A sample data processing method, apparatus and system

Abstract

This invention discloses a sample data processing method, apparatus, and system. The method includes: acquiring vehicle source data collected by a vehicle-to-everything (V2X) device, wherein the vehicle source data includes vehicle user features extracted using a vehicle feature extractor; acquiring a set of platform source data transmitted by at least one third-party internet device, wherein the platform source data includes platform user features extracted using a platform feature extractor; distributing the vehicle source data and any set of platform source data to target devices, wherein the target devices include at least one of the following: a V2X device and a third-party internet device; and performing data alignment processing on the vehicle source data and at least one set of platform source data to generate training sample data, wherein the training sample data consists of feature data shared by the V2X device and at least one third-party internet platform. This invention solves the problem of low data acquisition efficiency in V2X devices in related technologies.

Description

Technical Field

[0001] This invention relates to the field of big data and data security in the Internet of Vehicles, and more specifically, to a sample data processing method, apparatus and system. Background Technology

[0002] Currently, the integration of vehicle-to-everything (V2X) big data with internet ecosystem data primarily involves OEMs purchasing raw data from internet platforms, then stitching together the purchased data and using the integrated data to build models. Here, "OEM" refers to car companies, as opposed to the underlying parts suppliers; "V2X data" encompasses a series of data within the vehicle's Internet of Things (IoT); and "vehicle internet ecosystem data" refers to the data generated by the ecosystem formed by car manufacturers, service companies, data companies, and repair parts suppliers. Many OEMs' reliance on their own capabilities to utilize raw data from internet platforms is often limited and involves data security risks, resulting in low data acquisition efficiency.

[0003] There is currently no effective solution to the above problems. Summary of the Invention

[0004] This invention provides a sample data processing method, apparatus, and system to at least address the technical problem of low data acquisition efficiency in vehicle networking devices in related technologies.

[0005] According to one aspect of the present invention, a sample data processing method is provided, comprising: acquiring vehicle source data collected by a vehicle-to-everything (V2X) device, wherein the vehicle source data includes vehicle user features extracted using a vehicle feature extractor; acquiring a set of platform source data transmitted by at least one third-party internet device, wherein the platform source data includes platform user features extracted using a platform feature extractor; distributing the vehicle source data and any set of platform source data to a target device, wherein the target device includes at least one of the following: a V2X device and a third-party internet device; performing data alignment processing on the vehicle source data and at least one set of platform source data to generate training sample data, wherein the training sample data consists of feature data shared by the V2X device and at least one third-party internet platform.

[0006] Optionally, after generating training sample data, the above method further includes: the target device using the training sample data to perform feature training on the local initialization model, generating a model evaluation index associated with the target device, wherein the initialization model is a machine learning model; and the target device using the model evaluation index to optimize the local feature extractor.

[0007] Optionally, before acquiring the vehicle source data collected by the vehicle networking device, the above method further includes: collecting metadata of different vehicles; a vehicle feature extractor deployed in the vehicle networking device extracts features from the metadata of different vehicles to obtain at least one set of vehicle user features of different vehicles; and integrating at least one set of vehicle user features of different vehicles to generate vehicle source data.

[0008] Optionally, in the process of generating vehicle source data, the above method further includes: performing differential privacy encryption on vehicle user characteristics, and distributing the encryption result as vehicle source data.

[0009] Optionally, before acquiring a set of platform source data transmitted by at least one third-party Internet device, the above method further includes: collecting metadata of at least one third-party Internet device; a platform feature extractor deployed in the third-party Internet device extracting features from the metadata of the corresponding third-party Internet device to acquire at least one set of platform user features of at least one third-party Internet device; and integrating at least one set of platform user features of at least one third-party Internet device to generate platform source data.

[0010] Optionally, in the process of generating platform source data, the above method further includes: performing differential privacy encryption on platform user characteristics, and distributing the encryption result as platform source data.

[0011] Optionally, vehicle source data and / or any set of platform source data can be distributed to different target devices in an unintentional transmission manner, enabling data exchange between the target devices.

