Pricing method based on AIGC service and related equipment

Abstract

The embodiment of the invention provides an AIGC service-based pricing method and related equipment, which are used for determining an efficient prompt pricing algorithm for an AIGC service platform. Comprising the steps of obtaining a generative artificial intelligence (GAI) model set, and determining all GAI sub-models and model prompt information in the GAI model set; obtaining a user prompt set and the prompt ambiguity of each user prompt intention; determining expected utility mapping according to the GAI model set and the user prompt set; according to the expected utility mapping, the model prompt information, all the GAI sub-models and the prompt ambiguity, determining a target GAI model, and determining a target prompt strategy; and determining isomorphic pricing information corresponding to the prompt pricing information under the same prompt ambiguity, thereby determining a target association relationship between the target prompt strategy and prompt cost information under the target GAI model according to the isomorphic pricing information, and determining target pricing information corresponding to the prompt pricing information under different prompt ambiguity according to the target association relationship.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of artificial intelligence, and in particular, to a pricing method and related equipment based on AIGC services. Background Art

[0002] Generative artificial intelligence (GAI) has achieved remarkable success in generating diverse and multimodal content based on user prompts. Accordingly, artificial intelligence-generated content (AIGC) services have also developed rapidly, such as the popularity of AIGC service platforms such as Monica, Poe, and ChatGpt. In practical applications, AIGC service platforms usually integrate multiple GAI models, such as GPT-4, GPT-4o, etc., to provide users with flexible purchase options and combinations. Users will purchase AIGC services based on usage indicators (such as the number of prompts, the number of tokens, etc.), interact with GAI models of different performance, and generate the required content through prompts. Therefore, this type of emerging AIGC service provides a new paradigm for the promotion and application of artificial intelligence and machine learning technologies, and promotes a surge in related research and deployment. Although the existing technologies have made preliminary explorations on GAI models, AIGC services, etc., they all focus on the application and performance optimization of GAI-related technologies, and there are obvious research and technical gaps in user usage decisions and the prompt pricing mechanism of service platforms.

[0003] Therefore, how to combine user decisions and pricing mechanisms is a technical problem that needs to be solved urgently. Summary of the invention

[0004] The embodiment of the present application provides a pricing method and related equipment based on AIGC service, which is used to combine user decision-making and pricing mechanism, so as to determine an efficient prompt pricing algorithm for the AIGC service platform.

[0005] The first aspect of the embodiment of the present application provides a pricing method based on AIGC service, which is applied to the AIGC service platform. The method includes:

[0006] Obtain a generative artificial intelligence GAI model set, determine all GAI sub-models in the GAI model set, and model prompt information corresponding to the GAI sub-models respectively; wherein the model prompt information includes prompt pricing information and prompt cost information;

[0007] Obtaining a user prompt set corresponding to the user and a prompt ambiguity of each user prompt diagram in the user prompt set; wherein the user prompt diagram is used to characterize the user prompt strategy, and the prompt ambiguity is used to characterize the probability of deriving any user prompt diagram;

[0008] Determine an expected utility mapping according to the GAI model set and the user prompt set; wherein the expected utility mapping is used to characterize the expected utility of the generated content of any user prompt diagram under any GAI sub-model, and the generated content is associated with the model prompt information;

[0009] Determine a target GAI model among all the GAI sub-models according to the expected utility mapping, the model prompt information, all the GAI sub-models and the prompt ambiguity, and determine a target prompt strategy in the user prompt diagram;

[0010] Determine the isomorphic pricing information corresponding to the prompt pricing information under the same prompt ambiguity, so as to determine the target association relationship between the target prompt strategy and the prompt cost information under the target GAI model according to the isomorphic pricing information, so as to determine the target pricing information corresponding to the prompt pricing information under different prompt ambiguities according to the target association relationship.

[0011] Optionally, the acquiring a user prompt set corresponding to the user and a prompt ambiguity of each user prompt diagram in the user prompt set includes:

[0012] Obtain the user suggestion map for any user, and construct a user intention space based on all user suggestion maps of all users;

[0013] In the user intention space, determining the user prompt set corresponding to all the user prompt diagrams; wherein the user prompt set includes user prompt tasks distributed under the condition of the user prompt diagram;

[0014] The prompt ambiguity for any user prompt in the user prompt set is determined.

[0015] Optionally, determining an expected utility mapping according to the GAI model set and the user prompt set includes:

[0016] Determine the user-generated content under the user suggestion diagram in the user intention space in any GAI sub-model; wherein the user-generated content is used to represent the model-generated content obtained by the user from any GAI sub-model;

[0017] In the GAI model set, the expected utility corresponding to the user-generated content is determined, and all expected utilities are obtained to obtain the expected utility mapping.

[0018] Optionally, determining a target GAI model in all the GAI sub-models according to the expected utility mapping, the model prompt information, all the GAI sub-models and the prompt ambiguity, and determining a target prompt strategy in the user prompt diagram includes:

[0019] Constructing a policy association relationship between the user prompt diagram and the user prompt strategy based on a context learning algorithm; wherein the policy association relationship is used to characterize the prompt ambiguity between the user prompt strategy and the user prompt diagram corresponding to the user prompt strategy;

[0020] Determine expected benefit information associated with the prompt ambiguity according to the strategy association relationship, the prompt pricing information corresponding to any GAI sub-model, the expected utility mapping corresponding to any GAI sub-model, and the user prompt strategy;

[0021] Determine the target prompt strategy according to the expected benefit information, the prompt pricing information corresponding to any GAI sub-model, the expected utility mapping corresponding to any GAI sub-model, and the prompt ambiguity;

[0022] The target GAI model is determined according to the target prompt strategy, the prompt pricing information corresponding to any GAI sub-model, the expected utility mapping corresponding to any GAI sub-model, and the prompt ambiguity.

[0023] Optionally, the determining of isomorphic pricing information corresponding to the prompt pricing information under the same prompt ambiguity, determining a target association relationship between the target prompt strategy and the prompt cost information under the target GAI model according to the isomorphic pricing information, and determining target pricing information corresponding to the prompt pricing information under different prompt ambiguities according to the target association relationship, includes:

[0024] Under the same prompt ambiguity, determining the isomorphic pricing information according to the corresponding GAI sub-model, the prompt cost information, the target GAI model, the expected utility map and the user prompt diagram;

[0025] Associating the target prompt strategy with the prompt cost information in the GAI model set under the AIGC service platform to perform a prompt pricing algorithm under the different prompt ambiguities, and obtaining an unconstrained segmented association relationship between the prompt pricing information and the prompt cost information corresponding to any GAI sub-model;

[0026] Determine an upper limit function of prompt pricing corresponding to the target prompt strategy according to the unconstrained segment association relationship between any two GAI sub-models;

[0027] The target pricing information is determined according to the upper limit function of the prompt pricing.

