CN115329864A

CN115329864A - Method and device for training recommendation model and electronic equipment

Info

Publication number: CN115329864A
Application number: CN202210967132.6A
Authority: CN
Inventors: 李航; 张德; 刘文炎; 万俊成
Original assignee: Beijing Youzhuju Network Technology Co Ltd
Current assignee: Beijing Youzhuju Network Technology Co Ltd
Priority date: 2022-08-11
Filing date: 2022-08-11
Publication date: 2022-11-11

Abstract

The embodiment of the disclosure provides a method and a device for training a recommendation model and electronic equipment. The method may include determining data requested to be deleted in the original data in response to receiving a data deletion request from a user. The method may further include retrieving remaining data for the user based on the data requested to be deleted. Additionally, the method may further include training the recommendation model using the residual data. By the technical scheme of the embodiment of the disclosure, longer model training time and a large amount of calculation overhead generated by retraining the recommended model can be avoided, and the time for updating the model is shortened. In addition, because the training of the recommendation model forgets the data which the user desires to delete, the performance of the recommendation model is improved, and the recommendation result is easier to hit the preference of the user, so that the user experience is improved.

Description

Method and device for training recommendation model and electronic equipment

Technical Field

Embodiments of the present disclosure relate to the field of data processing, and more particularly, to a method, an apparatus, and an electronic device for training a recommendation model.

Background

The recommendation model relies on machine learning techniques such as deep neural networks to simulate complex interactions between a user and items of interest (e.g., products, videos, movies, news stories, etc.) so that the user's items of interest can be predicted. However, it may be necessary or desirable to intentionally forget some training data for the recommendation model. When there is a need for deletion of part of the historical behavior data of the user from the recommendation system, conventional recommendation systems typically retrain the recommendation model based on the data remaining after deletion, which results in a large amount of computational overhead and long model training time.

Disclosure of Invention

The embodiment of the disclosure provides a method and a device for training a recommendation model and electronic equipment.

In a first aspect, an embodiment of the present disclosure provides a method for training a recommendation model. The method may include determining data requested to be deleted in the original data in response to receiving a data deletion request from a user. The method may further include retrieving remaining data for the user based on the data requested to be deleted. Additionally, the method may further include training the recommendation model using the residual data.

In a second aspect, an embodiment of the present disclosure provides an apparatus for training a recommendation model, where the apparatus may include: the data deleting module is configured to respond to a data deleting request of a user and determine the data requested to be deleted in the original data; a remaining data acquisition module configured to acquire remaining data for the user based on the data requested to be deleted; and a recommendation model training module configured to train the recommendation model using the residual data.

In a third aspect, an embodiment of the present disclosure provides an electronic device, including: a processor; and a memory coupled with the processor, the memory having instructions stored therein that, when executed by the processor, cause the electronic device to perform acts comprising: in response to receiving a data deletion request of a user, determining data requested to be deleted in original data; obtaining remaining data for the user based on the data requested to be deleted; and training the recommendation model using the residual data.

In a fourth aspect, the disclosed embodiments provide a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs any of the steps of the method according to the first aspect.

By the technical scheme of the embodiment of the disclosure, longer model training time and a large amount of calculation overhead generated by retraining the recommended model can be avoided, and the time for updating the model is shortened. In addition, because the training of the recommendation model forgets the data which the user desires to delete, the performance of the recommendation model is improved, and the recommendation result is easier to hit the preference of the user, so that the user experience is improved.

This summary is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. This summary is not intended to identify key features or essential features of the disclosure, nor is it intended to be used to limit the scope of the disclosure.

Drawings

The foregoing and other objects, features and advantages of the disclosure will be apparent from the following more particular descriptions of exemplary embodiments of the disclosure as illustrated in the accompanying drawings wherein like reference numbers generally represent the same or similar parts throughout the exemplary embodiments of the disclosure. In the drawings:

FIG. 1 illustrates a schematic diagram of an example environment in which embodiments of the present disclosure can be implemented;

FIG. 2 shows a schematic diagram of a detailed example environment for training and applying models, in accordance with embodiments of the present disclosure;

FIG. 3 shows a flow diagram of a process for training a recommendation model, in accordance with an embodiment of the present disclosure;

FIG. 4 shows a schematic diagram of raw data of multiple users maintained at the server side, according to an embodiment of the present disclosure;

FIG. 5 shows a flowchart of a process for training a recommendation model, in accordance with an embodiment of the present disclosure;

FIG. 6 shows a schematic diagram of an apparatus for training a recommendation model, in accordance with an embodiment of the present disclosure; and

FIG. 7 shows a schematic block diagram of an example device that may be used to implement embodiments of the present disclosure.