[0012] According to another aspect of the present invention, a sample data processing apparatus is also provided, comprising: a first acquisition module, configured to acquire vehicle source data collected by a vehicle networking device, wherein the vehicle source data includes vehicle user features extracted using a vehicle feature extractor; a second acquisition module, configured to acquire a set of platform source data transmitted by at least one third-party Internet device, wherein the platform source data includes platform user features extracted using a platform feature extractor; a distribution module, configured to distribute the vehicle source data and any set of platform source data to target devices, wherein the target devices include at least one of the following: a vehicle networking device and a third-party Internet device; and a processing module, configured to perform data alignment processing on the vehicle source data and at least one set of platform source data to generate training sample data, wherein the training sample data consists of feature data shared by the vehicle networking device and at least one third-party Internet platform.

[0013] According to another aspect of the present invention, a sample data processing system is also provided, comprising: a vehicle-to-everything (V2X) device for collecting vehicle source data, wherein the vehicle source data includes vehicle user features extracted using a vehicle feature extractor; and at least one third-party internet device for collecting platform source data, wherein the platform source data includes platform user features extracted using a platform feature extractor; wherein the V2X device and the at least one third-party internet device distribute the locally collected vehicle source data and any set of platform source data, thereby enabling data exchange between the V2X device and the at least one third-party internet device, and both the V2X device and the at least one third-party internet device perform data alignment processing on the vehicle source data and at least one set of platform source data to generate training sample data, wherein the training sample data consists of feature data shared by the V2X device and the at least one third-party internet platform.

[0014] According to another aspect of the present invention, a vehicle is also provided, comprising: one or more processors; a storage device for storing one or more programs; and a method for processing sample data such that the one or more processors execute any one of the programs when the one or more programs are executed by the one or more processors.

[0015] In this embodiment of the invention, firstly, vehicle source data collected by the vehicle-to-everything (V2X) device is acquired, wherein the vehicle source data includes vehicle user features extracted using a vehicle feature extractor; secondly, a set of platform source data transmitted by at least one third-party internet device is acquired, wherein the platform source data includes platform user features extracted using a platform feature extractor; the vehicle source data and any set of platform source data are distributed to target devices, wherein the target devices include at least one of the following: the V2X device and the third-party internet device; the vehicle source data and at least one set of platform source data are subjected to data alignment processing to generate training sample data, wherein the training sample data consists of feature data shared by the V2X device and at least one third-party internet platform. It is noteworthy that by combining differential privacy and unintended transmission data encryption technology, the data can be fully utilized while ensuring the privacy of multiple parties, thus solving the problems of low data security and poor reliability in the joint use of V2X data and internet ecosystem data. Attached Figure Description

[0016] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0017] Figure 1 This is a flowchart of a sample data processing method according to an embodiment of the present invention;

[0018] Figure 2 This is a schematic diagram of an optional initialization process according to an embodiment of the present invention;

[0019] Figure 3 This is a schematic diagram of an optional feature extraction process according to an embodiment of the present invention;

[0020] Figure 4 This is a schematic diagram of an optional iterative stage according to an embodiment of the present invention;

[0021] Figure 5 This is a first schematic diagram of a data exchange process according to an embodiment of the present invention;

[0022] Figure 6 This is a second schematic diagram of a data exchange process according to an embodiment of the present invention;

[0023] Figure 7 This is a third schematic diagram of a data exchange process according to an embodiment of the present invention;

[0024] Figure 8 This is a fourth schematic diagram of a data exchange process according to an embodiment of the present invention;

[0025] Figure 9 This is a schematic diagram of a data processing device according to an embodiment of the present invention. Detailed Implementation

[0026] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0027] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0028] Example 1

[0029] According to an embodiment of the present invention, a method embodiment of a multi-party joint model training method and system based on unintentional transmission is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0030] Figure 1 This is a flowchart of a sample data processing method according to an embodiment of the present invention, such as... Figure 1 As shown, the method includes the following steps:

[0031] Step S102: Obtain vehicle source data collected by the vehicle networking device.