[0028] Optionally, obtaining the unconstrained segmented association relationship between the prompt pricing information and the prompt cost information corresponding to any GAI sub-model includes:

[0029] According to the prompt pricing information and the expected utility mapping under any GAI sub-model, a first function solution for the prompt ambiguity is obtained; according to the upper limit value and the number value of the target prompt strategy, a second function solution for the prompt ambiguity is obtained;

[0030] The unconstrained segment association relationship is established according to the first function solution and the second function solution, and the maximum value of the difference between the prompt pricing information and the prompt cost information.

[0031] Optionally, determining an upper limit function of prompt pricing corresponding to the target prompt strategy according to the unconstrained segment association relationship between any two GAI sub-models includes:

[0032] Determine a minimum value of the number of times corresponding to the target prompt strategy according to the target prompt strategy, the number of times of the target prompt strategy, the expected utility mapping, the prompt ambiguity and the prompt pricing information under any GAI sub-model;

[0033] An upper limit function of the prompt pricing of the target prompt strategy is determined according to the minimum value of the frequency value of the target prompt strategy, the unconstrained segment association relationship, the prompt ambiguity, the expected utility mapping under any two GAI sub-models and the prompt pricing information.

[0034] A second aspect of the embodiment of the present application provides a pricing system based on AIGC services, including:

[0035] An acquisition unit, used to acquire a generative artificial intelligence GAI model set, determine all GAI sub-models in the GAI model set, and model prompt information corresponding to the GAI sub-models respectively; wherein the model prompt information includes prompt pricing information and prompt cost information;

[0036] The acquisition unit is further used to acquire a user prompt set corresponding to the user and a prompt ambiguity of each user prompt diagram in the user prompt set; wherein the user prompt diagram is used to characterize the user prompt strategy, and the prompt ambiguity is used to characterize the probability of deriving any user prompt diagram;

[0037] A determination unit, configured to determine an expected utility mapping according to the GAI model set and the user prompt set; wherein the expected utility mapping is used to characterize the expected utility of the generated content of any user prompt diagram under any GAI sub-model, and the generated content is associated with the model prompt information;

[0038] The determining unit is further configured to determine a target GAI model among all the GAI sub-models according to the expected utility mapping, the model prompt information, all the GAI sub-models and the prompt ambiguity, and determine a target prompt strategy in the user prompt diagram;

[0039] The determination unit 902 is also used to determine the isomorphic pricing information corresponding to the prompt pricing information under the same prompt ambiguity, so as to determine the target association relationship between the target prompt strategy and the prompt cost information under the target GAI model according to the isomorphic pricing information, so as to determine the target pricing information corresponding to the prompt pricing information under different prompt ambiguities according to the target association relationship.

[0040] The AIGC service-based pricing system provided in the second aspect of the embodiment of the present application is used to execute the AIGC service-based pricing method described in the first aspect.

[0041] A third aspect of the embodiment of the present application provides a pricing device based on AIGC service, including:

[0042] CPU, memory, input and output interfaces, wired or wireless network interfaces, and power supply;

[0043] The memory is a short-term storage memory or a persistent storage memory;

[0044] The central processor is configured to communicate with the memory and execute instruction operations in the memory to perform the pricing method based on AIGC service described in the first aspect.

[0045] A fourth aspect of an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium includes instructions. When the instructions are executed on a computer, the computer executes the pricing method based on the AIGC service described in the first aspect.

[0046] A fifth aspect of an embodiment of the present application provides a computer program product, which includes instructions. When the instructions are executed on a computer, the computer executes the pricing method based on the AIGC service described in the first aspect.

[0047] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages: through a pricing method based on AIGC services disclosed in the embodiments of the present application, it is possible to achieve policy interaction between users with different prompting capabilities and AIGC service receipts. For the AIGC service platform, under any user distribution, the optimal pricing mechanism and pricing algorithm for different GAI models are established to maximize the revenue of the AIGC service platform. At the same time, through the optimization of user and platform revenue, the positive feedback and virtuous cycle of AIGC services are achieved, which promotes the flow of resources and technological progress in related fields, promotes the application scenarios of AIGC, and improves practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present application. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0049] Figure 1 A schematic diagram of the framework of an AIGC service process disclosed in an embodiment of the present application;

[0050] Figure 2 An example diagram of a user prompt and intention disclosed in an embodiment of the present application;

[0051] Figure 3 A flowchart of a pricing method based on AIGC service disclosed in an embodiment of the present application;

[0052] Figure 4 A flowchart of another pricing method based on AIGC service disclosed in an embodiment of the present application;

[0053] Figure 5 A flowchart of another pricing method based on AIGC service disclosed in an embodiment of the present application;

[0054] Figure 6 An interaction diagram between a user and a platform under an AICG service process disclosed in an embodiment of the present application;

[0055] Figure 7 A schematic diagram of the relationship between an optimal prompt strategy for a user and prompt ambiguity disclosed in an embodiment of the present application;

[0056] Figure 8 A schematic diagram of the relationship between the optimal prompt strategy for a user and the prompt ambiguity under an optimal prompt price for a homogeneous scenario disclosed in an embodiment of the present application;

[0057] Fig. 9A schematic diagram of the structure of a pricing system based on AIGC service disclosed in an embodiment of the present application;

[0058] Fig.10 This is a structural diagram of a pricing device based on AIGC service disclosed in an embodiment of the present application. DETAILED DESCRIPTION

[0059] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0060] It should be noted that the descriptions involving "first", "second", etc. in this application are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in this field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0061] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0062] To facilitate the understanding and description of the AIGC service platform (hereinafter described as a platform) proposed in the technical solution of this application, please refer to Figure 1 and Figure 2 .in, Figure 1 A schematic diagram of the framework of an AIGC service process disclosed in an embodiment of the present application is shown in FIG. Figure 2 This is an example diagram of user prompts and intentions disclosed in an embodiment of the present application.

[0063] It should be noted that the embodiment of the present application discloses a user decision and pricing mechanism design scheme for AIGC services. By introducing the concept of prompt ambiguity, the user prompts are quantified in language. At the same time, the latent intention in the user prompt is associated with the generated content of the GAI model to establish a utility mapping. According to the quantification between user prompts and utility, the decision space for user model selection is simplified and the optimal number of user prompts for different GAI models is determined. Based on the analysis of user usage decisions, the embodiment of the present application formalizes the interaction between users and the platform in the AIGC service into a two-stage Stackelberg game. Through the correspondence between price and user decision, the platform's pricing mechanism design for each single GAI model is transformed into an unconstrained segmented form, and finally an efficient prompt pricing algorithm is established for the AIGC service platform. Specifically, AIGC services mainly involve multi-layer interactions between platforms, GAI models and users (such as Figure 1 shown):

[0064] 1) The platform announces the AIGC service pricing mechanism: Based on the performance and training cost differences between GAI models, the AIGC service platform announces the prompt prices of different models. Specifically, the billing of AIGC service usage mainly includes user-oriented subscriptions and token-based billing. Since such billing is calculated based on user usage, for the sake of simplicity, the price of a single prompt is considered as the minimum unit of AIGC service usage billing under different GAI models.