Detailed Description

It is understood that before the technical solutions disclosed in the embodiments of the present disclosure are used, the type, the use range, the use scene, etc. of the personal information related to the present disclosure should be informed to the user and obtain the authorization of the user through a proper manner according to the relevant laws and regulations.

For example, in response to receiving a user's active request, prompt information is sent to the user to explicitly prompt the user that the requested operation to be performed would require acquisition and use of personal information to the user. Thus, the user can autonomously select whether to provide personal information to software or hardware such as an electronic device, an application program, a server, or a storage medium that performs the operations of the disclosed technical solution, according to the prompt information.

As an optional but non-limiting implementation manner, in response to receiving an active request from the user, the manner of sending the prompt information to the user may be, for example, a pop-up window, and the prompt information may be presented in a text manner in the pop-up window. In addition, a selection control for providing personal information to the electronic device by the user's selection of "agreeing" or "disagreeing" can be carried in the pop-up window.

It is understood that the above notification and user authorization process is only illustrative and not limiting, and other ways of satisfying relevant laws and regulations may be applied to the implementation of the present disclosure.

It will be appreciated that the data involved in the subject technology, including but not limited to the data itself, the acquisition or use of the data, should comply with the requirements of the corresponding laws and regulations and related regulations.

The principles of the present disclosure will be described below with reference to a number of example embodiments shown in the drawings.

In describing embodiments of the present disclosure, the terms "include" and its derivatives should be interpreted as being inclusive, i.e., "including but not limited to. The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first," "second," and the like may refer to different or the same object. Other explicit and implicit definitions are also possible below.

In the description of the embodiments of the present disclosure, the term "raw data" may refer to historical data related to a user, which is stored on the server side, including but not limited to historical behavior data generated when the user uses an Application (APP) corresponding to the server. Furthermore, the term "feature data" generally refers to features that a machine learning module extracts from data.

As described above, with the continuous development of computer technology, machine learning technology is widely applied to various aspects of people's life. In order to better perform the recommendation task, the training process of the conventional recommendation model needs to be optimized. In conventional recommendation systems, there is often a need for controlled forgetting of user data maintained in the system, with motivation in two ways. In a first aspect, the recommendation model may reveal historical behavior information or other information generated by the user when browsing a web page or using APP. Therefore, users desire to delete at least a portion of this information in order to circumvent this privacy risk. In a second aspect, the performance of the recommended model may degrade rapidly due to the presence of noise in the training. By way of example, the training data may contain outdated instances, outliers, or instances contaminated by virus attacks. Therefore, the user desires to remove the effects of these data from the trained model to improve the experience of the recommendation system.

For this reason, conventional recommendation systems typically retrain the recommendation model based on the data remaining after deletion. However, this approach is very time consuming and results in a large computational overhead, which is not practical for a large-scale recommendation system in real life.

In addition, another conventional recommendation, recEraser, attempts to partition the training data into disjoint subsets during the initial training process. After data slicing, one submodel may be trained for each sliced data, and parameters of the ensemble model may be determined based on these submodels. When there is data that needs to be deleted, the RecEraser system may disregard the slice containing the data to be deleted when determining the parameters of the overall model. A disadvantage of this system is that the recommendation accuracy of the RecEraser system is degraded in case only a single or small amount of data needs to be deleted, whereas the precise algorithm of the RecEraser system may be forced to degrade to the above described retraining based on all remaining data in case data that does not arrive sequentially needs to be deleted. Therefore, the RecEraser system is not widely applicable to various application scenarios of the recommendation system.

According to the embodiment of the disclosure, a recommendation model training scheme based on a user deletion request is provided. The approach, upon receiving an indication or request from one or more users to delete private or noisy data, first trains or updates the recommendation model with the remaining data of the one or more users (i.e., data other than the data requiring deletion). Therefore, the method and the device avoid longer model training time and a large amount of calculation overhead generated by retraining the recommendation model, remarkably shorten the time for updating the model, and can provide an updated recommendation result for a user requesting to delete data in time, thereby solving the problems and/or other potential problems.

Embodiments of the present disclosure will be described in detail below in conjunction with example scenarios. It should be understood that this is done for illustrative purposes only and is not intended to limit the scope of the present disclosure in any way.

FIG. 1 shows a block diagram of an example system 100 for training a recommendation model in accordance with an embodiment of the present disclosure. It should be understood that the system 100 shown in fig. 1 is merely one example in which embodiments of the present disclosure may be implemented and is not intended to limit the scope of the present disclosure. Embodiments of the present disclosure are equally applicable to other systems or architectures.

