Content generation system and method, equipment and storage medium

Abstract

The invention discloses a content generation system, method and device and a storage medium, and relates to the technical field of artificial intelligence, the content generation system comprises a scene function layer, an intermediate interface layer and a model layer; the scene function layer obtains a content generation request input by a user; performing request analysis on the content generation request to obtain target request data; the intermediate interface layer determines an application function interface corresponding to the target request data; the intermediate interface layer calls the content generation model of the model layer based on the application function interface to generate content, and target content is obtained.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to content generation systems, methods, devices and storage media. Background Art

[0002] With the continuous progress of science and technology and the continuous development of society, it is foreseeable that AI big models will play an increasingly important role in the future. Not only in the field of scientific research, but also in industrial applications, daily life and many other aspects, such as automation and efficiency improvement: enterprises and organizations have a growing demand for automated processes and improved work efficiency; personalized services: consumers' demand for personalized recommendations and customized services has promoted the application of AI in recommendation systems, advertising, customer service and other aspects; emerging application areas: the demand for AI technology in medical care, education, transportation and other fields is gradually increasing.

[0003] Existing artificial intelligence models have certain limitations when faced with the diverse needs of users and are unable to meet the personalized needs of users. Summary of the invention

[0004] The main purpose of this application is to provide a content generation system, method, device and storage medium, aiming to solve the technical problem that the existing artificial intelligence model cannot meet the personalized needs of users for content generation.

[0005] To achieve the above objectives, the present application proposes a content generation system, which includes: a scene function layer, an intermediate interface layer, and a model layer;

[0006] The scene function layer is used to obtain content generation requests input by users;

[0007] The scene function layer is also used to parse the content generation request and obtain the target request data;

[0008] The middle interface layer is used to determine the application function interface corresponding to the target request data;

[0009] The intermediate interface layer is also used to generate content based on the application function interface calling the content generation model of the model layer to obtain the target content.

[0010] In addition, to achieve the above purpose, the present application also proposes a content generation method, which includes:

[0011] The scene function layer obtains the content input by the user to generate a request;

[0012] The scene function layer parses the content generation request to obtain the target request data;

[0013] The intermediate interface layer determines the application function interface corresponding to the target request data;

[0014] The intermediate interface layer generates content based on the application function interface by calling the content generation model of the model layer to obtain the target content.

[0015] In addition, to achieve the above objectives, the present application also proposes a content generation device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the content generation method as described above.

[0016] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the content generation method as described above are implemented.

[0017] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps of the content generation method as described above are implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0019] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0020] Figure 1 This is a functional structure diagram of the first embodiment of the content generation system of this application;

[0021] Figure 2 This is a schematic diagram of the workflow of Example 2 of the content generation system of this application;

[0022] Figure 3 A schematic diagram of the training process of a content generation model in one implementation of the content generation system of the present application;

[0023] Figure 4 A schematic diagram of the training process of a content generation model in one implementation of the content generation system of the present application;

[0024] Figure 5 This is a schematic diagram of the workflow of Example 1 of the content generation method of the present application;

[0025] Figure 6Schematic diagram of the device structure of the hardware operating environment involved in the content generation method in the embodiment of the present application.

[0026] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0027] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0028] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0029] The embodiment of the present application proposes a content generation system, which includes three layers, namely, a scene function layer, an intermediate interface layer, and a model layer. The scene function layer obtains the content generation request input by the user, and parses the content generation request to obtain the target request data; the intermediate interface layer determines the application function interface corresponding to the target request data; based on the application function interface, the corresponding content generation model in the model layer can be called to obtain the target content.

[0030] The content generation system of the present application can provide multiple language interfaces and artistic styles, allowing users to generate artworks by inputting descriptions, thereby improving the generation efficiency of the product and meeting the personalized needs of users.

[0031] Based on this, the embodiment of the present application provides a content generation system, referring to Figure 1 , Figure 1 This is a functional structure block diagram of the first embodiment of the content generation system of this application.

