CN111131227B - Data processing method and device - Google Patents

Abstract

The embodiment of the present invention discloses a data processing method and device, wherein the method includes: the server receives the first encrypted data sent by the client, at least uses the first encrypted data to train to obtain an encrypted transformation model, and sends the encrypted transformation model to the client, And receive the decrypted transformation model obtained by decrypting the encrypted transformation model by the client using the preset private key. In the embodiment of the present invention, the server always uses encrypted conversion data to train the model, and the decryption private key is stored in the client. Therefore, the server cannot decrypt the user's conversion data, thereby ensuring the security of the user's private data and improving the client's experience. better.

Description

A data processing method and device

technical field

The present invention relates to the technical field of financial technology (Fintech), in particular to a data processing method and device.

Background technique

With the development of computer technology, more and more technologies are applied in the financial field, and the traditional financial industry is gradually transforming into financial technology (Fintech). However, due to the security and real-time requirements of the financial industry, the financial field also requires technology higher requirements. In the field of user conversion technology, the client entrusts the server to promote the client's products. After receiving the client's entrustment, the server can use the conversion model to predict users who may actually convert, and then promote products to these users. It can be seen from this that the conversion effect of the conversion model directly affects the delivery effect of the product. If you want the product to have a better delivery effect, you need to improve the conversion effect of the conversion model.

For the client, a conversion model with a better conversion effect usually means a higher conversion cost. In order to reduce the conversion cost, the client can also expose the conversion data to the server so that the server can use more comprehensive data to train the conversion model , improving the conversion performance of the conversion model. However, the conversion data includes the user's privacy data, such as the user's residential address, mobile phone number, ID number, etc. If the conversion data is exposed to the server, the user's privacy data may not be protected; and, the server itself There is a risk of leaking the user's private data. If the server sells the user's private data to other clients of the same type, the experience of the client will be poor and the client will suffer a great loss.

To sum up, there is an urgent need for a data processing method to solve the technical problems of insecure user privacy data and poor client experience caused by exposing and transforming data to the server in the prior art.

Contents of the invention

The embodiment of the present invention provides a data processing method, which is used to solve the technical problems of insecure private data of the user and poor experience of the client caused by exposing and transforming data to the server in the prior art.

In a first aspect, an embodiment of the present invention provides a data processing method, the method comprising:

The server receives the first encrypted data sent by the client, and the first encrypted data is data after the client uses the preset public key to encrypt the converted data; further, the server uses at least the first encrypted data An encrypted transformation model is obtained through training with encrypted data, and the encrypted transformation model is sent to the client, and the decryption transformation model sent by the client is received, and the decryption transformation model is used by the client using the preset public key The corresponding preset private key is obtained by decrypting the encrypted conversion model.

In the above design, after receiving the encrypted conversion data sent by the client, the server at least uses the encrypted conversion data to train the encrypted conversion model, and then the client decrypts the encrypted conversion model and sends it to the server; it can be seen that, in the conversion model During the training process, the server always uses encrypted conversion data for model training, and the decryption private key is stored on the client. Therefore, the server cannot decrypt the user's conversion data, thereby ensuring the security of the user's private data and improving the client's experience. better.

In a possible design, the server at least uses the first encrypted data to train an encrypted conversion model, including: the server obtains the click data corresponding to the first encrypted data, and uses the preset public key Encrypting the click data to obtain second encrypted data; the preset public key is sent by the client to the server; the server performs training according to the first encrypted data and the second encrypted data , to obtain the encrypted transformation model.

In the above design, by using encrypted click data and encrypted conversion data to jointly train the conversion model, the conversion effect of the conversion model can be improved.

In a possible design, before the server receives the first encrypted data sent by the client, it also receives a request message sent by the client; the request message includes a data range, and according to the request message, obtains a click data in the data range, and send the click data to the client, where the click data is used by the client to obtain the conversion data corresponding to the click data.

In the above design, by sending a request message from the client to the server, the client can determine the data range of the samples participating in the model training, realizing the initiative of the client to control the model training process; and, the client can set the data according to actual needs range, so that the solution can meet the needs of different clients and is more in line with actual scenarios.

In a possible design, the request message also includes a user name and a key; before the server sends the click data to the client, it also queries a preset correspondence table to determine the preset correspondence The user name exists in the table, and the key corresponds to the user name; the preset correspondence table includes the correspondence between the user name and the key of the client registered in the server in advance.

In the above design, the client can register with the server in advance. In this way, for the request message sent by the client, the server can first authenticate the client according to the registration information, and start the server after confirming that the client is safe and has not been tampered with. Model training, so as to ensure the security of model training and prevent client permissions from being tampered with.

In a second aspect, an embodiment of the present invention provides a data processing method, the method comprising:

The client uses a preset public key to encrypt the converted data to obtain first encrypted data, sends the first encrypted data to the server, and receives the encrypted conversion model sent by the server, the encrypted conversion model is the at least obtained by using the first encrypted data training, further, the client uses the preset private key corresponding to the preset public key to decrypt the encrypted conversion model to obtain a decrypted conversion model, and the decrypted The conversion model is sent to the server.