As shown in fig. 1, system 100 may include a computing device 120 located on the server side. Computing device 120 may be configured to receive user data from client 110. Further, the computing device 120 may determine, through the recommendation model 130 disposed therein, recommendations that may be of interest to the user of the client 110, and feed back the recommendations to the client 110.

In some embodiments, the user data obtained by the computing device 120 may be historical behavior data generated by the user while browsing a web page or using APP, and the recommendation model 130 may determine recommendations for which the user may be interested based on a large amount of user data (a sufficient amount of user data for the user is maintained on the server side).

In the present disclosure, recommendation model 130 may be designed to perform recommendation tasks. As an example, a video recommendation model corresponding to video APP may be used to recommend videos of interest to a user based on the user's video viewing history data. Examples of recommendation models include, but are not limited to, deep neural networks of various types (DNNs), convolutional Neural Networks (CNNs), support Vector Machines (SVMs), decision trees, random forest models, and so forth. In implementations of the present disclosure, the recommendation model may also be referred to as "neural network", "learning model", "learning network", "model", and "network" for alternative use.

In some embodiments, computing device 120 may include, but is not limited to, a personal computer, a server computer, a handheld or laptop device, a mobile device (such as a mobile phone, a personal digital assistant, PDA, a media player, etc.), a consumer electronics product, a minicomputer, a mainframe computer, a cloud computing resource, and the like.

It should be understood that the devices and/or elements of the devices included in the system 100 are merely exemplary and are not intended to limit the scope of the present disclosure. It should be understood that system 100 may also include additional devices and/or units not shown. For example, in some embodiments, the computing device 120 of the system 100 may further include a storage unit (not shown) for storing the pre-input hyper-parameters and the like.

The training and use of the model in the computing device 120 will be described below with reference to fig. 2.

Fig. 2 shows a schematic diagram of a detailed example environment 200, according to an embodiment of the present disclosure. Similar to fig. 1, the example environment 200 may contain a computing device 220, user data 210 input to the computing device 220, and recommendation results 230 output from the computing device 220. The difference is that the example environment 200 may generally include a model training system 260 and a model application system 270. By way of example, model training system 260 and/or model application system 270 may be implemented in computing device 120 as shown in FIG. 1 or computing device 220 as shown in FIG. 2. It should be understood that the description of the structure and functionality of the example environment 200 is for exemplary purposes only and is not intended to limit the scope of the subject matter described herein. The subject matter described herein may be implemented in various structures and/or functions.

As previously described, the process of processing user data from the client 110 to determine recognition results such as items of interest to the user may be divided into two phases: a model training phase and a model application phase. As an example, as shown in FIG. 2, in the model training phase, the model training system 260 may train the recommendation model 240 using the user data set 250. It should be appreciated that the user data set 250 may be historical behavior data generated by a large number of users during use of an APP or browsing a web page. In the model application phase, the model application system 270 may receive the trained recommendation model 240. Thus, the recommendation model 240 loaded into the computing device 220 of the model application system 270 may determine the recommendation 230 based on the user data 210.

In other embodiments, the recommendation model 240 may be constructed as a learning network. In some embodiments, the learning network may include a plurality of networks, where each network may be a multi-layer neural network, which may be composed of a large number of neurons. Through the training process, the respective parameters of the neurons in each network can be determined. The parameters of the neurons in these networks are collectively referred to as parameters of the recommendation model 240.

The training process of the recommendation model 240 may be performed in an iterative manner until at least some of the parameters of the recommendation model 240 converge or until a predetermined number of iterations is reached, thereby obtaining final model parameters.

The technical solutions described above are only for example and do not limit the present disclosure. It should be understood that the various networks may also be arranged in other ways and connections. To explain the principle of the above scheme more clearly, the process of recommendation model training will be described in more detail below with reference to fig. 3.

FIG. 3 shows a flow diagram of a process 300 for training a recommendation model, according to an embodiment of the present disclosure. In some embodiments, process 300 may be implemented in computing device 120 in fig. 1 and computing device 220 in fig. 2. A process 300 of training a recommendation model in accordance with an embodiment of the present disclosure is now described with reference to fig. 3. For ease of understanding, the specific examples set forth in the following description are intended to be illustrative, and are not intended to limit the scope of the disclosure.

In step 302, computing device 120 may determine, in response to receiving a data deletion request from a user, data in the original data that was requested to be deleted. In some embodiments, the data corresponding to the data deletion request is historical data or noise data generated by the user while browsing a web page or using an application.

In order to clearly describe the determination process of "data requested to be deleted in original data" referred to in the present disclosure, this will now be described in detail with reference to fig. 4. FIG. 4 shows a schematic diagram of raw data 400 for multiple users maintained at the server side, according to an embodiment of the disclosure. As shown in fig. 4, the server side may generally maintain or maintain user data 410, user data 420, user data 430, user data 440, user data 450, user data 460, and each user data corresponds to only one user. It should be understood that the particular number of users shown in fig. 4 is merely exemplary, and that the server side may maintain user data for more users.