[0065] 2) User purchase of AIGC service usage: Each user prompt is intended to convey the user’s specific intent (i.e., task) to the GAI model, thereby generating the required content (e.g., Figure 2 As shown). In practice, user prompts are closely related to the user's prompting ability and the inherent complexity of the task, and it is often difficult to accurately convey the target intent. Therefore, the embodiment of the present application introduces the concept of prompt ambiguity to quantify the prompting ability of different users. Users with different prompting abilities will purchase the corresponding GAI model prompt usage according to their own circumstances and the AIGC service pricing mechanism to maximize the benefits of each user. Specifically, users professionally polish the text by proposing various intentions. For example, prompts such as avoiding long sentences, emphasizing key information, or emphasizing logical connections can be proposed.

[0066] 3) The platform sends GAI model usage prompts: Based on the user's payment situation, the platform sends the corresponding AIGC service usage prompts to the user's account for subsequent use by the user.

[0067] 4) Users prompt the GAI model to generate content: Users use the purchased prompts to interact with the GAI model, and guide the GAI model to generate target content through single or multiple prompts under each task.

[0068] To understand the above Figure 1 and Figure 2 As shown, and to solve the technical problems described above, please refer to Figure 3 , Figure 3 This is a flowchart of a pricing method based on AIGC service disclosed in an embodiment of the present application, including steps 301 to 305.

[0069] 301. Obtain a generative artificial intelligence GAI model set, determine all GAI sub-models in the GAI model set, and model prompt information corresponding to each GAI sub-model.

[0070] based on Figure 1 and Figure 2 As shown, in order to establish the design of user decision-making and pricing mechanism between the AIGC service platform and users, it is necessary to obtain the GAI model set on the AIGC service platform in advance. Therefore, all GAI sub-models and model prompt information contained in the GAI sub-model are determined in the GAI model set. Furthermore, the model prompt information includes prompt pricing information and prompt cost information.

[0071] In one specific embodiment, the GAI model set includes multiple GAI sub-models, such as GPT4 or GPT-4o, etc., and the GAI sub-model is not specifically limited here. At the same time, each GAI sub-model (it can also be understood as a separate GAI model. This embodiment is to distinguish it from the GAI model set. Each GAI sub-model will be described as a GAI model later) can set a single prompt price (i.e., the prompt pricing information described above) and a single prompt cost (i.e., the prompt cost information described above) for the current GAI model. It should be noted that for the AIGC service platform, each prompt received by the GAI model will generate corresponding computing and communication resource costs. To simplify the device, it is understood as a prompt cost. It should also be noted that for the convenience of subsequent descriptions, the AIGC service platform will be referred to as the platform, and it will not be repeated later.

[0072] 302. Obtain a user prompt set corresponding to the user and a prompt ambiguity of each user prompt diagram in the user prompt set.

[0073] For any user, the user prompt set of the user and the prompt ambiguity of each user prompt diagram in the user prompt set can be obtained. It should be noted that the user prompt diagram is used to characterize the user prompt strategy, and the prompt ambiguity is used to characterize the probability of deriving any user prompt diagram.

[0074] In one specific embodiment, the user's suggestion diagram (task) can be summarized, wherein the suggestion diagram (task) can be understood as Figure 2 As shown, for example, the task of polishing text, or realizing text recognition, translation and sentence construction, etc. By summarizing the prompt diagram (task) into the user intention space, it can be understood that any user's prompt can be described as a specific distribution independently generated under the condition of user intention. Therefore, based on the above prompt space, prompt ambiguity can be introduced. For user prompts, when the probability of its satisfaction is related to user intention and user prompt, the corresponding probability can be called prompt ambiguity, which refers to the probability of inferring user intention from user prompts.

[0075] 303. Determine an expected utility mapping according to the GAI model set and the user prompt set.

[0076] Based on step 302, the expected utility mapping can be determined according to the GAI model set and the user prompt set. It should be noted that the expected utility mapping is used to characterize the expected utility of the generated content of any user prompt diagram under any GAI sub-model, and the generated content is associated with the model prompt information.

[0077] In one specific embodiment, since in the GAI model set, users can obtain utility from the generated content of each GAI model (each GAI model can output relevant content based on the user input content). Among them, the utility depends on the comprehensive performance of the GAI model in various tasks. It should be noted that "utility", in this embodiment, such as translating a paragraph of text, is the level of translated text (which can be understood as the level of translation into faithfulness, expressiveness and elegance). Or, it can be the prediction accuracy of emotion recognition. Therefore, it is possible to establish an expected utility mapping between generated content and intent space.

[0078] 304. Determine a target GAI model among all GAI sub-models according to the expected utility mapping, the model prompt information, all GAI sub-models and the prompt ambiguity, and determine a target prompt strategy in the user prompt diagram.

[0079] Based on step 303, the target GAI model can be determined among all GAI sub-models according to the expected utility mapping, model prompt information, all GAI sub-models and prompt ambiguity, thereby determining the target prompt strategy in the user prompt diagram.

[0080] In one specific embodiment, in combination Figure 1 and Figure 2 In the second stage of the AIGC process, the user aims to determine the prompt decision of the GAI model and the prompt strategy corresponding to the prompt decision, so as to maximize its expected benefits. Therefore, in-context learning can be used to construct a theoretical relationship between the user prompt strategy and the user intention, so as to determine the correlation between the user's prompt strategy and the user intention carried by the prompt strategy. Therefore, for any user's prompt strategy, it can be quantified through probabilistic user prompts to characterize the user's expected benefits, and then formulate a usage decision problem with arbitrary fuzziness in the second stage of the AIGC process.

[0081] Furthermore, based on the above process, the optimal prompt strategy for a user with a determined prompt ambiguity under the user decision of a given GAI model of any user can be determined, that is, the target prompt strategy mentioned above. Furthermore, the optimal selection strategy of the GAI model can also be obtained.

[0082] 305. Determine isomorphic pricing information corresponding to the prompt pricing information under the same prompt ambiguity, so as to determine the target association relationship between the target prompt strategy and the prompt cost information under the target GAI model according to the isomorphic pricing information, so as to determine the target pricing information corresponding to the prompt pricing information under different prompt ambiguities according to the target association relationship.