[0032] The vehicle source data includes vehicle user features extracted using a vehicle feature extractor.

[0033] The aforementioned vehicle-to-everything (V2X) devices refer to in-vehicle equipment, such as audio systems and navigation systems. The aforementioned vehicle source data refers to data that directly provides raw information. The aforementioned vehicle feature extractor is a commonly used tool in machine learning, pattern recognition, and image processing, used to extract desired feature information from initial data. The aforementioned vehicle user features may include user groups, user preferences, user trajectories, and a range of other user information.

[0034] Vehicle-to-everything (V2X) devices can effectively utilize dynamic vehicle information across an information network platform through wireless communication technology. They can provide various functions during vehicle operation to ensure safer distances between vehicles, reducing the likelihood of collisions. V2X devices can also assist drivers with real-time navigation and improve traffic efficiency through communication with other vehicles and network systems. The vehicle source data collected through V2X devices will play a crucial role in machine learning within this field.

[0035] Step S104: Obtain a set of platform source data transmitted by at least one third-party Internet device, wherein the platform source data includes: platform user features extracted using a platform feature extractor.

[0036] The aforementioned third-party internet devices are internet platform devices other than those used by OEMs (Original Equipment Manufacturers). OEMs include, but are not limited to, automotive companies. The data collected by these third-party internet platforms is called platform source data, and OEMs can purchase this platform source data from these platforms.

[0037] The aforementioned platform feature extractor is a tool used by third-party internet platforms to extract feature information. This extractor can extract platform user information from platform user features.

[0038] In one alternative embodiment, because internet platforms often struggle to provide raw data to big data owners, increasing the difficulty for OEMs to obtain platform source data, a combination of differential privacy and unintentional transmission encryption can be used to acquire platform source data. Differential privacy encryption, a cryptographic method, maximizes the accuracy of data queries while minimizing the chance of identifying query records when retrieving data from statistical databases. Unintentional transmission is a cryptographic protocol where the sender sends a message from a pool of pending messages to the receiver without subsequently knowing which message was sent; hence, it's also called an unintentional transmission protocol. This combination of differential privacy encryption and unintentional transmission enables internet platforms to provide raw data to big data owners.

[0039] Step S106: Distribute vehicle source data and any set of platform source data to target devices, wherein the target devices include at least one of the following: vehicle networking devices and third-party Internet devices.

[0040] The aforementioned target device can be a vehicle-to-everything (V2X) device, or it can be a car model within a third-party internet device that requires machine learning and model training.

[0041] In one optional embodiment, after obtaining the source data from a third-party platform, the platform source data and existing vehicle source data can be sent to the target device so that the target device can process the vehicle source data and platform source data to generate sample data.

[0042] Step S108: Perform data alignment processing on the vehicle source data and at least one set of platform source data to generate training sample data, wherein the training sample data consists of feature data shared by the vehicle networking device and at least one third-party Internet platform.

[0043] In one optional embodiment, vehicle source data and at least one set of platform source data can be obtained. By aligning the vehicle source data and the at least one set of platform source data, training sample data with common characteristics to the vehicle source data can be determined from the at least one set of platform source data. Here, alignment processing is a sample data processing method. By combining the vehicle source data with the screening of the at least one set of platform source data, more training sample data with higher relevance to vehicle-to-everything (V2X) devices can be obtained, thereby improving the accuracy of the training sample data.

[0044] Through the above steps, firstly, vehicle source data collected by the vehicle-to-everything (V2X) device is acquired. This vehicle source data includes vehicle user features extracted using a vehicle feature extractor. Secondly, a set of platform source data transmitted from at least one third-party internet device is acquired. This platform source data includes platform user features extracted using a platform feature extractor. Thirdly, the vehicle source data and any set of platform source data are distributed to target devices, where the target devices include at least one of the following: the V2X device and the third-party internet device. Finally, the vehicle source data and at least one set of platform source data are aligned to generate training sample data. This training sample data contains feature data shared by the V2X device and at least one third-party internet platform, thus improving data acquisition efficiency. It is noteworthy that the training sample data is obtained by filtering at least one set of platform source data based on the vehicle source data; therefore, its accuracy is high. Furthermore, due to the large quantity of platform source data, using vehicle source data and at least one set of platform source data can improve the data acquisition efficiency of the V2X device, thereby solving the technical problem of low data acquisition efficiency in V2X devices in related technologies.