When a user corresponding to the user data 410 sends a data deletion request to the server side through a given interface such as APP, the computing device 120 on the server side may determine which of the user's original data is the data requested to be deleted based on the data deletion request. In fig. 4, computing device 120 may determine, based on the data deletion request, that data 411 is the data requested to be deleted.

It should be appreciated that in the field of video recommendation, the raw data described herein may generally include label information for a plurality of users and videos that the users each have viewed, and the video recommendation model to be trained is configured to determine other videos associated with the label information that are not viewed by at least one of the users. It should also be understood that in the field of news message recommendation, the raw data described herein may generally include tag information for multiple users and news messages that the users each read, and the news message recommendation model to be trained is configured to determine other news messages associated with the tag information that at least one of the users did not read. It should also be understood that in the field of product recommendation, the raw data described herein may generally include label information for a plurality of users and products that the users have each searched, and the product recommendation model to be trained is configured to determine other products associated with the label information that at least one of the users has not searched.

Returning to FIG. 3, at step 304, computing device 120 may retrieve the remaining data for the user based on the data requested to be deleted. In other words, in fig. 4, computing device 120 may obtain remaining data 412 for user data 410 based on data 411 that is requested to be deleted. In some embodiments, to retrieve the remaining data 412, the computing device 120 may determine the remaining data 412 by removing the data 411 that was requested to be deleted from the user data 410. In some embodiments, when the user indicates to delete all of the user data 410, the computing device 120 may determine the remaining data 412 by random initialization. In this way, the influence of the data requested to be deleted on the model can be eliminated as much as possible in the subsequent training of the recommendation model.

Returning again to FIG. 3, at step 306, computing device 120 may train recommendation model 130 using residual data 412. In some embodiments, to train the recommendation model 130, the computing device 120 may obtain parameters of the original recommendation model determined by training with the raw data 400 and determine the parameters as initial parameters of the recommendation model 130. In turn, the computing device 120 may update the initial parameters of the recommendation model 130 with the remaining data 412. By way of example, the residual data 412 may be utilized to train the recommendation model 130 for a round. In this way, the training process of the recommendation model 130 can be hot-started by using the parameters of the original recommendation model, so that the training time is shortened, and since the residual data 412 better reflects the preference of the user corresponding to the user data 410, one or more rounds of training on the recommendation model 130 directly using the residual data 412 can quickly and effectively adjust the parameters of the model. In addition, because the training of the recommendation model forgets the data which the user desires to delete, the recommendation result of the recommendation model is easier to hit the preference of the user, and therefore the user experience is improved.

In some embodiments, process 300 may also include operations, for example, computing device 120 may obtain other data from the raw data for users other than the user requesting the deletion of data described above, and use the other data to train recommendation model 130 that has been trained using the remaining data. In other words, as shown in FIG. 4, the computing device 120 may obtain other data 413 from the raw data 400 and train the recommendation model 130 that has been trained using the remaining data 412 using the other data 413. In this way, since the other data 413 is training data once used for training the original recommendation model, it is more stable and facilitates convergence of the recommendation model 130, thereby shortening the training time of the recommendation model 130.

In some embodiments, the raw data includes at least a plurality of users and corresponding items of interest. As an example, in fig. 4, user data 410 in raw data 400 may contain the corresponding user and their items of interest, and so on. The term "item of interest" may refer to a personal preference embodied in the user's historical behavioral data. As an example, in an application scenario of video recommendation, when a user frequently watches videos related to time administration, an item of interest of the user may be determined as time administration. Of course the items of interest may have a finer classification. As an example, in an application scenario of product recommendation, when a user frequently searches for "leather shoe No. 43 _ man" in APP, its interested items may be determined as "man's shoe", "leather shoe", "43", and the like.

FIG. 5 shows a flow diagram of a portion of a detailed process 500 for training a recommendation model, according to an embodiment of the present disclosure. In some embodiments, process 500 may be implemented in computing device 120 in fig. 1 and computing device 220 in fig. 2. A partial detailed process 500 of training a recommendation model according to an embodiment of the present disclosure is now described with reference to fig. 5. For ease of understanding, the specific examples set forth in the following description are intended to be illustrative, and are not intended to limit the scope of the disclosure.

At step 501, the computing device 120 may leave unchanged at least a first feature data in the recommendation model 130 that corresponds to an item of interest of a group of users of the plurality of users. In some embodiments, the computing device 120 may fix the feature data of the item of interest in the recommendation model 130 and other model parameters in the recommendation model 130 in addition to the feature data of the user and the item of interest. It should be understood that the group of users may be one user or a small group of users.

