Information processing, shop diagnosis and intelligent assistant working method, system and device

By combining big models and causal inference technology, a Query template library and copywriting database suitable for store operation scenarios is built, which solves the problem of insufficient intelligence of e-commerce platform management tools, and realizes the precise diagnosis of store problems and the personality of operation strategies. The company has improved the operational efficiency of merchants.

CN118917411APending Publication Date: 2024-11-08HANGZHOU ALIBABA INT INTERNET IND CO LTD

Patent Information

Application Number
CN202410867589.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-28
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The management tools of existing e-commerce platforms lack intelligence, and it is difficult for merchants with insufficient experience to accurately locate store problems and adjust their business strategies through statistical data.

Method used

Combining large models and causal inference technology, a Query template library and copywriting database suitable for store operation scenarios is constructed, and analyses the store optimization items through causal inference algorithms, generate training sample sets to adjust and train pre-trained models, and provide intelligent response support. .

Benefits of technology

It improves the platform's intelligent service capabilities, helps merchants more accurately identify store problems and provide personalized operation suggestions, and improves store operation efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118917411A_ABST
    Figure CN118917411A_ABST
Patent Text Reader

Abstract

The embodiment of the invention provides an information processing, shop diagnosis and intelligent assistant working method, system and device. According to the scheme provided by the invention, the causal inference algorithm is utilized to construct the copywriting database (such as at least one factor of the preset problem obtained by performing attribution analysis on the preset problem by utilizing the causal inference algorithm stored in the copywriting database); and enabling the merchant side management tool by using the first preset model adjusted by using the training sample set related to the shop operation scene, so that the intelligent degree of the management tool is higher. The copywriting database provides data support and can retrieve at least one factor of a shop diagnosis result and / or a question inquired by a user and suggestion content associated with the factor; the first preset model provides intelligent response support, and based on the shop diagnosis result and / or the questions inquired by the user and the suggestion content associated with the at least one factor, the first preset model can output copywriting which more conforms to natural language logic and is suitable for the shop operation scene.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to an information processing method, a store diagnosis method, an intelligent assistant working method, a service system, an electronic device, a storage medium and a computer program product. Background Art

[0002] There are many merchants on the e-commerce platform. Merchants can manage the stores they operate on the e-commerce platform through the management tools provided by the platform. For example, they can use the management tool to obtain the store’s recent traffic data (such as the number of visits, the number of orders, the number of payments, etc.), and they can also use the management tool to set keywords for store products to increase the chance of product search, etc.

[0003] However, the existing management tools are not smart enough. Merchants can only understand some statistical data of the store through the management tools. For experienced merchants, these statistical data can accurately locate the problems of the store, and then adjust the business strategy based on rich experience. But for inexperienced merchants, seeing these statistical data cannot accurately locate the problems of the store, and they don’t know how to adjust the business strategy. Summary of the invention

[0004] The embodiments of the present application provide an information processing method, a store diagnosis method, an intelligent assistant working method, a service system, an electronic device, a storage medium and a computer program product to help merchants operate stores.

[0005] In the first embodiment of the present application, an information processing method is provided. The method comprises:

[0006] Determine the target problems for the target stores;

[0007] Retrieving at least one factor of the target problem in the document database;

[0008] Generate a prompt word Prompt based on the target question and the suggestion content associated with the at least one factor;

[0009] Outputting the text corresponding to the Prompt using the first preset model to help the user solve or improve the target problem;

[0010] Among them, the copywriting database stores: at least one factor of the preset problem obtained by attribution analysis of the preset problem using a causal inference algorithm; the first preset model is obtained by adjusting and training the pre-trained model based on the pre-trained model for the intelligent answering task in the store operation scenario.

[0011] In a second embodiment of the present application, an information processing method is provided. The method comprises:

[0012] Building a query template library, wherein the query template library includes multiple query templates, and each query template is associated with a preset question related to store operations;

[0013] Based on the Query template library, a document database is constructed; wherein the document database stores: at least one factor of the Query template obtained by performing attribution analysis on a preset question associated with the Query template using a causal inference algorithm;

[0014] Generate a training sample set according to the Query template library and the document database;

[0015] The pre-trained model is adjusted and trained using the training sample set to obtain a first preset model that can handle intelligent response tasks in store operation scenarios.

[0016] In the third embodiment of the present application, a store diagnosis method is provided. The method is applicable to a client, and specifically, the method includes:

[0017] In response to a diagnosis trigger event for a target store, triggering store diagnosis to present and / or play a diagnosed target problem;

[0018] Displaying and / or playing a copy, wherein the copy includes suggestion content to help the user solve or improve the target problem;

[0019] Among them, the copy is obtained by using a first preset model, which is obtained by adjusting and training the pre-trained model for the intelligent answering task in the store operation scenario on the basis of the pre-trained model; the suggested content is associated with at least one factor of the target problem, and at least one factor of the problem is obtained by attribution analysis of the target problem using a causal inference algorithm.

[0020] In the fourth embodiment of the present application, a method for operating a store smart assistant is provided. The method is applicable to a server, and specifically, the method includes:

[0021] Display the management page of the target store;

[0022] Monitor the user's operations on the management page to capture the user's target problems in store operations;

[0023] In response to a confirmation instruction triggered by the user regarding the target problem, display and / or play a copy, wherein the copy includes suggestion content to help the user solve or improve the target problem;

[0024] Among them, the copy is obtained by using a first preset model, which is obtained by adjusting and training the pre-trained model for the intelligent answering task in the store operation scenario on the basis of the pre-trained model; the suggested content is associated with at least one factor of the target problem, and at least one factor of the problem is obtained by attribution analysis of the target problem using a causal inference algorithm.

[0025] In the fifth embodiment of the present application, a service system is provided. The service system includes a client and a server. The client is used to implement the steps in the above-mentioned store diagnosis method embodiment or the steps in the above-mentioned store intelligent assistant working method embodiment. The server is used to implement the steps in the above-mentioned information processing method embodiments.

[0026] In a sixth embodiment of the present application, an electronic device is provided. The electronic device includes a memory and a processor. The memory is used to store executable instructions; the processor implements the steps in the above-mentioned method embodiments by running the executable instructions.

[0027] In a sixth embodiment of the present application, a computer-readable storage medium is provided, wherein the storage medium stores computer instructions, and when the instructions are executed by a processor, the steps in the above-mentioned method embodiments can be implemented.

[0028] The seventh embodiment of the present application further provides a computer program product, which includes a computer program or instructions. When the computer program or instructions are executed by a processor, the processor executes the steps in the above-mentioned method embodiments.