[0083] Based on step 304, the isomorphic pricing information corresponding to the prompt pricing information under the same prompt ambiguity can be determined, thereby determining the target association relationship between the target prompt strategy and the prompt cost information under the target GAI model based on the isomorphic pricing information. Therefore, the target pricing information corresponding to the prompt pricing information under different prompt ambiguities can be determined based on the target association relationship.

[0084] In one of the specific embodiments, the problem of designing the prompt pricing mechanism of the platform on the GAI model set can be formulated as a three-layer non-convex optimization based on the user's benefit maximization usage decision in the second stage of the AIGC service process. Specifically, the optimal pricing of the AIGC service platform is first determined in the scenario of homogeneous users (same prompt ambiguity). Thus, under the optimal prompt price of the homogeneous scenario, the user's optimal prompt strategy is associated with the platform's GAI model prompt cost, so as to design a prompt pricing algorithm in the scenario of heterogeneous users (different prompt ambiguities). Through the price prompt upper limit function, the prompt ambiguity distribution of any heterogeneous user group can be characterized by a probability density function, and an equivalent objective function is constructed at the same time, thereby determining the optimal prompt pricing algorithm of the AIGC service platform in the equivalent objective function, and thereby determining the target pricing information.

[0085] Through the pricing method based on AIGC service disclosed in this embodiment, the policy interaction between users with different prompting capabilities and AIGC service receipts can be realized. For the AIGC service platform, under any user distribution, the optimal pricing mechanism and pricing algorithm of different GAI models are established to maximize the revenue of the AIGC service platform. At the same time, through the optimization of user and platform revenue, the positive feedback and virtuous cycle of AIGC services are realized, which promotes the flow of resources and technological progress in related fields, promotes the application scenarios of AIGC, and improves practicality.

[0086] To facilitate understanding of the above Figure 3 In step 301 to step 305, in another feasible technical solution, please refer to Figure 4 , Figure 4 This is a flowchart of another pricing method based on AIGC service disclosed in an embodiment of the present application, including steps 401 to 412.

[0087] 401. Obtain a generative artificial intelligence GAI model set, determine all GAI sub-models in the GAI model set, and model prompt information corresponding to each GAI sub-model.

[0088] In this embodiment, step 401 is the same as the aforementioned Figure 3 Step 301 is similar, and will not be described in detail here. However, it should be noted that in this embodiment, the platform publishes the GAI model set M; prompts pricing (i.e., prompts pricing information) where p m Represents a single prompt price of the GAI model m∈M; prompt cost (i.e. prompt cost information) Among them C m ∈C represents the single prompt cost of GAI model m∈M (for the platform, each prompt received by GAI model m∈M will incur corresponding computing and communication resource costs. For simplicity, it is normalized to C here. m ).

[0089] 402. Obtain a user suggestion map for any user, and construct a user intention space based on all user suggestion maps of all users.

[0090] In the embodiment, steps 402 to 403 are the same as those described above. Figure 3 However, it should be noted that, in this embodiment, it is necessary to first obtain the user suggestion diagram for any user, and then construct the user intention space based on all user suggestion diagrams of all users.

[0091] In one specific embodiment, for any user, the proposed diagram (task) is represented as γ∈Γ, where Γ is the user intention space. It should be noted that Γ can be understood as a discrete complete set of the user intention space. Correspondingly, γ is the user proposed diagram.

[0092] 403. In the user intention space, determine a user prompt set corresponding to all user suggestion diagrams, and determine the prompt ambiguity for any user suggestion diagram in the user prompt set.

[0093] Based on step 402, a user prompt set corresponding to all user suggestion diagrams can be determined in the user intention space. Furthermore, the prompt ambiguity for any user suggestion diagram can also be determined in the user prompt set. It should be noted that the user prompt set includes user prompt tasks distributed under the condition of the user suggestion diagram.

[0094] In one specific embodiment, X γ Denote as the set of user prompts under the intent γ∈Γ. In particular, any user prompt x∈X γ Independently generated under the specific distribution of the intention γ∈Γ. Based on the above prompt space, the prompt ambiguity ε∈[0,1] is further introduced. For the user prompt x∈X γ , when it satisfies Prob(γ|x)≥1-ε, it is recorded as ε fuzzy, which means that the probability of correctly inferring the target intention from the user prompt x is greater than or equal to 1-ε. It should be noted that the PROB function refers to the function of the corresponding probability that the value falls within the specified interval, and returns the corresponding probability that the value in the area falls within the specified interval.

[0095] 404. Determine user-generated content under the user suggestion diagram in the user intent space in any GAI sub-model.

[0096] It should be noted that steps 404 to 405 in this embodiment are similar to those in the above Figure 3 The step 303 in the example is similar. Specifically, the user-generated content under the user suggestion diagram in the user intention space in any GAI sub-model can be determined. It should also be noted that the user-generated content is used to represent the model-generated content obtained by the user from any GAI sub-model.

[0097] In one specific embodiment, the utility that a user obtains from the content generated by the GAI model m∈M depends on the comprehensive performance of the model on various tasks, such as the prediction accuracy of emotion recognition.

[0098] 405. In the GAI model set, determine the expected utility corresponding to the user-generated content, and obtain all expected utilities to obtain an expected utility mapping.

[0099] Therefore, based on step 404, the expected utility corresponding to the user-generated content can be determined in the GAI model set, so as to obtain all expected utilities and obtain an expected utility mapping.

[0100] In one specific embodiment, for a GAI model m∈M, U m represents the expected utility mapping between the generated content and the intention space, namely, U m =Ε γ∈Γ [U m (γ)], where U m (γ) represents the expected utility of the generated content of the GAI model m∈M under the intention γ∈Γ. In short, it can also be understood as the average of all expected utilities, etc.

[0101] 406. Construct a strategic association relationship between the user prompt diagram and the user prompt strategy based on a context learning algorithm.

[0102] It should be noted that steps 406 to 409 in this embodiment are similar to the aforementioned Figure 3 Based on step 405, a policy association relationship between the user prompt diagram and the user prompt strategy can be constructed through a context learning algorithm. It should be noted that the policy association relationship is used to characterize the prompt ambiguity between the user prompt strategy and the user prompt diagram corresponding to the user prompt strategy.