At step 502, the computing device 120 may update the second feature data for the set of users to minimize the loss function value of the recommendation model 130. In some embodiments, the computing device 120 may update the feature data for the set of users with a second order optimizer to minimize the loss of residual data by the recommendation model 130.

At step 503, computing device 120 may determine whether all users have been traversed, and if not, continue with step 501, and if so, proceed to step 504.

At step 504, the computing device 120 may leave unchanged at least the second characteristic data for the set of users in the recommendation model 130. In some embodiments, the computing device 120 may fix the feature data of the user in the recommendation model 130 and other model parameters in the recommendation model 130 in addition to the feature data of the user and the item of interest.

At step 505, the computing device 120 may update the first characterization data to minimize the loss function value. In some embodiments, the computing device 120 may update the feature data for the item of interest with a second order optimizer to minimize the loss of residual data by the recommendation model 130.

At step 506, computing device 120 may determine whether all items of interest have been traversed, and if not, continue with step 504, and if so, proceed to step 507.

At step 507, computing device 120 may return feature data for the user and the item of interest, thereby training recommendation model 130.

Therefore, the model training process can be accelerated in the alternating optimization mode, and the model training time is shortened. In addition, a second-order optimizer is utilized in the model training mode, and the second-order optimizer is combined with the hot start embodiment to improve the efficiency and effect of model training more remarkably.

With the above embodiments, the present disclosure avoids the long model training time and large computational overhead resulting from retraining the recommended model (forgetting the model). In addition, the training process of the recommended model can be remarkably accelerated by only utilizing the residual data in cooperation with the training mode of hot start, so that the training time of the model can be shortened and the performance of the model can be improved under the condition of utilizing other data in cooperation with the alternate optimization and the second-order optimization mode. In addition, because the training of the recommendation model forgets the data which the user desires to delete, the performance of the recommendation model is improved, and the recommendation result is easier to hit the preference of the user.

The present disclosure also provides a model training device. In particular, fig. 6 shows a schematic diagram of an apparatus 600 for training a recommendation model according to an embodiment of the present disclosure. As shown in fig. 6, the apparatus 600 may include at least a deleted data determination module 602, a remaining data acquisition module 604, and a recommendation model training module 606. The deleted data determination module 602 may determine the data requested to be deleted in the original data upon receiving a data deletion request from a user. The remaining data acquisition module 604 may acquire the remaining data for the user based on the data requested to be deleted. Further, the recommendation model training module 606 may train the recommendation model using the remaining data.

In some embodiments, recommendation model training module 606 may include: a raw parameter acquisition module configured to acquire parameters of a raw recommendation model determined by training using raw data; an initial parameter determination module configured to determine a parameter as an initial parameter of the recommendation model; and a recommendation model update module configured to update initial parameters of the recommendation model with the remaining data.

In some embodiments, the apparatus 600 may further comprise: a data acquisition module configured to acquire data of a user other than the user from the raw data; and a recommendation module retraining module configured to train a recommendation model trained using the remaining data using the other data.

In some embodiments, the remaining data acquisition module 604 may be configured to determine the remaining data by removing the data requested to be deleted from the original data.

In some embodiments, the raw data may include at least a plurality of users and corresponding items of interest, and recommendation model training module 606 may be configured to: maintaining unchanged at least feature data in the recommendation model corresponding to items of interest of a group of users of the plurality of users; and updating the feature data of the set of users to minimize the loss function value of the recommended model.

In some embodiments, the recommendation model training module 604 may be further configured to: keeping unchanged feature data of at least a group of users in the recommendation model; and updating the feature data corresponding to the items of interest for the group of users to minimize the loss function values.

In some embodiments, updating the feature data for the set of users and updating the feature data corresponding to the item of interest for the set of users may both be performed using a second order optimizer.

In some embodiments, the data corresponding to the data deletion request may be historical data or noise data generated by the user while browsing a web page or using an application.

In some embodiments, the raw data may include tag information for videos that a plurality of users each viewed, and the recommendation model is configured to determine other videos associated with the tag information that are not viewed by at least one of the plurality of users.

In some embodiments, the raw data may include tag information for items of interest that were viewed or accessed by each of the plurality of users, and the recommendation model is configured to determine other items of interest associated with the tag information that were not viewed or accessed by at least one of the plurality of users.

In some embodiments, the items of interest may be products, subject articles, pictures, self media accounts, and the like. The product may correspond to, for example, a page of some item displayed on a shopping website, etc. The subject article may correspond to, for example, news, essay, novel works, and the like. The pictures may correspond to, for example, electronic photographs, screenshots, vector graphics, and the like.