In the above design, the client sends the encrypted conversion data to the server, and after receiving the encrypted conversion model sent by the server, decrypts to obtain the decrypted conversion model; it can be seen that the encrypted conversion data and the decrypted conversion model are transmitted through the client For the server, the data interaction process can be completed on the premise of ensuring the security of the user's private data, and the server can also use the encrypted conversion data provided by the client to train an encrypted conversion model with a better conversion effect, thereby improving the conversion effect of the user.

In a possible design, before the client uses a preset public key to encrypt the conversion data, it also sends a request message to the server, the request message includes a data range, and the request message is used for the The server acquires click data satisfying the data range, receives the click data sent by the server, and acquires the conversion data corresponding to the click data.

In a third aspect, an embodiment of the present invention provides a data processing device, the device comprising:

A transceiver module, configured to receive the first encrypted data sent by the client, where the first encrypted data is the data after the client uses the preset public key to encrypt the converted data;

A training module, configured to at least use the first encrypted data to train an encrypted conversion model;

The transceiver module is further configured to send the encryption conversion model to the client, and receive the decryption conversion model sent by the client, and the decryption conversion model uses the preset public key for the client The corresponding preset private key is obtained by decrypting the encrypted conversion model.

In a possible design, the training module is specifically configured to: obtain click data corresponding to the first encrypted data, and use the preset public key to encrypt the click data to obtain second encrypted data; The preset public key is sent from the client to the server; training is performed according to the first encrypted data and the second encrypted data to obtain the encrypted conversion model.

In a possible design, the device further includes a processing module, before the transceiver module receives the first encrypted data sent by the client, it is further configured to: receive a request message sent by the client; Including a data range; correspondingly, the processing module is configured to: acquire click data satisfying the data range according to the request message; the transceiver module is also configured to: send the click data to the client, the The click data is used by the client to obtain the conversion data corresponding to the click data.

In a possible design, the request message also includes a user name and a key; before the sending and receiving module sends the click data to the client, the processing module is also configured to: query a preset correspondence table , determining that the user name exists in the preset correspondence table, and the key corresponds to the user name; the preset correspondence table includes the correspondence between the user name and the key of the client registered in the server in advance relation.

In a fourth aspect, a schematic structural diagram of a data processing device provided by an embodiment of the present invention includes:

An encryption module, configured to encrypt the converted data using a preset public key to obtain the first encrypted data;

A transceiver module, configured to send the first encrypted data to a server, and receive an encrypted conversion model sent by the server, where the encrypted conversion model is trained by the server using at least the first encrypted data;

A decryption module, configured to use a preset private key corresponding to the preset public key to decrypt the encrypted conversion model to obtain a decrypted conversion model;

The transceiver module is further configured to send the decrypted conversion model to the server.

In a possible design, before the encryption module encrypts the converted data with a preset public key, the transceiver module is further configured to: send a request message to the server, the request message includes a data range, and the The request message is used for the server to obtain the click data satisfying the data range; and, to receive the click data sent by the server; correspondingly, the encryption module is also used for: obtaining the click data corresponding to the click data The conversion data.

In a fifth aspect, an embodiment of the present invention provides a computing device, including at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program, and when the program is executed by the processing unit, the The processing unit executes the data processing method described in any one of the first aspect or the second aspect.

In a sixth aspect, an embodiment of the present invention provides a computer-readable storage medium, which stores a computer program executable by a computing device, and when the program runs on the computing device, the computing device executes the above-mentioned first Any data processing method described in the first aspect or the second aspect.

These or other aspects of the present invention will be more clearly understood in the description of the following embodiments.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings that need to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For Those skilled in the art can also obtain other drawings based on these drawings without any creative effort.

FIG. 1 is a schematic diagram of a possible system architecture provided by an embodiment of the present invention;

FIG. 2 is a schematic diagram of an interaction process corresponding to a data processing method provided by an embodiment of the present invention;

FIG. 3 is a schematic structural diagram of a data processing device provided by an embodiment of the present invention;

FIG. 4 is a schematic structural diagram of another data processing device provided by an embodiment of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some of the embodiments of the present invention, rather than all of them. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

The technical field of financial technology (Fintech) usually involves a variety of transactions. For example, bank transactions can include card sales transactions, deposit transactions, loan transactions, insurance transactions, wealth management transactions, etc. The daily transaction volume of banks can reach thousands or even several Thousands of strokes. Taking loan transactions as an example, in order to ensure the conversion effect of loan transactions, banks can usually entrust the server to advertise loan products to users; therefore, the server needs to use the conversion model to predict the target users who are interested in loan products, so as to deliver loan products to Target users to obtain better conversion effects.