[0029] The pre-trained model mentioned in each embodiment of the present application refers to a machine learning model with a large number of parameters generated by training with a self-supervised method using massive data, which provides excellent distributed feature representation capabilities and model generalization capabilities for downstream tasks (such as the intelligent response tasks in the store operation scenarios mentioned in each embodiment of the present application). The pre-trained model can also be called a generalized large model.

[0030] The technical solution provided by the embodiment of the present application combines a large model with causal inference technology, that is, a causal inference algorithm is used when constructing a copywriting database to perform attribution analysis on the preset problem associated with the query template to obtain at least one factor; the training samples used to adjust and train the pre-trained model are generated based on the query template library and the copywriting database. Using a training sample set composed of multiple such training samples, the pre-trained model is adjusted and trained to adjust it to the first preset model suitable for the store operation scenario.

[0031] The solution provided by the embodiment of the present application uses a causal inference algorithm to construct a copywriting database (such as at least one factor of a preset problem obtained by performing attribution analysis on the preset problem using a causal inference algorithm stored in the copywriting database), and the first preset model adjusted using a training sample set related to the store operation scenario is used to empower the merchant-side management tool, so that the management tool is more intelligent. The copywriting database, the first preset model, etc. can be deployed on the server side, and the corresponding product form on the client side can be an intelligent assistant, through which the user can diagnose the target store, or ask questions related to the store operation, etc. The copywriting database provides data support for the intelligent assistant, and can retrieve at least one factor of the store diagnosis results and / or the questions asked by the user, and the recommended content associated with the factor; the adjusted and trained first preset model provides intelligent response support for the intelligent assistant. Based on the store diagnosis results and / or the questions asked by the user and the recommended content associated with at least one factor, the first preset model can output copywriting that is more in line with natural language logic and suitable for the store operation scenario to help users operate the store.

[0032] It can be seen that the technical solutions provided in the embodiments of the present application utilize large models and causal inference technology to enhance the intelligent service capabilities of the platform and help merchants operate their stores. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0034] Figure 1 A schematic diagram of the technical architecture corresponding to the technical solution provided in the embodiment of the present application;

[0035] Figure 2 A schematic diagram of a service system provided in accordance with an embodiment of the present application;

[0036] Figure 3 A flowchart of an information processing method provided in one embodiment of the present application;

[0037] Figure 4 A flowchart of an information processing method provided by another embodiment of the present application;

[0038] Figure 5 A flowchart of a store diagnosis method provided in one embodiment of the present application;

[0039] Figure 6a A first interface schematic diagram corresponding to the technical solution provided in one embodiment of the present application;

[0040] Figure 6b A second interface schematic diagram corresponding to the technical solution provided in one embodiment of the present application;

[0041] Figure 6c A third interface schematic diagram corresponding to the technical solution provided in one embodiment of the present application;

[0042] Figure 7 A flowchart of a store smart assistant working method provided in one embodiment of the present application;

[0043] Figure 8 A structural block diagram of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0044] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0045] In some processes described in the specification, claims and the above-mentioned figures of this application, multiple operations appearing in a specific order are included, and these operations may not be executed or executed in parallel in the order in which they appear in this article. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit "first" and "second" to different types. The embodiments described below are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of this application.

[0046] Before introducing various embodiments, some technical terms appearing in this document are briefly introduced.

[0047] Query: A query is a message sent by a search engine or database to find a file, website, record, or series of records in a database.

[0048] Pre-trained model: refers to a machine learning model with a large number of parameters generated by using a self-supervised method to train massive amounts of data. It provides excellent distributed feature representation and model generalization capabilities for downstream tasks. Pre-trained models can be called generalized large models.

[0049] SFT: Supervised Finetuning, supervised fine-tuning, is to fine-tune a pre-trained model based on a large pre-trained model for a task in a specific scenario. During the fine-tuning process, the pre-trained model will adjust the parameters and structure of the pre-trained model according to the characteristics of the task to improve the performance of the pre-trained model on the task.

[0050] LoRA: A model Fine Tuning method that freezes the model parameters of the pre-trained model and injects a trainable low-rank matrix into the attention layer of the Transformer architecture (a model that uses the attention mechanism to increase the model training speed), greatly reducing the number of trainable parameters for downstream tasks and increasing the training speed.

[0051] SHAP: Shapley Value, a method for calculating the causality between features and labels. By enumerating different feature combinations, it calculates the impact of adding a feature to a certain feature combination on the prediction result compared to not adding this feature. This impact is the marginal benefit of this feature for this feature combination.

[0052] Tars: ICUB large language model, a deep learning model trained based on massive text data.

[0053] ab: Inquiry, a question asked by a buyer or seller to the other party about the terms of the transaction in order to buy or sell a certain commodity.

[0054] rfq: Request for quotation.

[0055] In order to facilitate the understanding of the technical solution, the underlying technical architecture on which each embodiment of the present application is based is described. The technical solutions provided in the following embodiments are all implemented based on this technical architecture. Figure 1 As shown in the figure, the technical architecture diagram of the technical solution provided by this application. Figure 1 As shown, the technical architecture of the technical solution provided by this application includes:

[0056] 1. Build a Query template library ( Figure 1 Winning bid number 101)

[0057] In an achievable technical solution, building a Query template library may include:

[0058] S11, determining multiple basic Query templates;

[0059] S12, using the second preset model, expanding the multiple basic Query templates to obtain multiple extended Query templates;

[0060] S13, evaluating the multiple extended Query templates, and retaining the extended Query templates that pass the evaluation;

[0061] The Query template library includes a plurality of basic Query templates and a plurality of extended Query templates that have passed evaluation.

[0062] For example, the multiple basic query templates in S11 above may include but are not limited to: multiple basic query templates covering all traffic nodes, multiple basic query templates related to store optimization items, multiple basic query templates related to store operation problems, etc. Among them, in the e-commerce scenario, traffic nodes may include but are not limited to: keyword search, user access, adding to shopping cart, inquiry, ordering, payment, etc. Store optimization items may include but are not limited to: keywords of goods in the store, pictures of goods in the store, promoted goods in the store, applicable groups of goods, etc. Store operation problems may include but are not limited to: how to increase the chance of being searched, how to increase the payment amount, how to increase the store visit volume, etc. The store mentioned in this article can be an online store on an e-commerce platform or an online store on a social platform, etc., and an online store can also be called an "online store" or "electronic store". E-commerce platforms or social platforms provide merchants with technical support such as operating stores on the Internet, displaying and selling goods or services in online stores. E-commerce platforms or social platforms use Internet technology to allow consumers to browse product information, make comparisons, place orders, pay for goods, and wait for the goods to be delivered to the designated address.