[0032] like Figure 1 As shown, in this embodiment, the content generation system includes: a scene function layer, an intermediate interface layer and a model layer;

[0033] Among them, the scene function layer is used to obtain the content generation request input by the user;

[0034] The scene function layer is also used to parse the content generation request and obtain the target request data;

[0035] The middle interface layer is used to determine the application function interface corresponding to the target request data;

[0036] The intermediate interface layer is also used to generate content based on the application function interface calling the content generation model of the model layer to obtain the target content.

[0037] It should be noted that the content generation system of the embodiment of the present application can be divided into three layers, wherein the upper layer is the scene function layer, which can interact with the user to obtain the content generation request input by the user; the middle layer is the intermediate interface layer, which includes multiple application function interfaces, which can be parsed based on the content generation request input by the user to determine which application function interfaces need to be called to complete the product function request; the bottom layer is the model layer, which can include at least one AI large model. Through the application function interface in the intermediate interface layer, the call to the AI ​​large model in the model layer can be realized, thereby realizing the production of the target content.

[0038] It is understandable that the above-mentioned content generation request may include description text, description picture, target content style, etc. Among them, the description text may be obtained by voice instructions, tabs, etc. input by the user, or may be directly input by the user, and the embodiment of the present application does not limit this.

[0039] It should be understood that based on the scene function layer, the scene functions required by the user can be determined, such as painting, sculpture, digital art, etc.; based on the scene functions required by the user, the corresponding middle layer interface can be determined, thereby realizing application functions such as image generation optimization effects, automatic classification of pictures and albums, image generation of dynamic image effects, avatar generation of animation effects, and image generation divided into spatial video effects.

[0040] It can be understood that based on the determined middle-layer interface, the large model functions that need to be called can be determined, such as image generation, image super-resolution, text generation, subtitle translation and other functions.

[0041] It should be noted that when the scene function layer receives the user's content generation request, it can parse the content generation request and determine the key data related to the target content, that is, the target request data. These key data can be used to characterize the model operation that the user wants to perform. By parsing the target request data through the intermediate interface layer, the application function interface required to complete the content generation request can be determined.

[0042] In one implementation method of the embodiment of the present application, the embodiment of the present application can define only the scene function interface for the intermediate interface layer, such as defining the image generation function interface as Interface A, defining the image generation oil painting effect function interface as Interface B, defining the image generation dynamic image effect function interface as Interface C, defining the avatar generation animation effect function interface as Interface D, etc. The embodiment of the present application does not impose any restrictions on this.

[0043] It can be understood that the embodiment of the present application encapsulates several application function interfaces in the intermediate interface layer, and these function interfaces are respectively based on several content generation models in the model layer. By parsing the user's content generation request to determine the application function interface to be called, the content generation model in the bottom layer can be called based on the target request data and the application function interface, thereby realizing the generation of the target content. Through the multiple application function interfaces encapsulated in the intermediate interface layer, the personalized needs of different users can be met, making the generated art products more diversified, and improving the applicability of the content generation system of the present application.

[0044] The content generation system of the embodiment of the present application includes a scene function layer, an intermediate interface layer and a model layer; the scene function layer obtains the content generation request input by the user; the content generation request is parsed to obtain the target request data; the intermediate interface layer determines the application function interface corresponding to the target request data; the intermediate interface layer calls the content generation model of the model layer based on the application function interface to generate content and obtain the target content. Since the target request data is determined by parsing the content generation request input by the user, and the application function interface to be called is determined, the call of the corresponding content generation model in the model layer is realized, the applicability and personalization capabilities of the content generation system are improved, and the target content generated by the system can be closer to the needs of users.

[0045] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated in the following. Figure 2 ,The intermediate interface layer is used to determine the interface calling logic of the target request data;

[0046] The intermediate interface layer is also used to determine the application function interface and the interface calling sequence required to execute the content generation request according to the interface calling logic.

[0047] It is understandable that, for the user's target request data, if only one application function interface is involved, the application function interface can be directly called to implement the call to the content generation model in the model layer.