For ease of understanding, the following embodiments of the present invention use promotion of advertising products to users as an example to describe the data processing method in the embodiments of the present invention. It should be noted that although the following embodiments of the present invention use the promotion process of advertising products as an example The data processing method in the embodiment of the present invention is described, but it can be understood that the data processing method in the embodiment of the present invention can also be applied in any other promotion scenario or conversion scenario, without limitation.

In the promotion scenario of advertising products, the server refers to the advertising platform. After the server selects the target users for the advertising products through the conversion model, it can expose the advertisements of the advertising products to the target users. If the target users are interested in the advertising products, they can click The advertisement enters the purchase page of the advertising product, so that the server can record the pre-conversion data (such as exposure data and click data) during the advertisement delivery process; correspondingly, the client refers to the advertiser, because the actual conversion process of the advertising product passes through the client Control, therefore, if the target user purchases the advertised product on the purchase page of the advertised product, the client can record the conversion data (such as purchase data) during the advertisement delivery process, and the conversion data is associated with the user's actual purchase data, which can be Identify the actual conversion results of users to the advertised product. It can be seen that the exposure data in the exposure stage and the click data in the click stage are maintained by the server, and the conversion data in the conversion stage are maintained by the client. Therefore, since only exposure data and click data are stored in the server, the server can only use exposure data and click data to train the conversion model. Obviously, if the server wants to train a conversion model with better effect, it only needs to use Clicking data is not enough.

In order to solve the above problems, in a possible implementation, the client can send the conversion data to the server in plain text, so that the server can jointly use the exposure data, click data and conversion data to jointly train to obtain conversions with better conversion effects Model. However, since the conversion data includes the user's privacy data, this method may lead to leakage of the user's privacy data because the conversion data is not protected. It can be seen that in the field of advertising delivery technology, there is an antagonistic relationship between the conversion effect of the conversion model and the protection of the user's private data. If you want to train a conversion model with a better effect, the security of the user's private data cannot be obtained. Guarantee, if you want to protect the user's private data, the effect of the conversion model is not good.

Based on this, an embodiment of the present invention proposes a data processing method to improve the conversion effect of the conversion model on the premise of protecting the security of the user's private data.

FIG. 1 is a schematic diagram of a possible system architecture provided by an embodiment of the present invention. As shown in FIG. 1, the system architecture includes a server 110 and at least one client, such as a client 121, a client 122, and a client 123; wherein , at least one client may communicate with the server 110 respectively, for example, the communication connection may be implemented in a wired manner, or the communication connection may also be implemented in a wireless manner, which is not specifically limited.

As shown in Figure 1, at least one client can also be included in the system architecture, such as client 131, client 132 and client 133, and at least one client can also be respectively connected to the server 110 in communication; wherein, the client can be Refers to terminal devices, such as mobile phones, notebooks, tablet computers, Ipads, etc., without limitation.

In a specific implementation, at least one client may send an entrustment request to the server 110, and the entrustment request is used for the server to convert the advertising product; correspondingly, after receiving the entrustment request, the server 110 may use the conversion model to predict the at least one client , selecting a target user terminal interested in the advertising product from at least one user terminal, and delivering an advertisement of the advertising product to the target user terminal.

As an example, the server 110 can be a global wide area network (World Wide Web, web) server, and a web browser can be provided on any user end, and the server 110 can display the advertisement of the advertisement product through the web browser on the target user end To the target user, and can record the exposure data of the advertised product. Correspondingly, if the target user is interested in the advertised product, he can click the preset button or preset link on the advertisement of the advertised product to enter the product introduction page or purchase page of the advertised product; the server 110 can record the target user's interest in the advertised product. Click data for operations on advertising pages. Moreover, since the purchase page of the advertised product is maintained by the client marketing the advertised product, the client can record the conversion data of the target user to the advertised product.

It should be noted that the above is only an exemplary and simple description, and the delivery methods of the advertising products listed are only for the convenience of explaining the solution, and do not constitute a limitation on the solution; in specific implementation, the server can also use other methods Advertising products can be placed, for example, through smart TVs, video playback applications, radio or radio, etc., without limitation.

Based on the system architecture shown in FIG. 1, FIG. 2 is a schematic flowchart corresponding to a data processing method provided by an embodiment of the present invention, the method comprising:

Step 201, the client uses a preset public key to encrypt the converted data to obtain first encrypted data.

In the embodiment of the present invention, there may be multiple ways for the client to obtain the first encrypted data, and two possible implementation ways are introduced below:

In a possible implementation, if the client determines to launch a certain advertising product, it can obtain the conversion data corresponding to the advertising product from its built-in database, and can use the preset public key to perform conversion data corresponding to the advertising product. Encrypt to obtain the first encrypted data. Wherein, the preset public key may be a public key determined using a first preset encryption algorithm, and the first preset encryption algorithm may be set by those skilled in the art based on experience.

For example, if the advertising product is advertising product A, the client can obtain the conversion data of all users who purchased advertising product A, and can use the preset public key to encrypt the conversion data of each user to obtain multiple encrypted Conversion data; further, the client can set an identifier for each piece of encrypted conversion data, and the identifier of the conversion data is used to uniquely identify a piece of conversion data.