Fig. 7 illustrates a schematic block diagram of an example device 700 that may be used to implement embodiments of the present disclosure. For example, computing device 120 as shown in FIG. 1 and computing device 220 as shown in FIG. 2 may be implemented by device 700. As shown, device 700 includes a Central Processing Unit (CPU) 701 that may perform various appropriate actions and processes in accordance with computer program instructions stored in a Read Only Memory (ROM) 702 or computer program instructions loaded from a storage unit 708 into a Random Access Memory (RAM) 703. In the RAM703, various programs and data required for the operation of the device 700 can be stored. The CPU 701, ROM702, and RAM703 are connected to each other via a bus 704. An input/output (I/O) interface 705 is also connected to bus 704.

Various components in the device 700 are connected to the I/O interface 705, including: an input unit 706 such as a keyboard, a mouse, or the like; an output unit 707 such as various types of displays, speakers, and the like; a storage unit 708 such as a magnetic disk, optical disk, or the like; and a communication unit 709 such as a network card, modem, wireless communication transceiver, etc. The communication unit 709 allows the device 700 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks. It is to be understood that the present disclosure may display, with the output unit 707, real-time dynamic change information of user satisfaction, key factor identification information of group users or individual users of satisfaction, optimization policy information, policy implementation effect evaluation information, and the like.

The processing unit 701 may be implemented by one or more processing circuits. The processing unit 701 may be configured to perform the various processes and processes described above, such as the

processes

300, 500. For example, in some embodiments, the

processes

300, 500 may be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 708. In some embodiments, part or all of a computer program may be loaded onto and/or installed onto device 700 via ROM702 and/or communications unit 709. When the computer program is loaded into RAM703 and executed by CPU 701, one or more of the steps of

processes

300, 500 described above may be performed.

The present disclosure may be systems, methods, and/or computer program products. The computer program product may include a computer-readable storage medium having computer-readable program instructions embodied thereon for carrying out various aspects of the present disclosure.

The computer readable storage medium may be a tangible device that can hold and store the instructions for use by the instruction execution device. The computer readable storage medium may be, for example, but not limited to, an electronic memory device, a magnetic memory device, an optical memory device, an electromagnetic memory device, a semiconductor memory device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a Static Random Access Memory (SRAM), a portable compact disc read-only memory (CD-ROM), a Digital Versatile Disc (DVD), a memory stick, a floppy disk, a mechanical coding device, such as punch cards or in-groove projection structures having instructions stored thereon, and any suitable combination of the foregoing. Computer-readable storage media as used herein is not to be interpreted as a transitory signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., optical pulses through a fiber optic cable), or an electrical signal transmitted through an electrical wire.

The computer-readable program instructions described herein may be downloaded from a computer-readable storage medium to a respective computing/processing device, or to an external computer or external storage device via a network, such as the internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers and/or edge servers. The network adapter card or network interface in each computing/processing device receives computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in the respective computing/processing device.

The computer program instructions for carrying out operations of the present disclosure may be assembler instructions, instruction Set Architecture (ISA) instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider). In some embodiments, the electronic circuitry that can execute the computer-readable program instructions implements aspects of the present disclosure by utilizing the state information of the computer-readable program instructions to personalize the electronic circuitry, such as a programmable logic circuit, a Field Programmable Gate Array (FPGA), or a Programmable Logic Array (PLA).

Various aspects of the present disclosure are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the disclosure. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer-readable program instructions.

These computer-readable program instructions may be provided to a processing unit of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processing unit of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks. These computer-readable program instructions may also be stored in a computer-readable storage medium that can direct a computer, programmable data processing apparatus, and/or other devices to function in a particular manner, such that the computer-readable medium storing the instructions comprises an article of manufacture including instructions which implement the function/act specified in the flowchart and/or block diagram block or blocks.

The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer, other programmable apparatus or other devices implement the functions/acts specified in the flowchart and/or block diagram block or blocks.

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

According to one or more embodiments of the present disclosure. Example 1. A method of training a recommendation model, comprising: in response to receiving a data deletion request of a user, determining data requested to be deleted in original data; obtaining remaining data for the user based on the data requested to be deleted; and training the recommendation model using the residual data.

Example 2. The method of example 1, wherein training the recommendation model comprises: acquiring parameters of an original recommendation model determined by training of the original data; determining the parameters as initial parameters of the recommendation model; and updating the initial parameters of the recommendation model with the residual data.

Example 3. The method of example 1, further comprising: acquiring other data of other users except the user from the original data; and training the recommendation model trained using the remaining data using the other data.

Example 4. The method of example 1, wherein obtaining the remaining data comprises: determining the remaining data by removing the data requested to be deleted from the original data.