[0048] It should be noted that in actual applications, the target request data often involves multiple application function interfaces. In this case, there may be a sequence between different application function interfaces. For example, if a user wants to generate a 1920*1082 image of a specified person with an animation effect, the user can first call the Wenshengtu function interface, and use the content generation model corresponding to the Wenshengtu function interface to generate a picture containing the specified person according to the description text entered by the user; by inputting the picture containing the specified person into the Wenshengtu function interface, the content generation model corresponding to the picture generation animation picture function interface can further generate a specified person image containing an animation effect; by inputting the specified person image containing an animation effect into the picture adjustment function interface, the content generation model corresponding to the picture adjustment function interface can be adjusted based on the resolution, size, background, etc. entered by the user, until a specified person image containing an animation effect that meets the user's needs is generated.

[0049] Therefore, in one implementation method of the embodiment of the present application, when a user needs to generate an oil painting based on text, a logical mapping relationship can be defined in the application function interface in the intermediate interface layer. For example, for the text-generated image function interface, a logical mapping relationship can be set for it and the picture-generated oil painting function interface, so that the picture generated by the content generation model corresponding to the text-generated image can be used as the input of the picture generation optimization function interface, thereby generating the target content required by the user.

[0050] Specifically, the intermediate interface layer can determine the product type required to generate the target content based on the target request data, and at the same time, according to the constraint parameters input by the user in the target request data, the interface to be called and the interface call logic between each interface to be called can be determined between each application function interface based on the product type and the constraint parameters, and then the interface call sequence corresponding to the target content division is generated. In the interface call sequence, several application function interfaces can be included, and there is a call sequence between each application function interface. For adjacent interfaces in the interface call sequence, they can be made into a preceding interface and a succeeding interface, wherein the first target content generated after the preceding interface calls the content generation model will be used as the input of the succeeding interface, so that the succeeding interface can input the first target content into the content generation model corresponding to the subsequent interface to generate the second target content.

[0051] It should be noted that the above-mentioned constraint parameters may include the size of the generated image, the size of the generated sculpture, the type of generated image, etc., and the embodiments of the present application do not limit this.

[0052] It can be understood that the preceding interface is the application function interface called first among the adjacent interfaces in the interface call sequence, and the following interface is the application function interface called later among the adjacent interfaces in the interface call sequence. Specifically, the intermediate interface layer is also used to determine the application function interface required to execute the content generation request and the interface call sequence composed of the application function interfaces based on the interface call logic;

[0053] The intermediate interface layer is also used to determine adjacent interfaces in the interface call sequence as preceding interfaces and succeeding interfaces;

[0054] The intermediate interface layer is also used to determine the interface calling order in the interface calling sequence and use the first target content corresponding to the preceding interface as the input of the succeeding interface.

[0055] In the embodiment of the present application, the middle interface layer determines the interface call logic of the target request data; according to the interface call logic, the application function interface and the interface call sequence required to execute the content generation request are determined. Since the interface call logic of the target request data is determined when multiple application function interfaces need to be called, the order of the application function interfaces to be called is realized, which optimizes the execution function sequence of the system and improves the application capability of the system.

[0056] Based on the first embodiment and / or the second embodiment of the present application, in the third embodiment of the present application, the same or similar contents as those of the above-mentioned first embodiment and / or the second embodiment can refer to the above introduction, and will not be repeated later. Figure 3 In the embodiment of the present application, a plurality of content generation models are deployed in the model layer. In order to train each content generation model in the model layer, the training method of the content generation model includes:

[0057] Inputting the random vector into the generator of the content generation model to generate samples, and obtaining the target simulation sample;

[0058] Inputting the target reference sample and the target simulation sample into the discriminator of the content generation model to perform sample judgment and obtain a sample judgment result;

[0059] Perform difference analysis on sample judgment results to obtain sample prediction probability;

[0060] Adjust the parameters of the generator or discriminator based on the sample prediction probability;

[0061] When the content generation model is adjusted to a convergence condition, a trained content generation model is obtained.