In the embodiment of the present invention, the identification of the converted data can be set by those skilled in the art based on experience, for example, it can be the user's ID card number, the user's phone number, etc., which is not specifically limited.

In an example, the preset public key may be a public key determined using a differential privacy encryption algorithm. Since the differential privacy encryption algorithm can encrypt different feature values of the same feature label into irregular encrypted feature values, the differential privacy encryption algorithm is used After the conversion data of the user is encrypted, the server 110 cannot determine whether the user has successfully purchased the advertising product according to the encrypted conversion data, thereby ensuring the security of the user's private data. For example, user 1 is a user who has purchased advertising product A in the last 2 months, and users 2 to 5 are users who have clicked on advertising product A in the last 2 months but have not purchased advertising product A, then the conversion results For this feature label, after encrypting the conversion data of users 1 to 5 using the preset public key determined by the differential privacy encryption algorithm, the encrypted feature value corresponding to the conversion result of user 1 is 11100001, and the encrypted feature value corresponding to the conversion result of user 2 is 11100001. The feature value is 11110001, the encrypted feature value corresponding to the conversion result of user 3 is 10110001, the encrypted feature value corresponding to the conversion result of user 4 is 10110101, and the encrypted feature value corresponding to the conversion result of user 5 is 10110011; obviously, even if the server 110 Knowing that there is one user among users 1 to 5 who purchased the advertising product A within the last 2 months, the server 110 cannot determine which user purchased the advertising product A according to the encrypted feature value of the conversion results of users 1 to 5 .

In another possible implementation, the client can first send a request message to the server 110, and the request message includes the data range; correspondingly, after the server 110 obtains the request message, it can obtain the data from its built-in database according to the request message. Click data that meets the data range, and can send the click data to the client; in this way, after receiving the click data sent by the server 110, the client can obtain the conversion data corresponding to the click data from its built-in database, and then use The preset public key encrypts the converted data to obtain the first encrypted data. In this implementation, by sending a request message from the client to the server, the client can determine the data range of the samples participating in the model training, realizing the initiative of the client to control the model training process; and, the client can according to actual needs Set the data range, so that the solution can meet the needs of different clients and is more in line with the actual scene.

Wherein, the data range can be set by those skilled in the art based on experience, for example, it can be a preset time period, purchase amount, order area, etc., and is not limited.

For example, if the advertisement product is advertisement product A, and the data range in the request message is the preset time period 2019.7.29-2019.8.29, the server 110 can obtain a The click data of all users who have clicked on the advertising product A within a month's time period, and an identifier can be set for each user's click data. The identifier can be the user's ID number, phone number, etc., without limitation. In this way, the click data sent by the server 110 to the client may be composed of multiple pieces of user click data. When the client receives multiple pieces of user click data, it can query the built-in database to obtain information that matches the identifiers of the multiple user click data. The user's conversion data as conversion data corresponding to the click data.

It should be noted that the above is only an exemplary simple description, and does not constitute a limitation on the solution; in specific implementation, the server 110 may only obtain the click data of the advertisement of the advertising product within the data range, and store the click data Send to the client, or you can only obtain the exposure data of the advertisement of the advertising product within the data range, and send the exposure data to the client, or you can also obtain the exposure data and click data of the advertisement of the advertising product within the data range , and send the exposure data and click data to the client, without limitation.

Furthermore, for any piece of corresponding click data and conversion data (that is, the click data and conversion data of the same user), the client can directly use the preset public key to encrypt the conversion data to obtain a piece of encrypted data, or it can first The click data and conversion data are mixed, and then the mixed click data and conversion data are encrypted to obtain a piece of encrypted data. Correspondingly, if the server 110 sends exposure data and click data to the client at the same time, the client may first mix the corresponding exposure data, click data, and conversion data, and then mix the mixed exposure data, click data, and conversion data. The data is encrypted to obtain a piece of encrypted data.

In one possible scenario, the user may need to trigger multiple operations on the advertisement page of the advertised product before entering the purchase page of the advertised product. For example, the conversion process of the advertisement of Weiweidai product is as follows: Enter the user information filling page. After the user information is completed, it will enter the enterprise information filling page. After the enterprise information is completed, it will enter the quota application page. After the quota application is completed, it will enter the purchase page of Weiweidai products; therefore , the click data corresponding to the advertisement of Weiweidai products includes the click data of the user on the user information filling page, the click data of the user on the enterprise information filling page, and the click data of the user on the quota application page. In this scenario, the click data sent by the server 110 to the client may refer to the click data of the last page, that is, the click operation information that enters the purchase page of the advertised product after the click, and in the above example refers to the click data on the quota application page. Click Data.