[0103] In one specific embodiment, please refer to Figure 6 , Figure 6 This is an interaction diagram between a user and a platform under an AICG service process disclosed in an embodiment of the present application. In the second stage of the AICG process, the user aims to make a GAI model decision s∈M and a corresponding prompt strategy n∈N + , in order to maximize its expected return π user It should be noted that s is the GAI model selected in M, n is the number of prompts, N + Represents a positive integer. First, we use context learning (In-contextLearning) to build a user prompt strategy n∈N + Theoretical relationship with target intention γ∈Γ:

[0104] Prob(γ 1 =γ,γ 2 =γ,...,γ k =γ|x 1 ,x 2 ,...,x k )≥1-ε k , where x k is the kth tip of the user, γ kis the intention carried by the kth prompt, for any k∈N + , which can be quantified through probabilistic user prompts.

[0105] 407. Determine expected benefit information associated with prompt ambiguity according to the strategy association relationship, prompt pricing information corresponding to any GAI sub-model, expected utility mapping corresponding to any GAI sub-model, and user prompt strategy.

[0106] Based on step 406, the expected benefit information associated with the prompt ambiguity can be determined according to the strategy association relationship, the prompt pricing information corresponding to any GAI sub-model, the expected utility mapping corresponding to any GAI sub-model, and the user prompt strategy.

[0107] In one specific embodiment, the expected return of the user can be represented as π user (s,n,ε)=(1-ε n )·U s -n·p s , thereby specifying an arbitrary ambiguity ε∈[0,1] in the AIGC process stage 2 (participating Figure 6 ) usage decision problem.

[0108] Specifically, there are

[0109] Among them, n * This is the target prompt strategy described later.

[0110] 408. Determine a target prompt strategy according to expected benefit information, prompt pricing information corresponding to any GAI sub-model, expected utility mapping corresponding to any GAI sub-model, and prompt ambiguity.

[0111] Therefore, based on step 407, it is possible to determine and give the optimal prompt strategy for users with prompt ambiguity ε=[0,1] under any user GAI model decision s∈M, that is, the target prompt strategy.

[0112]

[0113] It should be noted that n * This is the target prompting strategy described above.

[0114] 409. Determine a target GAI model according to a target prompt strategy, prompt pricing information corresponding to any GAI sub-model, expected utility mapping corresponding to any GAI sub-model, and prompt ambiguity.

[0115] In combination with the above steps 407 and 408, the target GAI model can be determined according to the target prompt strategy, the prompt pricing information corresponding to any GAI sub-model, the expected utility mapping corresponding to any GAI sub-model, and the prompt ambiguity.

[0116] In one specific embodiment, the selection strategy of the optimal GAI model can be obtained according to the equation. Specifically,

[0117]

[0118] Among them, s * This is the target GAI model described above.

[0119] Further, see Figure 7 , Figure 7 The following is a schematic diagram of the relationship between the optimal prompt strategy and prompt ambiguity of a user disclosed in the embodiment of the present application. By analyzing the optimal usage decision of the user, the relationship between the optimal prompt strategy and prompt ambiguity of the user under different pricing mechanisms is further established, such as Figure 7 As shown in Figure 2, this lays the foundation for the design of prompt pricing for the first phase of the AIGC service platform. Specifically, when the prompt price p of the GAI model corresponding to the user decision s = 0, the user's optimal prompt strategy is infinite; when p s ∈(0,U s / 4], the user's optimal prompt strategy n * (s,ε) first increases and then decreases with the increase of the hint ambiguity ε; when p s ∈(U s / 4,U s ), the user's optimal prompt strategy n * (s,ε) decreases continuously with the increase of the hint ambiguity ε; when p s ∈[U s ,∞), the user's optimal prompt strategy n * (s,ε) is always zero.

[0120] 410. Under the same prompt ambiguity, determine the isomorphic pricing information according to the corresponding GAI sub-model, prompt cost information, target GAI model, expected utility mapping and user prompt diagram.

[0121] Therefore, based on the above steps, isomorphic pricing information can be determined under the same prompt ambiguity according to any GAI sub-model, prompt cost information, target GAI model, expected utility mapping and user prompt diagram.

[0122] In one specific embodiment, the problem of designing the prompt pricing mechanism of the platform on the GAI model set M can be formulated as a three-layer non-convex optimization based on the user's decision to maximize the benefits in the second stage of the AIGC service process. Specifically, there are:

[0123]

[0124] Based on this, we first set the optimal pricing of the AIGC service platform in the scenario of homogeneous users (same prompt ambiguity).

[0125]

[0126] Among them, m * = arg max s∈M n * (s,ε)(ε σ(s,ε)-1 (1-ε)U s -C s ),

[0127] In short, σ(m, ε) in this embodiment is a simplified representation of a mathematical calculation formula and has no actual meaning.

[0128] 411. Associate the target prompt strategy with the prompt cost information in the GAI model set under the AIGC service platform to perform a prompt pricing algorithm under different prompt ambiguities, and obtain an unconstrained segmented association relationship between the prompt pricing information and the prompt cost information corresponding to any GAI sub-model.

[0129] Therefore, based on step 410, the target prompt strategy can be associated with the prompt cost information in the GAI model set under the AIGC service platform, so as to perform a prompt pricing algorithm under different prompt ambiguities and obtain an unconstrained segmented association relationship between the prompt pricing information and the prompt cost information under any GAI sub-model.

[0130] In one specific embodiment, in combination with the above description, under the optimal prompt price of the isomorphic scenario, the user's optimal prompt strategy is associated with the GAI model prompt cost of the platform (such as Figure 8 As shown, Figure 8 A schematic diagram of the relationship between the optimal prompt strategy and prompt ambiguity of a user under the optimal prompt price in a homogeneous scenario disclosed in an embodiment of the present application is provided to design a prompt pricing algorithm in a heterogeneous user (different prompt ambiguity) scenario.

[0131] Specifically, the single GAI model pricing design problem is first transformed into an unconstrained segmented optimization.

[0132] Thus, an unconstrained segmented association relationship between the prompt pricing information and the prompt cost information is established.

[0133] 412. Determine an upper limit function of prompt pricing corresponding to the target prompt strategy based on the unconstrained segmented association relationship between any two GAI sub-models, and determine the target pricing information based on the upper limit function of prompt pricing.

[0134] Based on step 411, the upper limit function of the prompt pricing corresponding to the target prompt strategy can be determined according to the unconstrained segment association relationship between any two GAI sub-models, thereby determining the target pricing information according to the upper limit function of the prompt pricing.

[0135] In one specific embodiment, a prompt price cap for user model decision is established between any two GAI models.

[0136] It should be noted that m′ and m are any two GAI models. It should also be noted that τ is a function code, specifically,

[0137]

[0138] Therefore, through the price prompt upper limit function, an equivalent objective function can be constructed to obtain the final target pricing information. Figure 5 The embodiment shown.

[0139] Furthermore, through the price prompt upper limit function, the following GAI model set M = {M L ,M H} as an example (i.e., ordinary and advanced GAI model types, such as GPT-4o and GPT-4).