[0062] It should be noted that in the content generation model of the present application, a generator and a discriminator may be included, wherein the generator is a generative neural network responsible for generating a sample close to reality from a random vector of an input (Input) and outputting it (Output) to the discriminator for judgment. The generation target of the generator is to generate a target simulation sample that can be used to deceive the discriminator, so that it cannot distinguish between the target simulation sample and the target reference sample. In the generator of the embodiment of the present application, one or more hidden layers may be included, and a nonlinear activation function may be used. The discriminator of the embodiment of the present application is a classification neural network for judging whether the input data comes from the real data distribution, and its optimization goal is to accurately distinguish the target reference sample from the target simulation sample generated by the generator. The input of the discriminator may be at least one pair of data, i.e., the target reference sample and the target simulation sample, and through the processing of the discriminator, a corresponding probability value may be output, which may be used to judge the probability that the target reference sample and the target simulation sample are reference samples.

[0063] It can be understood that the above-mentioned target reference samples are real samples used for reference, and the above-mentioned target simulation samples are false samples generated by the generator.

[0064] It should be explained that the discriminator can judge each input sample (target reference sample and target simulation sample) based on the target discriminant function of sample judgment, and output the sample probability value. When the sample probability value is greater than a preset probability threshold (such as 95%, 90%, etc., which is not limited in the embodiment of the present application), the sample judgment result can be considered to be an expected probability (which can be represented by 1), that is, the input sample is a reference sample; if the sample probability value is not greater than the preset probability threshold, the sample judgment result can be considered to be an unexpected probability (which can be represented by 0), that is, the input sample is a simulation sample. Specifically, the target reference sample and the target simulation sample are input into the discriminator of the content generation model for sample judgment, and the steps of obtaining the sample judgment result include:

[0065] Obtain target discriminant function for sample judgment;

[0066] Based on the target discriminant function, each target reference sample and target simulation sample input to the discriminator is judged to obtain a sample output probability value; the sample output probability value is used to characterize the probability that the input sample is a true sample;

[0067] The sample judgment result is determined based on the sample output probability value.

[0068] It is understandable that the target discriminant function may be a Sigmoid function or other functions, and the embodiment of the present application does not limit this.

[0069] It should be noted that the Sigmoid function is an activation function commonly used in machine learning and deep learning, and can be used for the output layer in binary classification problems. In the embodiments of the present application, the Sigmoid function is used as an example to illustrate the solution of the present application.

[0070] Specifically, the representation of the Sigmoid function can be shown as follows:

[0071]

[0072] In one implementation of the embodiment of the present application, for the sample judgment results obtained by evaluation, a difference analysis can be performed on the expected probability and the unexpected probability therein, so as to evaluate and optimize the performance of the generator and the performer, thereby improving the generation quality of the target simulation sample and the discrimination ability of the discriminator. Specifically, the step of performing a difference analysis on the sample judgment results to obtain the sample prediction probability includes:

[0073] Obtain the expected sample output probability value corresponding to the expected result and the undesired sample output probability value corresponding to the undesired result;

[0074] Determine the expected term of the function based on the expected result and the expected sample output probability value;

[0075] Determine the function's undesirable term based on the undesirable result and the undesirable sample output probability value;

[0076] The sample prediction probability is obtained according to the loss difference between the expected term of the function and the undesired term of the function. Specifically, the weight difference loss function in the embodiment of the present application can be as follows:

[0077]

[0078] in, Used to represent the loss value of the sample; y is used to represent the true label, and its value can be 0 or 1; Used to represent the output probability value obtained based on the Sigmoid function.

[0079] It should be noted that the expected term of the above function is When y is 1, that is, when the input sample is a real sample, the loss value calculation term can be converted to At this time, if the output probability value The closer it is to 1, the smaller the loss; the non-expected term of the above function is When y is 0, that is, when the input sample is not a real sample, the loss value calculation term can be converted to At this time, if the output probability value The closer it is to 0, the smaller the loss.