In one example, the client can register an account on the server 110 in advance, and the account includes a user name and a key. After the server 110 creates permissions for the user name and key registered by the client, the information sent by the client to the server 110 can be It carries the user name and key, so that the server 110 can authenticate the client before processing the data; correspondingly, the server 110 can store the user names and keys of all pre-registered clients in the preset correspondence table. In a specific implementation, the request message sent by the client to the server 110 may include the user name, key, and data range at the same time. In this way, after receiving the request message, the server 110 can first analyze the user name and key of the client, and then query the preset Set the correspondence table to determine whether the username exists in the preset correspondence table, if the username does not exist in the preset correspondence table, the server 110 can send an abnormal response message to the client; if the preset correspondence table exists If the user name is the user name, the server 110 can inquire whether the key corresponds to the user name. If it does not correspond, it means that the client forges the user name or key, so the server 110 can warn the client. If it is corresponding, the server 110 The click data of the data range can be obtained and sent to the client.

In the above example, the client can pre-register with the server. In this way, for the request message sent by the client, the server can first authenticate the client according to the registration information, and start the server after determining that the client is safe and has not been tampered with. Model training, so as to ensure the security of model training and prevent client permissions from being tampered with.

Step 202, the client sends the first encrypted data to the server.

In a specific implementation, the client may generate a message according to the Internet Protocol Address (Internet Protocol Address, IP) of the client, the IP of the server 110 and the first encrypted data, and may send the message to the server 110 . Wherein, if the client and the server 110 are connected by a wired method (such as optical fiber, cable, etc.), the client can send the message to the server 110 by a wired method; if the client and the server 110 are connected by a wireless method (such as WIFI, Bluetooth, etc. ) connection, the client can send the message to the server 110 in a wireless manner.

Step 203: After receiving the first encrypted data sent by the client, the server at least uses the first encrypted data to train to obtain an encrypted conversion model.

In a specific implementation, the process of training the model by the server 110 can be set based on a specific business scenario. For example, if the server 110 only has the function of model training, the server 110 can obtain an encrypted conversion model based only on the first encrypted data training sent by the client. If the server 110 also has the function of data storage, then the server 110 can use the first encrypted data sent by the client and the data stored in the server 110 to jointly train to obtain an encrypted conversion model; if the server 110 also has the function of interacting with a third party, then The server 110 may first obtain the data in the third party from the third party, and then use the first encrypted data sent by the client and the data in the third party to jointly train to obtain an encrypted conversion model, and so on.

In a possible implementation, after receiving the first encrypted data sent by the client, the server 110 may first obtain the click data corresponding to the first encrypted data, and encrypt the click data with a preset public key to obtain the second encrypted data. Encrypt the data, and then use the first encrypted data and the second encrypted data to jointly train to obtain an encrypted conversion model. Wherein, the preset public key may be sent to the server 110 by the client in advance, or may be carried by the client in a message corresponding to the first encrypted data and sent to the server 110, which is not limited.

Specifically, the first encrypted data received by the server 110 may include multiple pieces of encrypted conversion data, and each piece of encrypted conversion data has a unique identifier, such as the user's ID number, the user's mobile phone number, and so on. In this way, the server 110 may first obtain the click data satisfying the data range from the built-in database, and then determine the click data corresponding to each piece of encrypted conversion data according to the identifier of the encrypted conversion data and the identifier of the click data.

It should be noted that the click data corresponding to the conversion data obtained by the server 110 may include the user's click data on all pages of the advertising product. For example, for Weiweidai products, the click data obtained by the server 110 according to the data range may only include The user's click data on the quota application page, and the click data obtained according to the conversion data may include the user's click data on the user information filling page, the user's click data on the enterprise information filling page, and the user's quota application. Click data on the page.

Further, for any corresponding conversion data and click data, the server 110 may use a preset public key to encrypt the click data to obtain a piece of encrypted click data, and may set an identifier for the piece of encrypted click data; wherein, The identifier of the piece of encrypted click data matches the identifier of the corresponding encrypted conversion data. It can be seen that the second encrypted data is composed of multiple pieces of encrypted click data, and each piece of encrypted click data may correspond to one piece of encrypted conversion data.

In the embodiment of the present invention, the type of the encryption conversion model can be set by those skilled in the art based on experience, for example, it can be a classification model, or it can also be a neural network model, or it can also be a regression model, which is not limited in the embodiment of the present invention .

It should be noted that the encrypted conversion model in step 203 can be any one or more of the encrypted purchase conversion model, encrypted click conversion model and encrypted comprehensive conversion model; specifically, if the encrypted click data and The encrypted conversion data is trained to obtain the encrypted purchase conversion model, and the purchase conversion model can predict the probability of the user who clicks on the advertisement of the advertised product to purchase the advertised product; if the encrypted exposure data and encrypted click data are used to train the encrypted click conversion model, Then the click conversion model can predict the probability of the user clicking on the advertisement after viewing the advertisement page of the advertised product; if the encrypted comprehensive conversion model is obtained by training with encrypted exposure data, encrypted click data and encrypted conversion data, the comprehensive conversion model can predict The probability that a user who views the advertisement page of the advertised product purchases the advertised product.