[0045] Optionally, according to the above method, after generating training sample data, the method further includes: the target device using the training sample data to perform feature training on the local initialization model, generating a model evaluation index associated with the target device, wherein the initialization model is a machine learning model; and the target device using the model evaluation index to optimize the local feature extractor.

[0046] The above initialization model is a car model that has been set to the default state. Feature training is a common machine learning method in the field of artificial intelligence.

[0047] The model evaluation metrics mentioned above are the reference for analyzing and evaluating the trained model in the embodiments, and may include accuracy, recall, etc.

[0048] Figure 2 This is a schematic diagram of an initialization process according to an embodiment of the present invention, such as... Figure 2As shown, C is the owner of the vehicle network big data, and A and B are internet platforms. C now needs to rely on A and B for model applications. However, A and B cannot provide C with raw data. A and B possess feature extractors that can extract feature factors without practical meaning from the source data. The feature extractor is a device that extracts required information based on a certain feature, used to extract the required feature information from the initial data; the feature factors without practical meaning are numerical information of features not contained in the source data, and their use will not lead to information leakage. A and B perform differential privacy encryption on the feature factors and then input them into C in an inadvertent transmission manner. C uses the feature factors to perform prediction or classification models without exposing the selected data to A and B.

[0049] Figure 3 This is a schematic diagram of a feature extraction process according to an embodiment of the present invention, such as... Figure 3 As shown, A and B use existing feature extractors or create new feature extractors to produce feature A. i B j After adding noise using differential privacy, the data is input to C via unintentional transmission. Unintentional transmission enables data transfer from A to C and from B to C. During this process, neither A nor B knows which users C has selected, nor the data characteristics of which users C has not selected. C cannot identify user characteristics; it only uses user characteristics shared by A, B, and C to construct an initialization model that relies solely on the characteristics of A and B. This process allows A and B to provide raw data to C, thus solving the difficulty for internet platforms to provide raw data to big data owners.

[0050] Optionally, before acquiring the vehicle source data collected by the vehicle networking device, the above method further includes: collecting metadata of different vehicles; a vehicle feature extractor deployed in the vehicle networking device extracts features from the metadata of different vehicles to obtain at least one set of vehicle user features of different vehicles; and integrating at least one set of vehicle user features of different vehicles to generate vehicle source data.

[0051] The aforementioned metadata, also known as vehicle intermediary data, is data that describes the data, primarily information describing the data attributes.

[0052] The above vehicle source data is data that can directly provide the original vehicle information.

[0053] In one optional embodiment, the vehicle feature extractor in the vehicle networking device can extract features from the metadata of different vehicles to obtain the user features of different vehicles. Then, at least one set of vehicle user features from different vehicles can be integrated to generate vehicle source data. This vehicle source data will be used for corresponding data processing with the platform source data.

[0054] Optionally, in the process of generating vehicle source data, the method further includes: performing differential privacy encryption on vehicle user characteristics, and distributing the encryption result as vehicle source data.

[0055] In one alternative embodiment, because internet platforms often struggle to provide raw data to big data owners, increasing the difficulty for OEMs to obtain platform source data, a combination of differential privacy and inadvertent transmission encryption can be used to acquire platform source data. Differential privacy encryption, a cryptographic method, maximizes the accuracy of data queries while reducing the chance of identifying query records when querying data from statistical databases, achieving inadvertent transmission. Inadvertent transmission is the effect of a cryptographic protocol where the sender sends a message from a pool of pending messages to the receiver, without subsequently knowing which message was sent; this protocol is also called an inadvertent transmission protocol.

