA), encryption schemes based on unintentional transmission (OT), multi-party secure computation encryption schemes (MPC), and homomorphic encryption schemes (HE).

18. The method of claim 13, further comprising: Obtain encrypted data from the terminal device as the result of the entity alignment task performed by the terminal device according to the determined entity alignment scheme; as well as Vertical federated learning is performed based on the encrypted data.

19. The method of claim 13, further comprising: The frequency of obtaining network status is dynamically adjusted based on real-time changes in network status.

20. A method for entity alignment in vertical federated learning, comprising: Send user requests for entity alignment tasks based on vertical federated learning to the base station; as well as The system receives an entity alignment scheme from the base station, which is determined by the base station based on the user's request and the network status between the base station and the terminal device.

21. The method of claim 20, further comprising: The user's security and real-time requirements for the entity alignment task are obtained as the user requirements, wherein the security and real-time requirements are classified into levels according to a first standard; The entity alignment scheme is selected from a plurality of predetermined entity alignment schemes. Based on the characteristics of the entity alignment scheme and the first standard, the security level and real-time level of the entity alignment scheme meet the security requirements and real-time requirements, respectively.

22. The method according to claim 20 or 21, in, The network status includes bandwidth level, latency level, and packet loss rate level obtained by classifying each of the bandwidth, latency, and packet loss rate of the network between the base station and the terminal device according to a second standard. The entity alignment scheme is selected from a plurality of predetermined entity alignment schemes. Based on the characteristics of the entity alignment scheme and the second standard, the bandwidth requirements, latency requirements, and packet loss rate requirements of the entity alignment scheme are respectively satisfied by the bandwidth level, latency level, and packet loss rate level of the network between the base station and the terminal device.

23. The method according to claim 20, wherein, The entity alignment scheme is one or more of the following: Key exchange-based encryption schemes, B-RSA encryption schemes combined with blind signatures, OT encryption schemes based on unintentional transmission, MPC encryption schemes, and homomorphic encryption schemes (HE).

24. The method of claim 20, further comprising: Perform the entity alignment task with other terminal devices according to the entity alignment scheme described above; as well as Encrypted data, as a result of the entity alignment task, is sent to the base station for vertical federated learning.

25. A computer-readable storage medium comprising executable instructions that, when executed by an information processing apparatus, cause the information processing apparatus to perform the method according to any one of claims 13 to 24.

26. A computer program product comprising a computer program that, when executed by a processor, causes the processor to perform the method according to any one of claims 13 to 24.

Images