Step 204, the server sends the encrypted conversion model to the client.

In a specific implementation, the server 110 can obtain the model parameters of the encrypted conversion model, and can generate a message according to the encrypted model parameters, the IP of the server 110, and the IP of the client, and send the message to the client.

Step 205: After receiving the encrypted transformation model, the client uses the preset private key corresponding to the preset public key to decrypt the encrypted transformation model to obtain the decrypted transformation model.

In one example, the client may store the corresponding relationship between advertising products, preset public keys and preset private keys, and when generating conversion models corresponding to multiple advertising products, different preset public key pairs may be used respectively The conversion data corresponding to multiple advertising products are encrypted, and the preset public keys of the conversion data corresponding to any two advertising products are different; thus, after receiving the encrypted conversion model corresponding to an advertising product sent by the server 110, the client can Query the corresponding relationship to determine the preset private key corresponding to the advertising product, so as to decrypt the encrypted conversion model corresponding to the advertising product by using the preset private key corresponding to the advertising product.

In the embodiment of the present invention, the corresponding preset private key and preset public key may be the private key and public key determined using the differential privacy algorithm, or may also be the private key and public key determined using the homomorphic encryption algorithm, without limitation .

Step 206, the client sends the decrypted conversion model to the server.

Here, the server 110 may directly send the decrypted transformation model to the client, or in order to ensure the security of data transmission, it may first encrypt the decrypted transformation model using a second preset encryption algorithm, and then encrypt the encrypted transformation model The model is sent to the client; wherein, the second preset encryption algorithm can be set by those skilled in the art based on experience, for example, it can be a preset hash algorithm, a symmetric encryption algorithm, an asymmetric encryption algorithm, and the like.

It should be noted that the second preset encryption algorithm is pre-negotiated and determined between the client and the server 110, that is, the server 110 may store a decryption algorithm corresponding to the second preset encryption algorithm; Independently set, the server 110 does not store the decryption algorithm corresponding to the first preset encryption algorithm, so the server 110 cannot decrypt the encrypted conversion data, ensuring the security of the user's private data.

Step 207: After receiving the decrypted conversion model sent by the client, the server uses the decrypted conversion model to determine the target user, and delivers to the target user.

In specific implementation, after receiving the decrypted conversion model sent by the client, the server 110 can predict each user according to the decrypted conversion model. As a target user of the advertising product, in this way, the server 110 may deliver an advertisement of the advertising product to the target user; correspondingly, if it is determined that the probability that the user is converted by the advertising product is less than a preset threshold value, then the user may not be served the advertisement of the product advertise. For example, when the preset threshold is 80%, if the decrypted conversion model is a comprehensive conversion model, the server 110 can use the comprehensive conversion model to respectively determine the probabilities of the users corresponding to the user terminals 131 to 133 to purchase advertising products, if the user The probability that the user corresponding to the terminal 131 purchases the advertising product is 90%, the probability that the user corresponding to the user terminal 132 purchases the advertising product is 60%, and the probability that the user corresponding to the user terminal 133 purchases the advertising product is 85%, then the server 110 can send the user The user corresponding to the terminal 131 and the user corresponding to the user terminal 133 are two target users of the advertising product, so the advertisement of the advertising product can be delivered to the two target users through the web browser of the user terminal 131 and the web browser of the user terminal 133 respectively .

In the above embodiments of the present invention, the server receives the first encrypted data sent by the client, and the first encrypted data is the data after the client encrypts the converted data using the preset public key; further, The server at least uses the first encrypted data to train to obtain an encrypted conversion model, sends the encrypted conversion model to the client, and receives the decrypted conversion model sent by the client, and the decrypted conversion model is determined by the client obtained by decrypting the encrypted transformation model by using the preset private key corresponding to the preset public key. In the embodiment of the present invention, after receiving the encrypted conversion data sent by the client, the server at least uses the encrypted conversion data to train the encrypted conversion model, and then the client decrypts the encrypted conversion model and sends it to the server; During the training process of the model, the server always uses encrypted conversion data for model training, and the decryption private key is stored on the client. Therefore, the server cannot decrypt the conversion data of the user, thereby ensuring the security of the user's private data. The client's The experience is better.

With regard to the above-mentioned method flow, the embodiment of the present invention also provides a data processing device, and the specific content of the device can be implemented with reference to the above-mentioned method.

Fig. 3 is a schematic structural diagram of a data processing device provided by an embodiment of the present invention, including:

The transceiver module 301 is configured to receive the first encrypted data sent by the client, the first encrypted data is the data after the client uses the preset public key to encrypt the converted data;

A training module 302, configured to use at least the first encrypted data to train to obtain an encrypted transformation model;

The transceiver module 301 is further configured to send the encryption conversion model to the client, and receive the decryption conversion model sent by the client, the decryption conversion model is the client using the preset public It is obtained by decrypting the encrypted transformation model with the preset private key corresponding to the key.