Example 5. The method of any of examples 1-4, wherein the raw data includes at least a plurality of users and respective items of interest, and wherein training the recommendation model includes: keeping at least first feature data in the recommendation model unchanged, the first feature data corresponding to an item of interest of a group of users in the plurality of users; and updating the second feature data of the set of users to minimize a loss function value of the recommendation model.

Example 6. The method of example 5, wherein training the recommendation model further comprises: keeping at least second feature data in the recommendation model unchanged; and updating the first characteristic data to minimize the loss function value.

Example 7. The method of example 6, wherein updating the feature data for the set of users and updating the feature data corresponding to the item of interest for the set of users are both performed using a second order optimizer.

Example 8. The method of any of examples 1-4, wherein the data to which the data deletion request corresponds is historical data or noise data generated by the user while browsing a web page or using an application.

Example 9. The method of any of examples 1-4, wherein the raw data includes tag information for a plurality of users and videos that the plurality of users each viewed, and the recommendation model is configured to determine other videos associated with the tag information that are not viewed by at least one of the plurality of users.

Example 10. The method of any of examples 1-4, wherein the raw data includes tag information for a plurality of users and items of interest that the plurality of users each viewed or accessed, and the recommendation model is configured to determine other items of interest associated with the tag information that at least one of the plurality of users did not view or accessed.

Example 11. The method of example 10, wherein the item of interest comprises at least one of: producing a product; a subject article; a picture; and a self media account number.

Example 12. An apparatus to train a recommendation model, in accordance with one or more embodiments of the present disclosure, includes: a deleted data determination module configured to determine data requested to be deleted in the original data in response to receiving a data deletion request of a user; a remaining data acquisition module configured to acquire remaining data for the user based on the data requested to be deleted; and a recommendation model training module configured to train the recommendation model using the residual data.

Example 13. The apparatus of example 12, wherein the recommendation model training module comprises: the original parameter acquisition module is configured to acquire parameters of an original recommendation model determined by training of the original data; an initial parameter determination module configured to determine the parameter as an initial parameter of the recommendation model; and a recommendation model update module configured to update the initial parameters of the recommendation model with the remaining data.

Example 14. The apparatus of example 12, further comprising: a data acquisition module configured to acquire data of a user other than the user from the raw data; and a recommendation modularity retraining module configured to train the recommendation model trained using the remaining data using the other data.

Example 15 the apparatus of example 12, wherein the remaining data acquisition module is configured to determine the remaining data by removing the data requested to be deleted from the original data.

Example 16. The apparatus of any of examples 12-15, wherein the raw data includes at least a plurality of users and respective items of interest, and wherein the recommendation model training module is configured to: keeping at least first feature data in the recommendation model unchanged, the first feature data corresponding to an item of interest of a group of users in the plurality of users; and updating the second feature data of the set of users to minimize a loss function value of the recommendation model.

The apparatus of example 17. The apparatus of example 16, wherein the recommendation model training module is further configured to: keeping at least second feature data in the recommendation model unchanged; and updating the first characteristic data to minimize the loss function value.

Example 18. The apparatus of example 17, wherein updating the feature data for the set of users and updating the feature data corresponding to the item of interest for the set of users are each performed using a second order optimizer.

Example 19. The apparatus of any of examples 12-15, wherein the data to which the data deletion request corresponds is historical data or noise data generated by the user while browsing a web page or using an application.

Example 20. The apparatus of any of examples 12-15, wherein the raw data includes tag information for a plurality of users and videos that the plurality of users each viewed, and the recommendation model is configured to determine other videos associated with the tag information that are not viewed by at least one of the plurality of users.

Example 21. The apparatus of any of examples 12-15, wherein the raw data includes tag information for a plurality of users and items of interest that the plurality of users each viewed or accessed, and the recommendation model is configured to determine other items of interest associated with the tag information that at least one of the plurality of users did not view or accessed.

Example 22. The apparatus of example 21, wherein the item of interest comprises at least one of: producing a product; a subject article; a picture; and a self media account number.

Example 23, according to one or more embodiments of the present disclosure, an electronic device, comprising: a processor; and a memory coupled with the processor, the memory having instructions stored therein that, when executed by the processor, cause the electronic device to perform acts comprising: in response to receiving a data deletion request of a user, determining data requested to be deleted in original data; obtaining remaining data for the user based on the data requested to be deleted; and training the recommendation model using the residual data.

Example 24. The apparatus of example 23, wherein training the recommendation model comprises: acquiring parameters of an original recommendation model determined by training of the original data; determining the parameters as initial parameters of the recommended model; and updating the initial parameters of the recommendation model with the residual data.