[0056] Figure 4 This is a schematic diagram of an optional iterative stage according to an embodiment of the present invention. In one optional embodiment, an iterative stage is adopted. Figure 4 The method described herein, in which differential privacy encryption is a common technique in cryptography, such as... Figure 4 The steps of this method are as follows.

[0057] Among them, the Population Stability Index (PSI) is an indicator set to measure the stability of the target, and the differential privacy encryption coefficient is a coefficient used when using differential privacy for encryption. This coefficient can be set according to the data type to be encrypted.

[0058] In the first step, C periodically feeds the model to A and B. A and B then train feature extractors A and B and differential privacy encryption coefficients based on the model stability index PSI.

[0059] In the second step, C has already obtained n. c Each vehicle-to-everything (V2X) data point contains m data points. k (k = 1, ..., n) c From this vehicle-to-everything (V2X) data, n metadata entries can be extracted using feature extractor C. c User characteristics C k (k = 1, ..., n) c );

[0060] Third step, C will n c Each feature is differentially encrypted for privacy, and then sent to A and B unnoticed.

[0061] In the fourth step, A has already obtained n.a There are m objects, where each object has m... i (i = 1, ..., n) a From the metadata (n records), feature extractor A can extract n... a There are 1 set of user characteristics A, where each object possesses a set of user characteristics A. i (i = 1, ..., n) a );

[0062] Step 5, A will n a Each user's characteristics are encrypted using a differential privacy algorithm and then sent to B unintentionally. The data in B is then aligned with the data in B to generate S. b Each object has a characteristic, where S b <n a S b <n b S b <n c ;

[0063] Step 6, B has already obtained n b There are m objects, where each object has m... j (j = 1, ..., n) j ) data points, through feature extractor B, extract n b There are 3 objects, each possessing a set of user characteristics B. j (j = 1, ..., n) b );

[0064] Step 7, B will n b Each user's characteristics are encrypted using a differential privacy algorithm and then sent to A unintentionally. The data in A is then aligned with the information in A to generate S. a Each object has a characteristic, where S a <n a S a <n b S a <n c ;

[0065] Step 8, A will n a A i (i = 1, ..., n) a User features are associated with user features B and C as model input, and then the feature extractor A is optimized using the evaluation metrics output by the model.

[0066] Step 9, B will n b B i (i = 1, ..., n) bUser features are associated with user features A and C as model input, and then the feature extractor B is optimized using the evaluation metrics output by the model.

[0067] Step 10: Features A extracted by the new stage feature extractor from A and B. i B i The input is fed into C. C then optimizes its own feature extractor and iterates the model, where iteration is an activity of repeated feedback processes aimed at approximating the desired target or result.

[0068] During the iteration process, data A and B also inadvertently transmit their respective encrypted data for model training. During model training, the input that A can submit is... B and C are encrypted data, and A is real data. To make A i transpose and B′ j C′ k The Hadamard product is performed using the transpose of the matrix. The transpose of a matrix is ​​an algorithm in linear algebra that reverses the rows and columns. The Hadamard product is formed by multiplying corresponding elements to create a new matrix. The input submitted by B is... A and C are encrypted data, while B is real data. In each iteration cycle, A, B, and C are adjusted independently. Since none of the three parties can independently possess their metadata, each party only has 1 / 3 of the features that can be adjusted during the feature extraction stage.

[0069] Optionally, before acquiring a set of platform source data transmitted by at least one third-party Internet device, the method further includes: collecting metadata of at least one third-party Internet device; a platform feature extractor deployed in the third-party Internet device extracting features from the metadata of the corresponding third-party Internet device to acquire at least one set of platform user features of at least one third-party Internet device; and integrating at least one set of platform user features of at least one third-party Internet device to generate the platform source data.

[0070] The aforementioned metadata of third-party internet devices can also be called intermediary data of third-party internet devices. It is data that describes data, mainly information describing data attributes.

[0071] The source data mentioned above is data that can directly provide the platform's original information.