[0140] For any heterogeneous user group, the prompt ambiguity distribution is represented by the probability density function f(ε), which is modified to:

[0141]

[0142] At the same time, construct an equivalent objective function:

[0143]

[0144] Therefore, the optimal prompt pricing (OPP) algorithm of the AIGC service platform is given as follows (expanded in pseudo code):

[0145]

[0146]

[0147] It should also be noted that (1) other pricing and billing methods can be used, but the billing is still based on the user's usage of AIGC services, including but not limited to user-oriented subscriptions and billing based on tokens, number of prompts, etc. Or, (2) other prompt quantification concepts can be used, but the analysis and design are still based on the impact of user prompts on the quality of content generated by the GAI model, including but not limited to the empirical quantification of user prompts on AIGC. Or, (3) other prompts and model modalities can be used, but the user and model are still used for interaction, including but not limited to text, image and video generation. Or, (4) AIGC service processes can be used, but they still include a variety of differentiated service performance offerings, including but not limited to upper-level service platforms on different platforms, such as cross-platform AIGC service integration such as Claude and ChatGPT.

[0148] Through a pricing method based on AIGC service disclosed in this embodiment, by establishing the optimal two-layer decision of heterogeneous users in AIGC service and the design of prompt pricing mechanism of the platform under any user group distribution, the benefits of each user and platform are optimized, and the positive feedback and virtuous cycle of AIGC service are realized. By introducing the concept of prompt ambiguity to quantify the user's prompt ability, the optimal GAI model selection strategy and prompt strategy of the user are obtained, and on this basis, the two-stage Stackelberg game between the user and the platform is solved by reverse deduction, and the optimal prompt pricing mechanism and pricing algorithm of the platform are designed to achieve a win-win situation for heterogeneous users and AIGC service platform, and promote resource flow and technological development in related fields. Specifically, compared with the existing GAI model and AIGC technology, the embodiment of the present application is oriented to AIGC service, and the optimal user usage decision and platform pricing mechanism design are established. At the same time, the benefits of each user and platform in the AIGC service process are optimized, a benign and efficient AIGC service ecological cycle is realized, and resource flow and technological progress in related fields are promoted.

[0149] To facilitate understanding of the above Figure 4 Steps 411 to 412 are shown in FIG. Figure 5 , Figure 5 This is a flowchart of another pricing method based on AIGC service disclosed in an embodiment of the present application, including steps 501 to 504.

[0150] 501. According to the prompt pricing information and expected utility mapping under any GAI sub-model, a first function solution for the prompt ambiguity is obtained; according to the upper limit value and the frequency value of the target prompt strategy, a second function solution for the prompt ambiguity is obtained.

[0151] Specifically, in this embodiment, a first function solution for prompt ambiguity can be obtained based on the prompt pricing information and expected utility mapping under any GAI sub-model; and a second function solution for prompt ambiguity can be obtained based on the upper limit value and the number value of the target prompt strategy.

[0152] In one specific embodiment, according to Figure 4 The first half of the formula in step 411 obtains the first function solution (corresponding to the first function relationship The second half obtains the second function solution (corresponding to the second function relationship The details will not be elaborated here.

[0153] 502. Establish an unconstrained segment association relationship according to the first function solution and the second function solution, and the maximum value of the difference between the prompt pricing information and the prompt cost information.

[0154] Based on step 501, an unconstrained segment association relationship can be established according to the first function solution and the second function solution, and the maximum value of the difference between the prompt pricing information and the prompt cost information.

[0155] In one specific embodiment, Figure 4 As shown in step 411, according to the first functional relationship and the second functional relationship, Figure 4 The equation of step 411 is to establish an unconstrained segment association relationship

[0156] 503. Determine the minimum value of the frequency value corresponding to the target prompt strategy according to the target prompt strategy, the frequency value of the target prompt strategy, the expected utility mapping, the prompt ambiguity and the prompt pricing information under any GAI sub-model.

[0157] Based on step 502, the minimum value of the frequency value corresponding to the target prompt strategy can be determined according to the target prompt strategy, the frequency value of the target prompt strategy, the expected utility mapping, the prompt ambiguity and the prompt pricing information under any GAI sub-model.

[0158] In one specific embodiment, it can be combined with Figure 4 The equation of step 412 in , This determines the minimum value of the target prompt strategy.

[0159] 504. Determine the upper limit function of the prompt pricing of the target prompt strategy according to the minimum value of the frequency value of the target prompt strategy, the unconstrained segment association relationship, the prompt ambiguity, the expected utility mapping under any two GAI sub-models and the prompt pricing information.

[0160] Furthermore, based on step 503, the upper limit function of the prompt pricing of the target prompt strategy can be determined according to the minimum value of the frequency value of the target prompt strategy, the unconstrained segment association relationship, the prompt ambiguity, the expected utility mapping under any two GAI sub-models, and the prompt pricing information. Figure 4 The equation for step 412 is Determines the upper limit of the prompt price for its user model decision.

[0161] Through the pricing method based on AIGC service disclosed in this embodiment, compared with the existing AIGC service pricing mechanism, the embodiment of this application introduces the concept of prompt ambiguity to quantify heterogeneous user prompts and capabilities, realizes the formulation of personalized user usage strategies, and provides guidance for maximizing the benefits of users in the process of using AIGC services. At the same time, with the help of the prompt ambiguity distribution of user groups, the present invention establishes the optimal prompt pricing mechanism and the corresponding OPP algorithm for the platform.

[0162] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0163] See also Fig. 9 , Fig. 9 This is a structural diagram of a pricing system based on AIGC services disclosed in an embodiment of the present application.

[0164] The acquisition unit 901 is used to acquire a generative artificial intelligence GAI model set, determine all GAI sub-models in the GAI model set, and model prompt information corresponding to the GAI sub-models respectively; wherein the model prompt information includes prompt pricing information and prompt cost information;

[0165] The acquisition unit 901 is further used to acquire a user prompt set corresponding to the user and a prompt ambiguity of each user prompt diagram in the user prompt set; wherein the user prompt diagram is used to characterize the user prompt strategy, and the prompt ambiguity is used to characterize the probability of deriving any user prompt diagram;

[0166] A determination unit 902 is used to determine an expected utility mapping according to the GAI model set and the user prompt set; wherein the expected utility mapping is used to characterize the expected utility of the generated content of any user prompt diagram under any GAI sub-model, and the generated content is associated with the model prompt information;

[0167] The determination unit 902 is further configured to determine a target GAI model from among all GAI sub-models according to the expected utility mapping, the model prompt information, all GAI sub-models and the prompt ambiguity, and determine a target prompt strategy from among the user prompt diagram;

[0168] The determination unit 902 is also used to determine the isomorphic pricing information corresponding to the prompt pricing information under the same prompt ambiguity, so as to determine the target association relationship between the target prompt strategy and the prompt cost information under the target GAI model according to the isomorphic pricing information, so as to determine the target pricing information corresponding to the prompt pricing information under different prompt ambiguities according to the target association relationship.