[0063] Among them, multiple basic Query templates covering all traffic nodes can be divided into two major scenarios. The first category is multiple basic Query templates corresponding to the difference between the store's own traffic data and the traffic data of reference stores (such as stores in the same industry and / or stores of the same level); the second category is multiple basic Query templates corresponding to the difference between the store's traffic data in the first period and the traffic data in the second period in history.

[0064] After having multiple basic query templates, you can also use the second preset model to expand multiple extended query templates offline. The second preset model can be a large language model (LLM). The large language model is a deep learning model trained based on massive text data. It can not only generate natural language text, but also deeply understand the meaning of the text and handle various natural language tasks.

[0065] The multiple extended queries extended by the second preset model may not all meet the requirements. Therefore, the multiple extended queries output by the second preset model can be evaluated. The evaluation process may include but is not limited to: obtaining evaluation rules; using the evaluation rules to evaluate each extended query separately, and retaining the extended query templates that pass the evaluation. Furthermore, a manual participation interface can also be added to the evaluation process, such as displaying the extended query templates that pass the evaluation on the client interface, and users can also edit, delete, etc. the extended query templates that pass the evaluation.

[0066] Among them, this application does not make any specific restrictions on the evaluation rules, and may include multiple rule items, each of which can be manually set.

[0067] Finally, a query template library containing multiple basic query templates and multiple extended query templates that have passed the evaluation is obtained. When diagnosing a store, multiple query templates in the query template library can be sorted according to the store's diagnosis results (such as low store exposure, low click-through rate, low inquiry volume, etc.). For example, query templates that are closely related to the store's diagnosis results are ranked in front. In this way, the top few, such as 1, 2, 3 or 4, etc., can be used as query templates corresponding to the store's diagnosis results. Alternatively, through big data statistics, the common problems of stores in the same industry or at the same level are counted, and multiple query templates in the query template library are sorted based on the statistical results. In this way, the top few, such as 1, 2, 3 or 4, etc., can be used as recommended query templates.

[0068] In summary, the Query template library may include at least one of the following:

[0069] Query template corresponding to the reason why the data of any store on the platform differs from that of the reference store;

[0070] Query templates corresponding to the reasons for differences in data at different times for any store on the platform;

[0071] Query template corresponding to store management questions;

[0072] Query template corresponding to store optimization items;

[0073] A query template expanded using the second preset model.

[0074] 2. Build a copywriting database ( Figure 1 Winning bid number 102)

[0075] 2.1. Collect basic data, such as the attribute information of the stores on the platform, the traffic data of the stores, the operation data of the stores, etc. Among them, the attribute information of the stores may include but is not limited to: the industry to which the store belongs, the store level, etc. The traffic data of the store may include but is not limited to: the exposure rate, click-through rate, inquiry volume, average daily number of orders, etc. of the store's historical period (such as the past 30 days, 60 days, etc.). The operation data of the store may include but is not limited to: the number of store windows, the number of high-quality products, the number of self-marketing interactions, the amount of RFD, the amount of advertising consumption, the store traffic situation (including the number of people entering the store through different channels such as search, activities, paid advertising, live broadcasts, etc.), etc.

[0076] 2.2. Store optimization attribution

[0077] Causal inference is a statistical and reasoning method that aims to understand and infer the cause-effect relationship between events, variables or behaviors. This application plan is to use causal inference technology to analyze abnormal traffic nodes that cause store problems. These traffic nodes are store optimization items that can be optimized. Funnel analysis technology is used in attribution analysis. That is, funnel analysis technology is used to analyze the traffic data and operation data of stores on the platform to analyze store optimization items that solve or improve store problems.

[0078] The data funnel node is a key element in funnel analysis, representing the various traffic nodes experienced from the beginning of an event to the end of the event. These traffic nodes include the starting point, intermediate process nodes, and the end point. Specifically, the data funnel node settings will be different for different application scenarios. In the field of e-commerce, data funnel nodes may include: customer visits, adding to shopping carts, placing orders, payments, and so on. By analyzing the data of these nodes, you can understand the customer conversion situation at each stage, so as to find potential problems and optimization items. For example, if you find that the conversion rate of a node is low, it may mean that there is a problem with the node and further adjustment and optimization are needed. This node is also the store optimization item that the store needs to optimize.

[0079] Therefore, when performing funnel analysis, it is necessary to carefully determine and define the nodes of the data funnel to ensure the accuracy and effectiveness of the analysis. The embodiment of the present application does not specifically limit the determination and definition of the nodes of the data funnel, and specific analysis is required for specific scenarios. In specific practice, it can be determined and defined based on experience.

[0080] When performing funnel analysis, you can use the monotone stack method to perform attribution analysis of node optimization points in the traffic funnel. The monotone stack is often used to find the position of the first element on the left or right of an element in an array that is larger or smaller than it, thereby identifying the local maximum and minimum values ​​in the sequence.

[0081] For example, in the specific implementation, two increasing monotonic stacks can be maintained to find the node where the index (visit volume, number of inquiries, etc.) associated with the questions asked by the merchant or the problems diagnosed by the merchant's store is the most different from that of the reference store (another store in the same industry or the same level as the merchant's store), that is, the position of the previous minimum value of each position in an array. By using two monotonic increasing stacks, the abnormal node position can be effectively found, and this position is the key point of optimization, that is, the store optimization item.

[0082] 2.3 Factor ranking of store optimization items

[0083] Modeling is performed with store optimization items as the target, and a causal tree model is constructed based on multiple factors. The causal tree model can be a LightGBM model (LGB model for short). In specific implementation, multiple factors can be selected from multiple search ranking features in e-commerce search scenarios. In specific implementation, the causal tree model can be trained based on a preset training set. Among them, the preset training set can be constructed based on the store data of the stores on the platform. For factors that cannot be changed, such as store types and statistical characteristics, they will be deleted during training to avoid interfering with the split nodes of the causal tree model and affecting the feature importance effect.

[0084] A data set is constructed based on the store data of the stores on the platform (such as store traffic data and store operation data). This data set is different from the above training set, and the data set can be called a prediction set. After the causal tree model is trained with the training set, the data set is used to calculate the influence of each factor in the multiple factors on the prediction results of the causal tree model, and the influence degree corresponding to each of the multiple factors is obtained.