[0080] It can be understood that by further predicting the probability of target reference samples and content generated samples by measuring the difference loss function, the discriminator's discrimination ability can be improved.

[0081] In one implementation of the embodiment of the present application, Figure 4 As shown, in order to optimize the discriminator and the generator, in the iterative optimization process of the content generation model, the embodiment of the present application alternately fixes the first model parameters of the generator and the second model parameters of the discriminator to achieve alternating iterations of the generator and the discriminator until the content generation model reaches a convergence standard;

[0082] Specifically, the step of iteratively optimizing the generator or discriminator based on the sample prediction probability includes: when the first model parameters of the generator are fixed, iteratively optimizing the second model parameters of the discriminator based on the sample prediction probability; when the second model parameters of the discriminator are fixed, iteratively optimizing the first model parameters of the generator based on the sample prediction probability.

[0083] In the embodiment of the present application, a random vector is input into the generator of the content generation model for sample generation to obtain a target simulation sample; a target reference sample and a target simulation sample are input into the discriminator of the content generation model for sample judgment to obtain a sample judgment result; a difference analysis is performed on the sample judgment result to obtain a sample prediction probability; the generator or the discriminator is iteratively optimized based on the sample prediction probability; when the content generation model is iteratively optimized to the convergence condition, a trained content generation model is obtained. Since the training generative adversarial network is used as the content generation model of the present application, the accuracy of the model processing in the present application is improved.

[0084] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the content generation method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0085] Based on the above content generation system, the present application also provides a content generation method, such as Figure 5 As shown, in the embodiment of the present application, the method includes:

[0086] Step S10, obtaining a content generation request input by a user;

[0087] Step S20, parsing the content generation request to obtain target request data;

[0088] Step S30, determining the application function interface corresponding to the target request data;

[0089] Step S40: Based on the application function interface, the content generation model of the model layer is called to generate content to obtain target content.

[0090] The content generation method provided by this application is based on the content generation system in the above embodiment, and can solve the technical problem that the existing artificial intelligence model cannot meet the personalized needs of user content generation. Compared with the prior art, the beneficial effects of the content generation method provided by this application are the same as the beneficial effects of the content generation system provided by the above embodiment, and the other technical features in the content generation method are the same as the features disclosed by the above embodiment system, which will not be repeated here.

[0091] The present application provides a content generating device, which includes: at least one processor; and a memory that is communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the content generating method in the above-mentioned embodiment 1.

[0092] Reference below Figure 6 , which shows a schematic diagram of the structure of a content generating device suitable for implementing the embodiments of the present application. The content generating device in the embodiments of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 6 The content generating device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0093] like Figure 6As shown, the content generating device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the content generating device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the content generation device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a content generation device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0094] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0095] The content generation device provided by this application adopts the content generation method in the above embodiment, which can solve the technical problem that the existing artificial intelligence model cannot meet the personalized needs of user content generation. Compared with the prior art, the beneficial effects of the content generation device provided by this application are the same as the beneficial effects of the content generation method provided by the above embodiment, and the other technical features in the content generation device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0096] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0097] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0098] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the content generation method in the above-mentioned embodiment.

[0099] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0100] The computer-readable storage medium may be included in the content generating device, or may exist independently without being installed in the content generating device.

[0101] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the content generating device, the content generating device performs the following process:

[0102] Get the content input by the user to generate a request;

[0103] Parse the content generation request to obtain target request data;

[0104] Determine the application function interface corresponding to the target request data;

[0105] Based on the application function interface, the content generation model of the model layer is called to generate content and obtain the target content.

[0106] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a 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 may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0107] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0108] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0109] The readable storage medium provided by the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned content generation method, and can solve the technical problem that the existing artificial intelligence model cannot meet the personalized needs of user content generation. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as the beneficial effects of the content generation method provided by the above-mentioned embodiment, and will not be repeated here.