[0169] Exemplarily, the system includes:

[0170] The acquisition unit 901 is specifically used to acquire a user suggestion diagram for any user, and construct a user intention space based on all user suggestion diagrams of all users;

[0171] The determining unit 902 is specifically configured to determine a user prompt set corresponding to all user suggestion diagrams in the user intention space; wherein the user prompt set includes user prompt tasks distributed under the condition of the user suggestion diagram;

[0172] The determination unit 902 is further configured to determine the prompt ambiguity for any user prompt diagram in the user prompt set.

[0173] Exemplarily, the system includes:

[0174] Determine the user-generated content under the user suggestion diagram in the user intent space in any GAI sub-model; wherein the user-generated content is used to represent the model-generated content obtained by the user from any GAI sub-model;

[0175] The determination unit 902 is specifically configured to determine the expected utility corresponding to the user-generated content in the GAI model set, and obtain all expected utilities to obtain an expected utility mapping.

[0176] Exemplarily, the system further includes: a construction unit 903;

[0177] A construction unit 903 is used to construct a policy association relationship between a user prompting diagram and a user prompting strategy based on a context learning algorithm; wherein the policy association relationship is used to characterize the prompt ambiguity between the user prompting strategy and the user prompting diagram corresponding to the user prompting strategy;

[0178] The determining unit 902 is specifically configured to determine the expected benefit information associated with the prompt ambiguity according to the strategy association relationship, the prompt pricing information corresponding to any GAI sub-model, the expected utility mapping corresponding to any GAI sub-model, and the user prompt strategy;

[0179] The determination unit 902 is further configured to determine a target prompt strategy according to the expected benefit information, the prompt pricing information corresponding to any GAI sub-model, the expected utility mapping corresponding to any GAI sub-model, and the prompt ambiguity;

[0180] The determination unit 902 is further configured to determine a target GAI model according to a target prompting strategy, prompting pricing information corresponding to any GAI sub-model, an expected utility mapping corresponding to any GAI sub-model, and prompting ambiguity.

[0181] Exemplarily, the system includes:

[0182] The determining unit 902 is specifically configured to determine the isomorphic pricing information under the same prompt ambiguity according to any GAI sub-model, prompt cost information, target GAI model, expected utility mapping and user prompt diagram;

[0183] The acquisition unit 901 is specifically used to associate the target prompt strategy with the prompt cost information in the GAI model set under the AIGC service platform, so as to perform a prompt pricing algorithm under different prompt ambiguities, and obtain an unconstrained segmented association relationship between the prompt pricing information and the prompt cost information under any GAI sub-model;

[0184] The determining unit 902 is further configured to determine an upper limit function of the prompt pricing corresponding to the target prompt strategy according to the unconstrained segment association relationship between any two GAI sub-models;

[0185] The determining unit 902 is further configured to determine target pricing information according to the upper limit function of the prompt pricing.

[0186] Exemplarily, the system further includes: an establishing unit 904;

[0187] The acquisition unit 901 is specifically used to obtain a first function solution for the prompt ambiguity according to the prompt pricing information and the expected utility mapping under any GAI sub-model; and obtain a second function solution for the prompt ambiguity according to the upper limit value and the number value of the target prompt strategy;

[0188] The establishing unit 904 is used to establish an unconstrained segment association relationship according to the first function solution and the second function solution, and the maximum value of the difference between the prompt pricing information and the prompt cost information.

[0189] Exemplarily, the system includes:

[0190] The determining unit 902 is specifically configured to determine a minimum value of the number of times corresponding to the target prompt strategy according to the target prompt strategy, the number of times of the target prompt strategy, the expected utility mapping, the prompt ambiguity and the prompt pricing information under any GAI sub-model;

[0191] The determination unit 902 is also used to determine the upper limit function of the prompt pricing of the target prompt strategy based on the minimum value of the frequency value of the target prompt strategy, the unconstrained segment association relationship, the prompt ambiguity, the expected utility mapping under any two GAI sub-models and the prompt pricing information.

[0192] See below Fig.10 The structural diagram of a pricing device based on AIGC service disclosed in an embodiment of the present application includes:

[0193] CPU 1001, memory 1005, input / output interface 1004, wired or wireless network interface 1003 and power supply 1002;

[0194] The memory 1005 is a temporary storage memory or a permanent storage memory;

[0195] The CPU 1001 is configured to communicate with the memory 1005 and execute the instructions in the memory 1005 to perform the aforementioned Figures 3 to 5 A pricing method based on AIGC services in any of the illustrated embodiments.

[0196] The embodiment of the present application also provides a chip system, the chip system includes at least one processor and a communication interface, the communication interface and the at least one processor are interconnected through a line, and the at least one processor is used to run a computer program or instruction to execute the aforementioned Figures 3 to 5 A pricing method based on AIGC services in any of the illustrated embodiments.

[0197] The embodiment of the present application also provides a computer-readable storage medium, the computer-readable storage medium includes instructions, when the instructions are executed on a computer, the computer executes the aforementioned Figures 3 to 5 A pricing method based on AIGC services in any of the illustrated embodiments.

[0198] The present application also provides a computer program product including instructions, which, when executed on a computer, enables the computer to execute the aforementioned Figures 3 to 5 A pricing method based on AIGC services in any of the illustrated embodiments.

[0199] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0200] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0201] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0202] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0203] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, read-only memory), random access memory (RAM, random access memory), disk or optical disk and other media that can store program code.

Claims

1. A pricing method based on AIGC service, characterized in that: Applied to the AIGC service platform, the method includes: Obtain a generative artificial intelligence GAI model set, determine all GAI sub-models in the GAI model set, and model prompt information corresponding to the GAI sub-models respectively; wherein the model prompt information includes prompt pricing information and prompt cost information; Obtaining a user prompt set corresponding to the user and a prompt ambiguity of each user prompt diagram in the user prompt set; wherein the user prompt diagram is used to characterize the user prompt strategy, and the prompt ambiguity is used to characterize the probability of deriving any user prompt diagram; Determine an expected utility mapping according to the GAI model set and the user prompt set; wherein the expected utility mapping is used to characterize the expected utility of the generated content of any user prompt diagram under any GAI sub-model, and the generated content is associated with the model prompt information; Determine a target GAI model among all the GAI sub-models according to the expected utility mapping, the model prompt information, all the GAI sub-models and the prompt ambiguity, and determine a target prompt strategy in the user prompt diagram; Determine the isomorphic pricing information corresponding to the prompt pricing information under the same prompt ambiguity, so as to determine the target association relationship between the target prompt strategy and the prompt cost information under the target GAI model according to the isomorphic pricing information, so as to determine the target pricing information corresponding to the prompt pricing information under different prompt ambiguities according to the target association relationship.