[0085] Specifically, the above-mentioned calculated influence can be specifically a SHAP value (Shapley value). SHAP (Shapley AdditiveexPlanations) is a "model explanation" package that can explain the output of any machine learning model. All features are considered "contributors". For each prediction sample, the model generates a prediction value, and the SHAP value is the numerical value assigned to each feature in the sample. In this application, the SHAP value is the influence or contribution of each factor on the prediction result output by the causal tree model. To calculate the SHAP value, the overall idea is to enumerate different factor combinations and calculate the impact of adding a factor under a certain factor combination relative to not adding this factor on the prediction result. This impact is the marginal benefit of this factor for this factor combination. The formula is as follows:

[0086]

[0087] Among them, V(i): represents the Shapley value of factor i among multiple factors; N: represents the set of multiple factors; |N|: represents the number of multiple factors; S: is a subset of N; |S|: is the number of subsets of N; v(S) represents the value of multiple factors in the S set.

[0088] In the calculation of the Shapley value, all possible factor combinations S are traversed, and the additional contribution of factor i after adding it to combination S is calculated, that is, v(S∪*i+)-v(S). Then, the additional contributions of all combinations are weighted averaged, where the weight is determined according to the size of the combination.

[0089] The above formula means that the Shapley value is equal to the average contribution of a factor in all possible cooperative combinations. It takes into account the marginal contribution of each factor and determines the final distribution through weighted average. The Shapley value of a factor is defined as the contribution of the factor to the prediction results of the causal tree model. When there are many factors, the approximate value of the Shapley value can be calculated by factor sampling or factor independence assumption.

[0090] Because the store optimization item is used as the target modeling, a causal tree model is constructed based on multiple factors; and then the SHAP value of each factor in the multiple factors is calculated using the above algorithm. Therefore, after obtaining the SHAP values ​​of multiple factors, the multiple factors can be sorted according to the SHAP values; and then the factor ranking of the store optimization item is obtained.

[0091] There are many store optimization items, such as whether the set channels are used to increase the click-through rate, whether the keywords are accurate, etc. These are all store optimization items. Through the above technical means, the factor ranking of each store optimization item can be obtained.

[0092] In fact, there are differences in store operations for stores in different industries, stores at different levels (rated according to the platform rating rules), and stores at different stages (such as the growth stage or the stage of preparing for a level upgrade, etc.). In other words, the ranking of factors for the same store optimization item will be different for stores in different industries, levels, or stages. For example, stores on the platform are grouped according to the first and second level industries to which the store belongs, the level to which it belongs, whether it has advertising, whether it is in the growth stage, whether it is in the stage of preparing for a level upgrade, etc.

[0093] The primary industry and the secondary industry are two levels of industry classification, which constitute a classification system from broad to specific. The primary industry is the most basic level in the industry classification, which can include agriculture, forestry, animal husbandry, fishery, industry, construction, etc. The secondary industry is a further subdivided industry category based on these primary industries.

[0094] Therefore, this application proposes the concept of grouping the stores on the platform, and calculating the factor ranking of the corresponding store optimization items for the stores in different groups. That is, in the above technical solution, when constructing the data set, it will be constructed based on the store data belonging to the same group (such as store traffic data and store operation data). Then, using the data set constructed by the store data of the stores in the same group, the SHAP value of each factor is calculated, and then the factor ranking obtained based on the SHAP value ranking is the factor ranking of the store optimization items in the group. In short, using the above technical solution, for the unified store optimization items, the factor ranking of the store optimization items corresponding to different groups will be calculated. For ease of understanding, it can be expressed as the following Table 1:

[0095]

[0096] 2.4 Answer Database

[0097] The suggested content associated with the factor configuration is respectively configured, and the suggested content can be a task that can be performed by the store (such as changing keywords, placing advertisements, using auxiliary channels, etc.). Furthermore, copy templates under different Query templates can be constructed. For example, for service, advertising or transaction tasks, the copy template can specify that in addition to the suggested content, it is also necessary to provide a comparison of the target store and the reference store data, etc. For commodity and sub-account tasks, the copy template can specify that in addition to the suggested content, it also provides a list card of the optimizable objects (optimizable target commodities or sub-accounts, etc.), and so on.

[0098] In summary, the copywriting database can store: attribute information of stores on the platform, traffic data and operation data of stores on the platform, store optimization items associated with query templates, factor ranking of store optimization items, recommended content associated with factors, and a corpus knowledge base of copywriting template information corresponding to different query templates.

[0099] 3. Model Deployment Figure 1 Winning bid number 103)

[0100] 3.1. Generate training samples

[0101] Based on the document database obtained in the above part (ii) and the query template library obtained in the above part (i), call the third preset model to generate multiple training samples. Among them, the third preset model can be a large language model (LLM, LargeLanguage Model), etc., and this embodiment does not make specific limitations on this. For example, call the third preset model to generate 4000 or more (such as 5000, 6000, etc.) valid data offline as training samples. Among them, some training samples contain sensitive data (such as merchant privacy information), and the sensitive information needs to be desensitized, such as using a mask. <mask>At the same time, the prompt word can be adjusted so that the text output by the first preset model is in a specified format, such as JSON format (JavaScript Object Notation, a lightweight data exchange format), to facilitate the access between subsequent groups and jump connections.

[0102] For example, clearly state in the prompt that the model is expected to return the copy in JSON format. For example, provide a simple JSON format example. An example JSON output can be like this: {"Question": "What is the target question?", "Answer": "This is the corresponding copy."} Please provide your answer in this format.

[0103] 3.2 Model Adjustment Training

[0104] Using the training samples generated in 3.1 above, a fine-tuning method (such as the LoRA method) can be used to train the Fine Tuning pre-trained model to enhance the logic and accuracy of the output of the pre-trained model in the store operation scenario. The advantage of LoRA is that the LoRA method only trains a small number of parameters of the pre-trained model, and by adding a low-rank matrix to the attention layer of the Transformer, the model estimation accuracy is improved.

[0105] Among them, the pre-trained model can be a large language model (LLM). The pre-trained model has learned rich feature representation and pattern recognition capabilities. Through fine-tuning, these learned feature representations can be used to quickly and effectively adapt to tasks in specific scenarios (such as tasks in store operation scenarios). Fine-tuning is faster than training the model from scratch, which plays a role in accelerating the training process.

[0106] 4. Online Link Construction ( Figure 1 104 of 104)

[0107] In the specific implementation, online deployment can be completed based on Rlab (a high-level programming language for engineering application calculation) tools. Of course, online deployment can also be completed based on other tools. This application does not make specific restrictions on this. Here, only the use of Rlab tools is exemplified. Online link construction also requires the following but not limited to: EAS, Igraph, pre-trained model development tools (Tars Studio), PAI and Itag tool support. Among them, EAS (Elastic Algorithm Service, model online service) is a machine learning platform product to achieve one-stop model development and deployment applications, and provides model online services for online reasoning scenarios. It supports the deployment of model services in public resource groups or exclusive resource groups, and realizes real-time response to model loading and data requests based on heterogeneous hardware (such as CPU or GPU). Igraph is an online graph storage and query system that provides storage, query, update and calculation services for large-scale graph data. The pre-trained model development tool (Tars Studio) provides a visual development tool. PAI is a machine learning platform that allows developers to use artificial intelligence AI technology more efficiently, concisely and standardly. Itag provides data annotation services to help users complete data annotation tasks quickly and efficiently.