Optionally, the training module 302 is specifically used for:

Obtain the click data corresponding to the first encrypted data, and use the preset public key to encrypt the click data to obtain second encrypted data; the preset public key is sent to the server by the client ;

performing training according to the first encrypted data and the second encrypted data to obtain the encrypted conversion model.

Optionally, the device further includes a processing module 303, before the transceiver module 301 receives the first encrypted data sent by the client, it is further configured to:

receiving a request message sent by the client; the request message includes a data range;

The processing module 303 is used for:

Obtain click data satisfying the data range according to the request message;

The transceiver module 301 is also used for:

Sending the click data to the client, where the click data is used by the client to obtain the conversion data corresponding to the click data.

Optionally, the request message also includes a username and a key;

Before the transceiver module 301 sends the click data to the client, the processing module 303 is further configured to:

Querying the preset correspondence table to determine that the user name exists in the preset correspondence table, and the key corresponds to the user name; the preset correspondence table includes users of clients registered in the server in advance Correspondence between names and keys.

Fig. 4 is a schematic structural diagram of another data processing device provided by an embodiment of the present invention, including:

An encryption module 401, configured to encrypt the converted data using a preset public key to obtain first encrypted data;

A transceiver module 402, configured to send the first encrypted data to a server, and receive an encrypted transformation model sent by the server, where the encrypted transformation model is trained by the server using at least the first encrypted data;

Decryption module 403, configured to use the preset private key corresponding to the preset public key to decrypt the encrypted conversion model to obtain the decrypted conversion model;

The transceiving module 402 is further configured to send the decrypted transformation model to the server.

Optionally, before the encryption module 401 uses a preset public key to encrypt the conversion data, the transceiver module 402 is also used to:

Send a request message to the server, where the request message includes a data range, and the request message is used for the server to obtain click data satisfying the data range; and receive the click data sent by the server;

The encryption module 401 is also used for:

Acquiring the conversion data corresponding to the click data.

It can be seen from the above that in the above embodiments of the present invention, the server receives the first encrypted data sent by the client, and the first encrypted data is encrypted by the client using the preset public key to convert the data further, the server at least uses the first encrypted data training to obtain an encrypted conversion model, sends the encrypted conversion model to the client, and receives the decrypted conversion model sent by the client, the The decrypted transformation model is obtained by the client decrypting the encrypted transformation model using the preset private key corresponding to the preset public key. In the embodiment of the present invention, after receiving the encrypted conversion data sent by the client, the server at least uses the encrypted conversion data to train the encrypted conversion model, and then the client decrypts the encrypted conversion model and sends it to the server; During the training process of the model, the server always uses encrypted conversion data for model training, and the decryption private key is stored on the client. Therefore, the server cannot decrypt the conversion data of the user, thereby ensuring the security of the user's private data. The client's The experience is better.

Based on the same inventive concept, an embodiment of the present invention also provides a computing device, including at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program, and when the program is executed by the processing unit, The processing unit is made to execute any data processing method described in FIG. 2 .

Based on the same inventive concept, an embodiment of the present invention also provides a computer-readable storage medium, which stores a computer program executable by a computing device, and when the program is run on the computing device, the computing device executes the 2 Any of the data processing methods described above.

Those skilled in the art should understand that the embodiments of the present invention may be provided as methods or computer program products. Accordingly, the present invention can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and a combination of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing equipment produce a An apparatus for realizing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions The device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby The instructions provide steps for implementing the functions specified in the flow chart or blocks of the flowchart and/or the block or blocks of the block diagrams.

While preferred embodiments of the invention have been described, additional changes and modifications to these embodiments can be made by those skilled in the art once the basic inventive concept is appreciated. Therefore, it is intended that the appended claims be construed to cover the preferred embodiment as well as all changes and modifications which fall within the scope of the invention.

Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these modifications and variations.

Claims (8)

1. A method of data processing, the method comprising:

a server receives a request message sent by a client, wherein the request message comprises a data range;

the server acquires click data meeting the data range according to the request message, wherein the click data is data of click operation of a user on an advertisement page of an advertisement product;

the server sends the click data to the client, so that the client can acquire conversion data matched with the click data according to the click data, wherein the conversion data is data for the user to purchase the advertisement product after entering a purchase page through the advertisement page;

the server receives first encrypted data sent by the client, wherein the first encrypted data is data obtained by encrypting the converted data by the client by using a preset public key;

the server obtains an encrypted purchase conversion model by using the first encrypted data and the click data for training;

the server sends the encrypted purchase conversion model to the client, wherein the encrypted purchase conversion model is used for the client to decrypt by using a preset private key corresponding to the preset public key to obtain the purchase conversion model;

and the server receives the purchase conversion model sent by the client and predicts the probability of the user clicking the advertisement page of the advertisement product purchasing the advertisement product by using the purchase conversion model, and if the probability of the user clicking the advertisement page of the advertisement product purchasing the advertisement product is determined to be greater than or equal to a preset threshold value, the user is taken as a target user of the advertisement product.