2. The pricing method based on AIGC service according to claim 1, characterized in that: The obtaining of a user prompt set corresponding to the user and a prompt ambiguity of each user prompt diagram in the user prompt set includes: Obtain the user suggestion map for any user, and construct a user intention space based on all user suggestion maps of all users; In the user intention space, determining the user prompt set corresponding to all the user prompt diagrams; wherein the user prompt set includes user prompt tasks distributed under the condition of the user prompt diagram; The prompt ambiguity for any user prompt in the user prompt set is determined.

3. The pricing method based on AIGC service according to claim 2, characterized in that: The determining of the expected utility mapping according to the GAI model set and the user prompt set includes: Determine the user-generated content under the user suggestion diagram in the user intention space in any GAI sub-model; wherein the user-generated content is used to represent the model-generated content obtained by the user from any GAI sub-model; In the GAI model set, the expected utility corresponding to the user-generated content is determined, and all expected utilities are obtained to obtain the expected utility mapping.

4. The pricing method based on AIGC service according to claim 3, characterized in that: The step of determining a target GAI model from among all the GAI sub-models according to the expected utility mapping, the model prompt information, all the GAI sub-models, and the prompt ambiguity, and determining a target prompt strategy from among the user prompt diagram includes: Constructing a policy association relationship between the user prompt diagram and the user prompt strategy based on a context learning algorithm; wherein the policy association relationship is used to characterize the prompt ambiguity between the user prompt strategy and the user prompt diagram corresponding to the user prompt strategy; Determine expected benefit information associated with the prompt ambiguity according to the strategy association relationship, the prompt pricing information corresponding to any GAI sub-model, the expected utility mapping corresponding to any GAI sub-model, and the user prompt strategy; Determine the target prompt strategy according to the expected benefit information, the prompt pricing information corresponding to any GAI sub-model, the expected utility mapping corresponding to any GAI sub-model, and the prompt ambiguity; The target GAI model is determined according to the target prompt strategy, the prompt pricing information corresponding to any GAI sub-model, the expected utility mapping corresponding to any GAI sub-model, and the prompt ambiguity.

5. The pricing method based on AIGC service according to claim 4, characterized in that: The determining of isomorphic pricing information corresponding to the prompt pricing information under the same prompt ambiguity, determining a target association relationship between the target prompt strategy and the prompt cost information under the target GAI model according to the isomorphic pricing information, and determining target pricing information corresponding to the prompt pricing information under different prompt ambiguities according to the target association relationship, includes: Under the same prompt ambiguity, determining the isomorphic pricing information according to the corresponding GAI sub-model, the prompt cost information, the target GAI model, the expected utility map and the user prompt diagram; Associating the target prompt strategy with the prompt cost information in the GAI model set under the AIGC service platform to perform a prompt pricing algorithm under the different prompt ambiguities, and obtaining an unconstrained segmented association relationship between the prompt pricing information and the prompt cost information corresponding to any GAI sub-model; Determine an upper limit function of prompt pricing corresponding to the target prompt strategy according to the unconstrained segment association relationship between any two GAI sub-models; The target pricing information is determined according to the upper limit function of the prompt pricing.

6. The pricing method based on AIGC service according to claim 5, characterized in that: The obtaining of the unconstrained segmented association relationship between the prompt pricing information and the prompt cost information corresponding to any GAI sub-model includes: According to the prompt pricing information and the expected utility mapping under any GAI sub-model, a first function solution for the prompt ambiguity is obtained; according to the upper limit value and the number value of the target prompt strategy, a second function solution for the prompt ambiguity is obtained; The unconstrained segment association relationship is established according to the first function solution and the second function solution, and the maximum value of the difference between the prompt pricing information and the prompt cost information.

7. The pricing method based on AIGC service according to claim 6, characterized in that: The step of determining an upper limit function of the prompt pricing corresponding to the target prompt strategy according to the unconstrained segment association relationship between any two GAI sub-models includes: Determine a minimum value of the number of times corresponding to the target prompt strategy according to the target prompt strategy, the number of times of the target prompt strategy, the expected utility mapping, the prompt ambiguity and the prompt pricing information under any GAI sub-model; An upper limit function of the prompt pricing of the target prompt strategy is determined according to the minimum value of the frequency value of the target prompt strategy, the unconstrained segment association relationship, the prompt ambiguity, the expected utility mapping under any two GAI sub-models and the prompt pricing information.

8. A pricing system based on AIGC service, characterized in that: Applied to the AIGC service platform, the system includes: An acquisition unit, used to acquire a generative artificial intelligence GAI model set, determine all GAI sub-models in the GAI model set, and model prompt information corresponding to the GAI sub-models respectively; wherein the model prompt information includes prompt pricing information and prompt cost information; The acquisition unit is further used to acquire a user prompt set corresponding to the user and a prompt ambiguity of each user prompt diagram in the user prompt set; wherein the user prompt diagram is used to characterize the user prompt strategy, and the prompt ambiguity is used to characterize the probability of deriving any user prompt diagram; A determination unit, configured to determine an expected utility mapping according to the GAI model set and the user prompt set; wherein the expected utility mapping is used to characterize the expected utility of the generated content of any user prompt diagram under any GAI sub-model, and the generated content is associated with the model prompt information; The determining unit is further configured to determine a target GAI model among all the GAI sub-models according to the expected utility mapping, the model prompt information, all the GAI sub-models and the prompt ambiguity, and determine a target prompt strategy in the user prompt diagram; The determination unit is also used to determine the isomorphic pricing information corresponding to the prompt pricing information under the same prompt ambiguity, so as to determine the target association relationship between the target prompt strategy and the prompt cost information under the target GAI model according to the isomorphic pricing information, so as to determine the target pricing information corresponding to the prompt pricing information under different prompt ambiguities according to the target association relationship.

9. A pricing device based on AIGC service, characterized in that: The device comprises: CPU, memory, input and output interfaces, wired or wireless network interfaces, and power supply; The memory is a short-term storage memory or a persistent storage memory; The central processing unit is configured to communicate with the memory and execute instruction operations in the memory to perform the pricing method based on AIGC services as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes instructions, and when the instructions are executed on a computer, the computer executes the pricing method based on AIGC services according to any one of claims 1 to 7.

Images