[0072] In one alternative embodiment, different metadata features can be extracted using a vehicle feature extractor in a third-party internet device to obtain different platform user features. Then, at least one set of platform user features can be integrated to generate platform source data, which will be subsequently used for corresponding data processing with vehicle source data.

[0073] Optionally, in the process of generating the platform source data, the above method further includes: performing differential privacy encryption on the platform user characteristics, and distributing the encryption result as platform source data.

[0074] Since internet platforms often struggle to provide raw data to big data owners, they can employ a combination of differential privacy and stealth encryption to access the platform's source data. Differential privacy encryption, a cryptographic technique, maximizes the accuracy of data queries while minimizing the chance of identifying the query records when retrieving data from statistical databases. Through differential privacy encryption, internet platforms can then provide the encrypted results to big data owners.

[0075] Optionally, vehicle source data and / or any set of platform source data can be distributed to different target devices in an unintentional transmission manner, enabling data exchange between the target devices.

[0076] In this context, unintentional transmission refers to the effect of a cryptographic protocol in which the sender sends a message to the receiver from a list of pending messages, but subsequently remains unaware of which message was sent. This protocol is also known as the unintentional transmission protocol. In this embodiment of the invention, vehicle source data and / or any set of platform source data can be distributed to different target devices through unintentional transmission, thereby completing data exchange.

[0077] Figures 5-8 The data exchange process from A to C is described in detail.

[0078] Figure 5 This is a first schematic diagram of a data exchange process according to an optional embodiment of the present invention, such as... Figure 5 As shown, A has object 1, object 2, and object 3. Based on the existing data distribution, A uses differential privacy to insert noisy objects. Now, A needs to transmit the feature data of object 1, object 2, object 3, and the noisy objects to C in an unintentional manner.

[0079] Figure 6 This is a second schematic diagram of a data exchange process according to an optional embodiment of the present invention, such as... Figure 6 As shown, the first step is to protect the real names of objects that C does not have. A generates a software structure standard (Universally Unique Identifier, or UUID for short) using the model name and version number (modelName) + object (Object). The UUIDs are: data1, data2, data3, and data4. UUID is a software construction standard. Its purpose is to ensure that all elements in the distributed system have unique identification information without requiring the central control terminal to specify the identification information.

[0080] Figure 7 This is a third schematic diagram of a data exchange process according to an optional embodiment of the present invention, such as... Figure 7 As shown, in the second step, A generates a public key (e, n) and a private key pair using RSA, and sends the UUID data: data1, data2, data3, data4 and the public key to C. The public key algorithm mentioned above is an asymmetric encryption algorithm, including fast public key algorithms and traditional public key algorithms. The private key algorithm is a symmetric encryption algorithm. RSA is the first algorithm that can be used for both encryption and digital signatures. It is easy to understand and operate, and is the most widely studied public key algorithm.

[0081] Figure 8 This is a fourth schematic diagram of a data exchange process according to an optional embodiment of the present invention, as shown below. Figure 8 As shown, in the third step, C obtains the UUID and compares it with existing data to obtain the id of the same UUID, where id is the address.

[0082] In the fourth step, C uses the public key to generate a random number for each sequence number, uses the random number to perform polynomial encryption on the selected address [2, 3], and transmits the two encrypted messages to A.

[0083] Fifth, after receiving the encryption sequence number from C, A combines the previously generated private and public keys to generate an encryption polynomial. The encryption polynomial is then used to encrypt all of A's data, and the encrypted data is transmitted to C.

[0084] In the sixth step, after receiving the encrypted data, C uses the previously generated random number and public key to decrypt it, and can only effectively parse out the data corresponding to [2, 3].