2. The method of claim 1, wherein the server trains an encrypted purchase conversion model using the first encrypted data and the click data, comprising:

the server acquires click data corresponding to the first encrypted data, and encrypts the click data by using the preset public key to obtain second encrypted data; the preset public key is sent to the server by the client;

and the server trains according to the first encrypted data and the second encrypted data to obtain the encrypted purchase conversion model.

3. The method according to claim 1 or 2, characterized in that the request message further comprises a user name and a key;

before the server sends click data to the client, the method further comprises the following steps:

the server inquires a preset corresponding relation table, and determines that the user name exists in the preset corresponding relation table, and the secret key corresponds to the user name; the preset corresponding relation table comprises the corresponding relation between the user name and the key of the client registered in the server in advance.

4. A method of data processing, the method comprising:

a client sends a request message to a server, wherein the request message comprises a data range, the request message is used for the server to obtain click data meeting the data range, and the click data is data for clicking an advertisement page of an advertisement product by a user;

the client receives the click data sent by the server;

the client acquires conversion data matched with the click data according to the click data, wherein the conversion data is data for the user to purchase the advertisement product after entering a purchase page through the advertisement page;

the client encrypts the converted data by using a preset public key to obtain first encrypted data;

the client sends the first encrypted data to a server;

the client receives an encrypted purchase conversion model sent by the server, wherein the encrypted purchase conversion model is obtained by training the server by using the first encrypted data and the click data;

the client decrypts the encrypted purchase conversion model by using a preset private key corresponding to the preset public key to obtain the purchase conversion model;

the client sending the purchase conversion model to the server in order for the server to predict a probability of purchasing the advertised product by a user clicking on an advertising page of the advertised product using the purchase conversion model, including: and if the server determines that the probability that the user clicking the advertisement page of the advertisement product purchases the advertisement product is greater than or equal to a preset threshold value, taking the user as a target user of the advertisement product.

5. A data processing apparatus, characterized in that the apparatus comprises:

the system comprises a receiving and sending module, a sending and receiving module and a sending and receiving module, wherein the receiving and sending module is used for receiving a request message sent by a client, and the request message comprises a data range;

the processing module is used for acquiring click data meeting the data range according to the request message, wherein the click data is data for clicking the advertisement page of the advertisement product by a user;

the receiving and sending module is further configured to send the click data to the client, so that the client obtains conversion data matched with the click data according to the click data, and receives first encrypted data sent by the client, where the conversion data is data for the user to purchase the advertisement product after entering a purchase page through the advertisement page, and the first encrypted data is data obtained by encrypting the conversion data by using a preset public key by the client;

the training module is used for training by using the first encrypted data and the click data to obtain an encrypted purchase conversion model;

the receiving and sending module is further configured to send the encrypted purchase conversion model to the client, and receive the purchase conversion model sent by the client, where the purchase conversion model is obtained by the client decrypting the encrypted purchase conversion model by using a preset private key corresponding to the preset public key;

and the prediction module is used for predicting the probability of the users clicking the advertisement pages of the advertisement products for purchasing the advertisement products by using the purchase conversion model, and if the probability of the users clicking the advertisement pages of the advertisement products for purchasing the advertisement products is determined to be greater than or equal to a preset threshold value, the users are used as target users of the advertisement products.

6. A data processing apparatus, characterized in that the apparatus comprises:

the system comprises a receiving and sending module, a sending and receiving module and a sending and receiving module, wherein the receiving and sending module is used for sending a request message to a server, the request message comprises a data range, the request message is used for the server to obtain click data meeting the data range, and the click data is data for clicking an advertisement page of an advertisement product by a user; receiving the click data sent by the server;

the encryption module is used for acquiring conversion data matched with the click data according to the click data and encrypting the conversion data by using a preset public key to obtain first encrypted data; wherein the conversion data is data of purchasing the advertisement product after the user enters a purchase page through the advertisement page;

the transceiver module is further configured to send the first encrypted data to a server, and receive an encrypted purchase conversion model sent by the server, where the encrypted purchase conversion model is obtained by the server through training using the first encrypted data and the click data;

the decryption module is used for decrypting the encrypted purchase conversion model by using a preset private key corresponding to the preset public key to obtain the purchase conversion model;

the transceiver module is further configured to send the purchase conversion model to the server, so that the server predicts a probability that a user who clicks an advertisement page of the advertisement product purchases the advertisement product using the purchase conversion model, and if it is determined that the probability that the user who clicks the advertisement page of the advertisement product purchases the advertisement product is greater than or equal to a preset threshold, the user is taken as a target user of the advertisement product.

7. A computing device comprising at least one processing unit and at least one memory unit, wherein the memory unit stores a computer program that, when executed by the processing unit, causes the processing unit to perform the method of any of claims 1 to 4.

8. A computer-readable storage medium, storing a computer program executable by a computing device, the program, when executed on the computing device, causing the computing device to perform the method of any one of claims 1 to 4.

Images