[0108] The links built online may include but are not limited to:

[0109] S21. Determine the target Query.

[0110] S22. Rewrite the target Query.

[0111] The target query rewriting may include but is not limited to:

[0112] Query normalization: rewrite the target query through simple string processing, such as removing spaces, converting uppercase and lowercase letters, processing special characters, etc.

[0113] Query error correction: Actively correct target queries with spelling errors;

[0114] Query expansion: expand the query according to the semantics or intent of the target query, find a set of queries that are consistent with the target query, and search together with the target query;

[0115] etc.

[0116] S23. Acquire data based on Igraph.

[0117] For example, the acquired data may be data retrieved from a copywriting database. The data may include, but are not limited to, store data of the target store (such as traffic data and operation data of the store), store data of the reference store (such as traffic data and operation data of the reference store), and copywriting information. The copywriting information may include, but are not limited to, copywriting templates, query templates matching the target query after the rewriting process in S22, store optimization items associated with the matching query, at least one factor of the store optimization item, recommended content associated with the factor, and the like.

[0118] S24: pre-process the acquired data.

[0119] The present application does not specifically limit "preprocessing", which may be, but is not limited to, operations such as data cleaning, standardization, and format conversion. For example, useful data is filtered out from the acquired data through the data cleaning process. For example, unstructured data (such as text) may need to be structured through steps such as word segmentation and stem extraction. According to the input requirements of the pre-trained model, the data needs to be converted into an appropriate format, such as JSON format or text format.

[0120] S25, Prompt splicing

[0121] The matched Query template and the data after the above preprocessing can be spliced ​​into Prompt. The output (or response) quality of the first preset model (such as the first preset model can be a fine-tuned large language model) depends on the prompt word (ie Prompt). Simply put, Prompt is the question the user wants to ask, which is sent as input to the first preset model. The first preset model will try to understand this input and then output a suitable answer or response. Splicing Prompt here is to make the first preset model output more accurate and useful copy.

[0122] Prompt is an important part. Prompt is used to guide the behavior of the model. By adjusting Prompt, the quality of the model output copy can be improved. In the specific implementation, the designer can manually design the prompt splicing rules so that the splicing can be performed according to the preset splicing rules when executing prompt splicing.

[0123] 5. Front-end Function Deployment

[0124] With the technical architecture support of the above parts one, two, three and four, the corresponding functional interfaces can be deployed on the front-end application (APP) or browser web page, so that users can call the corresponding services of the back-end (such as the server with the above technical architecture) through the functional interface, such as store diagnosis services, intelligent question and answer services, etc.

[0125] The functional interface can be in the form of a functional control on the APP page or browser webpage, and the user can call the corresponding backend service by clicking the functional control. Alternatively, a specific logo is provided on the APP page or browser webpage, such as a smart assistant logo, which can be any pattern or animation, and is not limited here. The user can wake up the smart assistant function through voice or gesture, and the smart assistant logo changes the display style to let the user know that the smart assistant has been successfully woken up. The user can call the corresponding backend service by interacting with the smart assistant. That is, the product form presented by the above technical architecture on the client side can be a smart assistant logo.

[0126] The store diagnosis service may include: diagnosing the store, and providing store optimization items that are strongly related to the problem when the store has a problem. Furthermore, the store diagnosis service may also provide action suggestions for improving the store optimization items, etc.

[0127] The intelligent question-and-answer service may include: responding to behavior suggestion questions input by users, outputting action strategies and goals corresponding to the behavior suggestion questions, etc.

[0128] In summary, the technical solution provided by this application has at least the following technical effects:

[0129] 1. The technical solution provided in this application can rewrite and polish the output copy to achieve content personalization by adjusting and training the pre-trained model to be suitable for downstream tasks (such as intelligent response tasks in store operation scenarios).

[0130] 2. Using multiple factors as the starting point, establish causal tree models corresponding to different store optimization items, and then determine the factor ranking of different store optimization items. Factors are associated with suggested content, such as suggested content for tasks that merchants can perform.

[0131] 3. Because there are differences between different industries and different levels of stores, for the same problem (for example, how to increase the number of visits), the factor rankings of the same store optimization items corresponding to stores in different industries are different. For example, when solving the problem of visit volume for two stores in the machinery and equipment industry and the clothing industry, their corresponding factor rankings are different. Therefore, the technical solution of this application groups the stores on the platform; when determining the factor rankings of different store optimization items, the factor rankings of different store optimization items under different groups are determined according to the groups.

[0132] The technical solution provided by this application combines the pre-trained model with causal inference technology to realize the implementation of e-commerce scenarios. Specifically, the solution of this application constructs a Query template library suitable for store operation scenarios; and also constructs a copywriting database suitable for store operation scenarios based on the Query template library. In addition, when constructing the copywriting database, causal inference technology is also used to attribute and analyze at least one factor of each store operation-related problem, wherein the factor is associated with recommended content. In this way, a training sample set can be generated based on the Query template library and the copywriting database, and the training sample set can be used to adjust the pre-trained model for store operation scenarios, thereby enhancing the logic and accuracy of the output copy in the store operation scenario.

[0133] The technical solutions corresponding to the technical architecture mentioned above (such as Figure 1 The solutions corresponding to parts one, two, three and four in the above) can be deployed on the server side. The server side can be a server, a service cluster, a virtual server or a cloud, etc., which is not specifically limited in this embodiment. The server side provides corresponding functional services to the client side, such as store diagnosis, intelligent question and answer services, etc. Users can enter the management page provided by the platform to users through the browser or client application (APP) on the client device. The client device can be but not limited to: smart phones, smart wearable devices, tablet computers, laptops, desktop computers, etc. Figure 2 As shown, an embodiment of the present application provides a service system, which includes a server and a client.

[0134] The server 11 is used to build a query template library, the query template library includes multiple query templates, and each query template is associated with a preset question related to store operation; based on the query template library, a copy database is built; wherein the copy database stores: at least one factor of the query template is obtained by performing attribution analysis on the preset question associated with the query template using a causal inference algorithm; a training sample set is generated according to the query template library and the copy database; and the pre-trained model is adjusted and trained using the training sample set to obtain a first preset model that can handle intelligent response tasks in store operation scenarios.