[0085] Example 2

[0086] According to another aspect of the invention, a sample data processing apparatus is also provided. Figure 9 This is a schematic diagram of a data processing apparatus according to an embodiment of the present invention, such as... Figure 9 As shown, the device includes the following components:

[0087] The first acquisition module 902 is used to acquire vehicle source data collected by the vehicle networking device, wherein the vehicle source data includes: vehicle user features extracted by the vehicle feature extractor;

[0088] The second acquisition module 904 is used to acquire a set of platform source data transmitted by at least one third-party Internet device, wherein the platform source data includes: platform user features extracted using a platform feature extractor;

[0089] The distribution module 906 is used to distribute vehicle source data and any set of platform source data to target devices, wherein the target devices include at least one of the following: vehicle networking devices and third-party Internet devices;

[0090] The processing module 908 is used to perform data alignment processing on vehicle source data and at least one set of platform source data to generate training sample data, wherein the training sample data is feature data shared by vehicle networking devices and at least one third-party Internet platform.

[0091] Optionally, the first acquisition module 902 includes: a training unit, used to perform feature training on the local initialization model after the target device uses training sample data, and generate a model evaluation index associated with the target device, wherein the initialization model is a machine learning model; and an evaluation unit, used to optimize the local feature extractor after the target device uses the model evaluation index.

[0092] Optionally, the first acquisition module 902 includes: a first acquisition unit, used to acquire metadata of different vehicles before acquiring vehicle source data acquired by the vehicle networking device; a first extraction unit, used by a vehicle feature extractor deployed in the vehicle networking device to extract features from the metadata of different vehicles and acquire at least one set of vehicle user features of different vehicles; and a first integration unit, used to integrate at least one set of vehicle user features of different vehicles to generate vehicle source data.

[0093] Optionally, the first acquisition module 902 includes: a first encryption unit, used to perform differential privacy encryption on vehicle user features during the generation of vehicle source data, and distribute the encryption result as vehicle source data.

[0094] Optionally, the second acquisition module 904 includes: a second acquisition unit, used to acquire metadata of at least one third-party Internet device before acquiring a set of platform source data transmitted by at least one third-party Internet device; a second extraction unit, used for a platform feature extractor deployed in the third-party Internet device to extract features from the metadata of the corresponding third-party Internet device, and to acquire at least one set of platform user features of at least one third-party Internet device; and a second integration unit, used to integrate at least one set of platform user features of at least one third-party Internet device to generate platform source data.

[0095] Optionally, the second acquisition module 904 includes: a second encryption unit, which performs differential privacy encryption on platform user features during the generation of platform source data, and distributes the encryption result as platform source data.

[0096] Optionally, the distribution module 906 includes: a sending unit, used to distribute vehicle source data and / or any set of platform source data to different target devices in an inadvertent transmission manner, so that data exchange can be completed between the target devices.

[0097] Example 3

[0098] According to another aspect of the present invention, a data processing system is also provided, the system comprising the following components:

[0099] Vehicle-to-everything (V2X) devices are used to collect vehicle source data, which includes vehicle user features extracted using a vehicle feature extractor; at least one third-party internet device is used to collect platform source data, which includes platform user features extracted using a platform feature extractor.

[0100] In this process, the vehicle-to-everything (V2X) device and at least one third-party internet device distribute locally collected vehicle source data and any set of platform source data, enabling data exchange between the V2X device and at least one third-party internet device. Furthermore, both the V2X device and at least one third-party internet device perform data alignment processing on the vehicle source data and at least one set of platform source data to generate training sample data. The training sample data consists of feature data shared by the V2X device and the at least one third-party internet platform.

[0101] Example 4

[0102] According to another aspect of the present invention, a vehicle is also provided, comprising: one or more processors; a storage device for storing one or more programs; and a method for processing sample data such that the one or more processors execute any one of the programs when the one or more programs are executed by the one or more processors.

[0103] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0104] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0105] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0106] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0107] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0108] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0109] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for processing sample data, characterized in that, include: Acquire vehicle source data collected by vehicle networking devices, wherein the vehicle source data includes vehicle user features extracted using a vehicle feature extractor; Acquire a set of platform source data transmitted by at least one third-party Internet device, wherein the platform source data includes: platform user features extracted using a platform feature extractor; The vehicle source data and any set of platform source data are distributed to the target device, wherein the target device includes at least one of the following: the vehicle networking device and the third-party Internet device; The vehicle source data and at least one set of platform source data are aligned to generate training sample data, wherein the training sample data consists of feature data shared by the vehicle networking device and the at least one third-party Internet platform. Before acquiring a set of platform source data transmitted by at least one third-party internet device, the method further includes: collecting metadata of at least one third-party internet device; the platform feature extractor deployed in the third-party internet device extracting features from the metadata of the corresponding third-party internet device to acquire at least one set of platform user features of the at least one third-party internet device; and integrating the at least one set of platform user features of the at least one third-party internet device to generate the platform source data.