[0110] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned content generation method when executed by a processor.

[0111] The computer program product provided by this application can solve the technical problem that the existing artificial intelligence model cannot meet the personalized needs of user content generation. Compared with the prior art, the beneficial effects of the computer program product provided by this application are the same as the beneficial effects of the content generation method provided by the above embodiment, which will not be repeated here.

[0112] The above are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A content generation system, characterized in that: The content generation system includes: a scene function layer, an intermediate interface layer and a model layer; The scene function layer is used to obtain a content generation request input by a user, and perform request parsing on the content generation request to obtain target request data; The intermediate interface layer is used to determine the application function interface corresponding to the target request data, and call the content generation model of the model layer based on the application function interface to generate content to obtain target content.

2. The content generation system according to claim 1, characterized in that: The intermediate interface layer is used to determine the interface calling logic of the target request data; The intermediate interface layer is further used to determine the application function interface and the interface calling sequence required to execute the content generation request according to the interface calling logic.

3. The content generation system according to claim 2, characterized in that: The intermediate interface layer is further used to determine the application function interface required to execute the content generation request and the interface call sequence composed of the application function interface based on the interface call logic; The intermediate interface layer is further used to determine adjacent interfaces in the interface call sequence as a preceding interface and a succeeding interface; The intermediate interface layer is further used to determine the interface calling order in the interface calling sequence and use the first target content corresponding to the preceding interface as the input of the succeeding interface.

4. The content generation system according to any one of claims 1 to 3, characterized in that: The model layer deploys multiple content generation models, and the training method of the content generation model includes: Inputting the random vector into the generator of the content generation model to generate samples, thereby obtaining target simulation samples; Inputting the target reference sample and the target simulation sample into the discriminator of the content generation model to perform sample judgment and obtain a sample judgment result; Performing difference analysis on the sample judgment results to obtain sample prediction probability; Adjusting parameters of the generator or the discriminator based on the sample prediction probability; When the content generation model is adjusted to a convergence condition, the trained content generation model is obtained.

5. The content generation system according to claim 4, characterized in that: The step of inputting the target reference sample and the target simulation sample into the discriminator of the content generation model to perform sample judgment and obtain the sample judgment result includes: Obtain target discriminant function for sample judgment; Based on the target discriminant function, sample judgment is performed on each target reference sample and target simulation sample input to the discriminator to obtain a sample output probability value; the sample output probability value is used to characterize the probability that the input sample is a true sample; A sample judgment result is determined based on the sample output probability value.

6. The content generation system according to claim 4, characterized in that: The sample judgment result includes an expected result and an unexpected result. The step of performing difference analysis on the sample judgment result to obtain the sample prediction probability includes: Obtain the expected sample output probability value corresponding to the expected result and the undesired sample output probability value corresponding to the undesired result; Determine the expected term of the function based on the expected result and the expected sample output probability value; Determine an undesirable term of the function based on the undesirable result and the undesirable sample output probability value; The sample prediction probability is obtained according to the loss difference between the expected term of the function and the unexpected term of the function.

7. The content generation system according to claim 4, characterized in that: In an iterative optimization process of the content generation model, alternately fixing a first model parameter of the generator and a second model parameter of the discriminator; The step of adjusting parameters of the generator or the discriminator based on the sample prediction probability comprises: When the first model parameters of the generator are fixed, adjusting the second model parameters of the discriminator based on the sample prediction probability; When the second model parameters of the discriminator are fixed, the first model parameters of the generator are adjusted based on the sample prediction probability.

8. A content generation method based on the content generation system according to any one of claims 1 to 7, characterized in that: The method comprises: Obtaining a content generation request input by a user, and performing request parsing on the content generation request to obtain target request data; An application function interface corresponding to the target request data is determined, and based on the application function interface, a content generation model of the model layer is called to generate content to obtain target content.

9. A content generating device, characterized in that: The device is deployed with a content generation system as claimed in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the content generation method according to claim 8 are implemented.

Images