[0135] The client 12 is used to deploy a corresponding interface so that the user can call the server through the interface, retrieve the target copy material in the copy database, and output the copy corresponding to the copy material using the first preset model.

[0136] Furthermore, the service system includes a server 11 and a client 12 which are also used to implement the following functions. Specifically:

[0137] The client 12 is further configured to respond to a diagnosis trigger event for a target store, trigger the store diagnosis to present and / or play the diagnosed target problem, and send information about the target problem to the server; or display the management page of the target store; monitor the user's operation on the management page to capture the user's target problem in store operation; respond to a confirmation instruction triggered by the user for the target problem, and send information about the target problem to the server;

[0138] The server 11 is further configured to determine a target problem for a target store; retrieve at least one factor of the target problem from a copy database; generate a prompt word Prompt based on the target problem and the suggestion content associated with the at least one factor; and output a copy corresponding to the Prompt using a first preset model to help the user solve or improve the target problem;

[0139] The client 12 is further used to receive the text fed back by the server, and to display and / or play the text.

[0140] It should be added here that the executor of the above-mentioned "store diagnosis" can be the client 12. After the store diagnosis is triggered, the client 12 can obtain the store data of the reference store and / or the historical store data of the target store from the server 11, etc., and then call the local diagnostic program of the client 12 to perform store diagnosis on the target store. Alternatively, the executor of "store diagnosis" is the server 11. After the client 12 triggers the store diagnosis, the client 12 sends a target store diagnosis request to the server 11, so that the server 11 obtains the current store data, historical store data, store data of the reference store, etc. of the target store, and diagnoses the target store based on the acquired data to determine the target problem. The server 11 feeds back the determined target store to the client 12.

[0141] It can be seen that the service system provided in the embodiment of the present application utilizes the powerful computing power of the server 11 to build a Query template library and use causal inference technology to find at least one factor that causes store problems; it also uses a large model (i.e., the first preset model) to achieve the logic and accuracy of the output copy in the store operation scenario, so as to help users (or store merchants) grow.

[0142] It is important to note that the above is an introduction to the technical solution of this application using a store operation scenario as an example. In fact, the technical architecture provided by this application can also be applied to other specific scenarios, such as enterprise management scenarios, project management scenarios, etc.

[0143] The technical solution provided in this application will be described below in the form of a method embodiment.

[0144] Figure 3 The flowchart of the information processing method provided by an embodiment of the present application is shown. The execution subject of the method provided by this embodiment can be the server in the above service system. Figure 3 , the method may include:

[0145] 301. Determine target issues for target stores.

[0146] 302. In a document database, retrieve at least one factor of the target problem.

[0147] 303. Generate a prompt word Prompt based on the target question and the suggestion content associated with the at least one factor.

[0148] 304. Output the text corresponding to the Prompt using the first preset model to help the user solve or improve the target problem.

[0149] Among them, the copywriting database stores: at least one factor of the preset problem obtained by attribution analysis of the preset problem using a causal inference algorithm; the first preset model is obtained by adjusting and training the pre-trained model based on the pre-trained model for the intelligent answering task in the store operation scenario.

[0150] The above technical architecture lists an example of building a query template library. In fact, the query template corresponds to the question. In other words, the query template can be directly understood as a corresponding question template. In other words, in the specific implementation, the above construction of the query template library can also be "constructing a question library".

[0151] When determining the target problem of the target store in 301, the target problem of the target store may be determined by diagnosing the target store, or the target problem may be determined based on user operations. Determining the target problem based on operations may include but is not limited to: determining the target problem of the target store based on information input by the user, or determining the target problem of the target store based on a request triggered by a setting item by the user, or the network side recommending at least one recommended problem based on big data analysis, and the user operates to select one of them, and the selected recommended problem is the target problem.

[0152] That is, the above 301 "target store, determine the target problem" may include:

[0153] 3011. In response to user input, determine a target question for the target store based on the user input information; and / or

[0154] 3012. In response to a request triggered by a user for a setting item on a management page of a target store, determining the target problem for the setting item; and / or

[0155] 3013. Diagnose the target store and determine the target problem based on the diagnosis result; and / or

[0156] 3014. Obtain at least one recommended question recommended by the network side based on big data analysis, and in response to a selection event, use the selected recommended question as the target question.

[0157] In a specific implementation, the above 3013 “diagnosing the target store and determining the target problem based on the diagnosis result” may include:

[0158] Diagnose the gap between the target store and the reference store, and / or the data difference between the target store in different periods, and / or the gap between the store data of the target store and the set indicators, and obtain a diagnosis result;

[0159] The target problem is determined based on the diagnosis result.

[0160] The reference store can be: a store that belongs to the same industry as the target store and whose store data is in the median or high number of all stores in the industry. Alternatively, the reference store is: a store that belongs to the same industry as the target store, the same level, and whose store data is in the median or high number of all stores in the industry. Alternatively, the reference store is: a store that belongs to the same group as the target store, the same level, and whose store data is in the median or high number of all stores in the industry. Alternatively, the reference store is: a store specified by the user, etc.

[0161] The above-mentioned "data differences exist in different periods of the target store" may be: data differences exist between the store data of the first period and the store data of the second period of the target store. The second period may not overlap with the first period, and the second period is earlier than the first period. Alternatively, the second period partially overlaps with the first period, the start time of the second period is earlier than the start time of the first period, and the end time of the second period is earlier than the end time of the first period. Alternatively, the second period includes the first period. That is, the start time of the first period is a certain time within the second period, and the end time of the first period is the same as the end time of the second period.

[0162] The above-mentioned "set indicator" can be: a target indicator set by the user. For example, the user expects the store visit volume to increase by 10% in the future, and this 10% increase is the set indicator. Alternatively, the above-mentioned "set indicator" can also be: an indicator of a reference store. The reference store is as mentioned above.

[0163] Indicators may include store indicators and product indicators. Among them, store indicators include but are not limited to: store page views, conversion rate, fan activity, member activity, average customer spending (the average amount of money a user who purchases a product in the store spends on a product), etc. Conversion rates include: search conversion rate, UV (unique visitor, referring to people who visit and browse store pages through the Internet) order conversion rate, etc. Product indicators may include but are not limited to: product attention, product add-to-cart rate, product conversion rate, etc.

[0164] As can be seen from the above content, the document database stores: at least one factor of the query template obtained by performing attribution analysis on the preset question associated with the query template using a causal inference algorithm. Accordingly, the above step 302 "retrieve at least one factor of the target question in the document database" may include the following steps:

[0165] 3021. Determine a target Query according to the target question;

[0166] 3022. Retrieve a query template matching the target query from the document database, and obtain at least one factor of the matching query template;

[0167] When the Prompt is generated (ie, the above step 303 ), suggestion content is generated based on the association between the target Query and at least one factor of the matched Query template.