2. The method according to claim 1, characterized in that, After generating training sample data, the method further includes: The target device uses the training sample data to perform feature training on the local initialization model and generates a model evaluation index associated with the target device, wherein the initialization model is a machine learning model; The target device uses the model evaluation index to optimize its local feature extractor.

3. The method according to claim 1, characterized in that, Before acquiring the vehicle source data collected by the vehicle networking device, the method further includes: Collect metadata from different vehicles; The vehicle feature extractor deployed in the vehicle networking device extracts features from the metadata of the different vehicles to obtain at least one set of vehicle user features for the different vehicles. The vehicle source data is generated by integrating at least one set of vehicle user characteristics from the different vehicles.

4. The method according to claim 3, characterized in that, In the process of generating the vehicle source data, the method further includes: The vehicle user characteristics are subjected to differential privacy encryption, and the encryption result is distributed as the vehicle source data.

5. The method according to claim 1, characterized in that, In the process of generating the platform source data, the method further includes: The platform user characteristics are subjected to differential privacy encryption, and the encryption result is distributed as the platform source data.

6. The method according to claim 1, characterized in that, The vehicle source data and / or any set of platform source data are distributed to different target devices in an unintentional transmission manner, enabling data exchange between the target devices.

7. A sample data processing apparatus, characterized in that, include: The first acquisition module is used to acquire vehicle source data collected by the vehicle networking device, wherein the vehicle source data includes vehicle user features extracted using a vehicle feature extractor. The second acquisition module is used to acquire a set of platform source data transmitted by at least one third-party Internet device, wherein the platform source data includes: platform user features extracted using a platform feature extractor; The distribution module is used to distribute the vehicle source data and any set of platform source data to the target device, wherein the target device includes at least one of the following: the vehicle networking device and the third-party Internet device; The processing module is used to perform data alignment processing on the vehicle source data and at least one set of platform source data to generate training sample data, wherein the training sample data is feature data shared by the vehicle networking device and the at least one third-party Internet platform. The second acquisition module is further configured to: collect metadata of at least one third-party Internet device; the platform feature extractor deployed in the third-party Internet device performs feature extraction on the metadata of the corresponding third-party Internet device to obtain at least one set of platform user features of the at least one third-party Internet device; and integrate the at least one set of platform user features of the at least one third-party Internet device to generate the platform source data.

8. A sample data processing system, characterized in that, include: A vehicle-to-everything (V2X) device for collecting vehicle source data, wherein the vehicle source data includes vehicle user features extracted using a vehicle feature extractor; At least one third-party internet device is used to collect platform source data, wherein the platform source data includes: platform user features extracted using a platform feature extractor; The vehicle-to-everything (V2X) device and the at least one third-party internet device distribute the locally collected vehicle source data and any set of platform source data, enabling data exchange between the V2X device and the at least one third-party internet device. Furthermore, both the V2X device and the at least one third-party internet device perform data alignment processing on the vehicle source data and at least one set of platform source data to generate training sample data. The training sample data consists of shared feature data between the V2X device and the at least one third-party internet platform. The at least one third-party internet device is further used to collect metadata of at least one third-party internet device; the platform feature extractor deployed in the third-party internet device performs feature extraction on the metadata of the corresponding third-party internet device to obtain at least one set of platform user features of the at least one third-party internet device; and the at least one set of platform user features of the at least one third-party internet device are integrated to generate the platform source data.

9. A vehicle, characterized in that, include: One or more processors; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors perform the sample data processing method according to any one of claims 1-6.

Images