Example 25. The apparatus of example 23, further comprising: acquiring other data of other users except the user from the original data; and training the recommendation model trained using the remaining data using the other data.

Example 26. The device of example 23, wherein obtaining the remaining data comprises: determining the remaining data by removing the data requested to be deleted from the original data.

Example 27. The apparatus of any of examples 23-26, wherein the raw data includes at least a plurality of users and respective items of interest, and wherein training the recommendation model includes: maintaining unchanged at least feature data in the recommendation model corresponding to items of interest for a group of users in the plurality of users; and updating the feature data of the set of users to minimize the loss function value of the recommendation model.

Example 28. The apparatus of example 27, wherein training the recommendation model further comprises: keeping unchanged at least the feature data of the set of users in the recommendation model; and updating feature data corresponding to the set of user's items of interest to minimize the loss function value.

Example 29 the apparatus of example 28, wherein updating the feature data for the group of users and updating the feature data corresponding to the item of interest for the group of users are both performed using a second order optimizer.

Example 30. The device of any of examples 23-26, wherein the data to which the data deletion request corresponds is historical data or noise data generated by the user while browsing a web page or using an application.

Example 31. The apparatus of any of examples 23-26, wherein the raw data comprises tag information for a plurality of users and videos that the plurality of users each viewed, and the recommendation model is configured to determine other videos associated with the tag information that are not viewed by at least one of the plurality of users.

Example 32. The apparatus of any of examples 23-26, wherein the raw data includes tag information for a plurality of users and items of interest that the plurality of users each viewed or accessed, and the recommendation model is configured to determine other items of interest associated with the tag information that at least one of the plurality of users did not view or accessed.

Example 33. The device of example 32, wherein the item of interest comprises at least one of: producing a product; a subject article; a picture; and a self media account number.

Example 34. A computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method of any of examples 1-11.

Having described embodiments of the present disclosure, the foregoing description is intended to be exemplary, not exhaustive, and not limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein was chosen in order to best explain the principles of the embodiments, the practical application, or technical improvements to the technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A method of training a recommendation model, comprising:

in response to receiving a data deletion request of a user, determining data requested to be deleted in original data;

obtaining remaining data for the user based on the data requested to be deleted; and

training the recommendation model using the residual data.

2. The method of claim 1, wherein training the recommendation model comprises:

acquiring parameters of an original recommendation model determined by training of the original data;

determining the parameters as initial parameters of the recommendation model; and

updating the initial parameters of the recommendation model with the residual data.

3. The method of claim 1, further comprising:

acquiring other data of other users except the user from the original data; and

training the recommendation model trained using the remaining data using the other data.

4. The method of claim 1, wherein obtaining the remaining data comprises:

removing the data requested to be deleted from the original data to determine the remaining data.

5. The method of any of claims 1-4, wherein the raw data includes at least a plurality of users and respective items of interest, and wherein training the recommendation model comprises:

keeping at least first feature data in the recommendation model unchanged, the first feature data corresponding to an item of interest of a group of users in the plurality of users; and

updating the second feature data of the set of users to minimize a loss function value of the recommendation model.

6. The method of claim 5, wherein training the recommendation model further comprises:

keeping at least the second feature data in the recommendation model unchanged; and

updating the first characteristic data to minimize the loss function value.

7. The method of claim 6, wherein updating the feature data for the set of users and updating the feature data corresponding to the item of interest for the set of users are each performed using a second order optimizer.

8. The method of any of claims 1-4, wherein the data to which the data deletion request corresponds is historical data or noise data generated by the user while browsing a web page or using an application.

9. The method of any of claims 1-4, wherein the raw data includes tag information for a plurality of users and videos that the plurality of users each viewed, and the recommendation model is configured to determine other videos associated with the tag information that are not viewed by at least one of the plurality of users.

10. The method of any of claims 1-4, wherein the raw data includes tag information for a plurality of users and items of interest that the plurality of users each viewed or accessed, and the recommendation model is configured to determine other items of interest associated with the tag information that at least one of the plurality of users did not view or accessed.

11. The method of claim 10, wherein the item of interest comprises at least one of:

producing a product;

a subject article;

a picture; and

from the media account number.

12. An apparatus to train a recommendation model, comprising:

a deleted data determination module configured to determine data requested to be deleted in the original data in response to receiving a data deletion request of a user;

a remaining data acquisition module configured to acquire remaining data for the user based on the data requested to be deleted; and

a recommendation model training module configured to train the recommendation model using the residual data.

13. An electronic device, comprising:

a processor; and

a memory coupled with the processor, the memory having instructions stored therein that, when executed by the processor, cause the electronic device to perform the method of any of claims 1-11.

14. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1-11.