[0168] As mentioned above, a Query template library is pre-built, and the Query template library includes multiple Query templates. Accordingly, one possible implementation of the above step 3021 "determine the target Query according to the target question" is:

[0169] According to the target question, multiple Query templates in the Query template library are sorted;

[0170] The target query is determined based on the first ranked query template, or the first number of query templates ranked first are used as candidate queries for the user to select, and the target query is determined based on a candidate query selected by the user.

[0171] During specific implementation, the first number of items in the front ranking may be: 2, 3, or 5 items in the front ranking, etc., which is not specifically limited in this embodiment.

[0172] Furthermore, the document database stores: store optimization items that can solve or improve the preset problems and at least one factor of the store optimization items, obtained by performing attribution analysis on the preset problems associated with the query template using a causal inference algorithm. Accordingly, the above step 3022 "retrieve the query template matching the target query in the document database, and obtain at least one factor of the matching query template" may include:

[0173] Retrieving a query template matching the target query in the document database;

[0174] Acquire a store optimization item associated with the matched Query template and at least one factor of the store optimization item;

[0175] When the Prompt is generated (ie, the above step 303 ), suggestion content is generated based on the target Query, the store optimization item associated with the matched Query template, and at least one factor associated with the store optimization item.

[0176] When constructing the copy database, the present application scheme adopts a causal inference algorithm. That is, in the scheme provided by the embodiment of the present application, the following steps are included when constructing the copy database: using a causal inference algorithm to perform attribution analysis on the preset problem associated with the Query template, and obtain store optimization items that solve or improve the preset problem and at least one factor of the store optimization item. This part is an important content of the embodiment of the present application. In an achievable manner, the step of "using a causal inference algorithm to perform attribution analysis on the preset problem associated with the Query template, and obtain store optimization items that solve or improve the preset problem and at least one factor of the store optimization item" may include the following steps:

[0177] 305. Analyze the traffic data and operation data of the stores on the platform using funnel analysis technology for the preset problems associated with the query template, so as to analyze the store optimization items that solve or improve the preset problems;

[0178] 306. Modeling the store optimization item as a target, and constructing a causal tree model based on multiple factors;

[0179] 307. Build a data set based on the traffic data and operation data of the stores on the platform;

[0180] 308. Using the data set, calculate the influence of each of the multiple factors on the prediction result of the causal tree model to obtain the influence degree corresponding to each of the multiple factors;

[0181] 309. Sort the multiple factors according to the influences corresponding to the multiple factors to obtain the factor ranking of the store optimization item;

[0182] Among them, at least one factor of the store optimization item is: the factors ranked in the top second number in the factor ranking of the store optimization item.

[0183] For the specific implementation content of the above steps 305 to 309, please refer to the above description and will not be repeated here.

[0184] Furthermore, considering the differences between stores of different industries and different levels, the factor rankings (factor priorities) corresponding to the same store optimization item are different; the method provided in this embodiment also groups the stores on the platform to determine the factor rankings of store optimization items for different groups. That is, the method provided in this embodiment of the application may also include the following steps:

[0185] 310. Group the stores on the platform;

[0186] 311. When constructing the data set, the data set is constructed based on the traffic data and operation data of the stores in a group; and

[0187] 312 , using the data set to obtain the factor ranking corresponding to the store optimization item, to obtain the factor ranking of the store optimization item corresponding to the group;

[0188] Among them, when "obtaining the store optimization items associated with the matched Query template, and at least one factor of the store optimization items" (i.e., a lower step included in step 3022), determine the group to which the target store belongs; obtain the store optimization items associated with the matched Query template, and at least one factor of the store optimization items corresponding to the group to which the target store belongs.

[0189] The method provided in this embodiment is to deploy Figure 1 Based on the technical architecture shown, the article is written from the perspective of how to apply the support provided by the technical architecture to help users solve or improve store problems. In fact, the method provided in the embodiment of the present application also includes the following steps, which belong to the preliminary deployment tasks. Specifically, the method provided in this embodiment may also include the following steps:

[0190] 313. Build a Query template library, where the Query template library includes multiple Query templates;

[0191] 314. Building the document database based on the Query template library;

[0192] 315. Generate the training sample set according to the Query template library and the document database;

[0193] 316. Use the training sample set to adjust and train the pre-trained model.

[0194] For the specific implementation of the above steps 313 to 316, please refer to the above description, which will not be repeated here.

[0195] In an achievable embodiment, the Query template library includes at least one of the following: a Query template corresponding to the reason for the difference in data between any store on the platform and a reference store, a Query template corresponding to the reason for the difference in data between any store on the platform at different periods, a Query template corresponding to store management issues, a Query template corresponding to store optimization items, and a Query template expanded using a second preset model.

[0196] The copywriting database stores: attribute information of stores on the platform, traffic data and operation data of stores on the platform, store optimization items associated with query templates, factor ranking of store optimization items, recommended content associated with factors, and a corpus knowledge base of copywriting template information corresponding to different query templates.

[0197] The technical solution provided in this embodiment combines a large model with causal inference technology, that is, a causal inference algorithm is used when constructing a copywriting database to perform attribution analysis on the preset questions associated with the Query template to obtain at least one factor; the training samples used to adjust and train the pre-trained model are generated based on the Query template library and the copywriting database. Using a training sample set composed of multiple such training samples, the pre-trained model is adjusted and trained to adjust it to a first preset model suitable for store operation scenarios. The copywriting database provides data support, which can retrieve at least one factor of the store diagnosis results and / or the questions asked by the user, and the recommended content associated with the factor; the adjusted and trained first preset model provides intelligent response support. Based on the store diagnosis results and / or the questions asked by the user and the recommended content associated with at least one factor, the first preset model can output copywriting that is more in line with natural language logic and suitable for store operation scenarios to help users operate stores.

[0198] Another embodiment of the present application also provides an information processing method, more specifically a model training method. The execution subject of the method provided in the embodiment of the present application can be the server in the above service system. Specifically, Figure 4 As shown, the method includes:

[0199] 401. Construct a query template library, wherein the query template library includes multiple query templates, and each query template is associated with a preset question related to store operation.

[0200] 402. Construct a document database based on the Query template library; wherein the document database stores: at least one factor of the Query template obtained by performing attribution analysis on a preset question associated with the Query template using a causal inference algorithm.

[0201] 403. Generate a training sample set according to the Query template library and the document database.

[0202] 404. Use the training sample set to adjust and train the pre-trained model to obtain a first preset model that can handle intelligent response tasks in a store operation scenario.

[0203] When constructing the document database in step 402, the causal inference algorithm is used to perform attribution analysis on the preset questions associated with the query template to obtain at least one factor of the query template, which may include:

[0204] 4021. Analyze the traffic data and operation data of the stores on the platform using funnel analysis technology for the preset problems associated with the query template, so as to analyze the store optimization items that solve or improve the preset problems;

[0205] 4022. Modeling the store optimization item as a target, and constructing a causal tree model based on multiple factors;

[0206] 4023. Build a data set based on the traffic data and operation data of the stores on the platform;

[0207] 4024. Using the data set, calculate the influence of each of the multiple factors on the prediction result of the causal tree model to obtain the influence degree corresponding to each of the multiple factors;

[0208] 4025. Sort the multiple factors according to the influences corresponding to each of the multiple factors to obtain a factor ranking of the store optimization item.

[0209] Furthermore, the method provided in this embodiment may also include the following steps:

[0210] 405. Group the stores on the platform.

[0211] 406. When constructing the data set, it is constructed based on the traffic data and operation data of the stores in a group.

[0212] 407. The factor ranking of the store optimization items obtained by using the data set is the factor ranking of the store optimization items corresponding to the group.

[0213] In a specific embodiment, the above step 401 "building a document database based on the Query template library" may include:

[0214] Obtain and store attribute information of stores on the platform;

[0215] Obtain and store traffic data and operation data of stores on the platform;

[0216] Performing attribution analysis on preset questions associated with the Query templates in the Query template library using a causal inference algorithm to obtain at least one factor of the Query template;

[0217] determining and storing suggested content associated with the factor;

[0218] Construct a corpus knowledge base of copywriting template information corresponding to different query templates.

[0219] In a specific embodiment, the above step 403 "generating a training sample set according to the Query template library and the document database" may include:

[0220] Outputting a plurality of training samples based on the Query template library and the document database using a third preset model;

[0221] The training sample set includes the multiple training samples.

[0222] The above step 401 "building a query template library" may include:

[0223] 4011. Determine multiple basic Query templates;

[0224] 4012. Using the second preset model, the multiple basic Query templates are extended to obtain multiple extended Query templates;

[0225] 4013. Evaluate the multiple extended Query templates and retain the extended Query templates that pass the evaluation;

[0226] The Query template library includes a plurality of basic Query templates and a plurality of extended Query templates that have passed evaluation.

[0227] For the specific content of each step in this embodiment, please refer to the above description and will not be repeated here.

[0228] The present application also provides a store diagnosis method in one embodiment. The execution subject of the method provided in this embodiment may be a client in the above service system. Figure 5 As shown, the method includes:

[0229] 501. In response to a diagnosis triggering event for a target store, triggering store diagnosis to present and / or play a diagnosed target problem;

[0230] 502. Display and / or play a copy, where the copy includes suggestion content that helps the user solve or improve the target problem;

[0231] Among them, the copy is obtained by using a first preset model, which is obtained by adjusting and training the pre-trained model for the intelligent answering task in the store operation scenario on the basis of the pre-trained model; the suggested content is associated with at least one factor of the target problem, and the at least one factor is obtained by attribution analysis of the target problem using a causal inference algorithm.

[0232] See also Figure 6a-6c As shown, the above diagnostic triggering events for the target store may include:

[0233] In response to a user's operation of a diagnostic control on the interface, generating a diagnostic trigger event; and / or

[0234] Responding to the user's store diagnosis voice, generating a diagnosis trigger event; and / or

[0235] In response to the store diagnosis instruction input by the user through the intelligent assistant interaction interface, a diagnosis trigger event is generated.

[0236] See also Figure 6a The interface example shown may be a store management interface, on which a diagnostic control is displayed, such as Figure 6a The "Store Diagnosis" shown, users can trigger the diagnosis by clicking on the "Store Diagnosis" control.

[0237] See also Figure 6b In the example interface shown, the user can first wake up the smart assistant by voice (such as "Hello, smart assistant"). At this time, the smart assistant logo 1 can be displayed on the interface, and then the user can issue a "store diagnosis" voice to trigger the diagnosis. The smart assistant logo can be any shape or pattern, which is not limited in this embodiment.

[0238] See also Figure 6c In the example interface shown, a smart assistant interaction interface 2 is displayed on the display screen, and an input box 3 is displayed on the smart interaction interface 2. The user can enter "store diagnosis" in the input box 3 and then click the confirmation or send control 4 (such as Figure 6c In ) to trigger store diagnosis.

[0239] The executor of the above-mentioned "store diagnosis" can be the client. After triggering the store diagnosis, the client can obtain the store data of the reference store and / or the historical store data of the target store from the server side, and then call the client's local diagnostic program to perform store diagnosis on the target store. Alternatively, the executor of "store diagnosis" is the server. After the client triggers the store diagnosis, the client sends a target store diagnosis request to the server so that the server can obtain the target store's current store data, historical store data, reference store data, etc., and diagnose the target store based on the acquired data to determine the target problem. The server will feedback the determined target store to the client.

[0240] Specifically, "diagnosing the target store" may include: diagnosing the gap between the target store and the reference store, and / or the data differences between the target store in different periods, and / or the gap between the store data of the target store and the set indicators, to obtain a diagnostic result.

[0241] The text displayed and / or played in the above 502 may be fed back by the server. For specific content, please refer to the above description and will not be elaborated here.

[0242] The present application also provides a method for working a store smart assistant in one embodiment. The execution subject of the method provided in this embodiment may be a client in the above service system. Figure 7 As shown, the method includes:

[0243] 601. Display the management page of the target store;

[0244] 602. Monitoring the user's operation on the management page to capture the user's target problems in store operation;

[0245] 603. In response to a confirmation instruction triggered by the user for the target problem, display and / or play a copy, where the copy includes suggestion content to help the user solve or improve the target problem;

[0246] Among them, the copy is obtained by using a first preset model, which is obtained by adjusting and training the pre-trained model for the intelligent answering task in the store operation scenario on the basis of the pre-trained model; the suggested content is associated with at least one factor of the target problem, and the at least one factor is obtained by attribution analysis of the target problem using a causal inference algorithm.

[0247] In the above 602, monitoring the user's operation on the management page to capture the user's target problems in store operation may include:

[0248] 6021. Responding to an input operation of a user on a management page, determining a target question based on the user input information; and / or

[0249] 6022. In response to a user's touch operation on a control on a management page, determine a target problem (e.g., how to set the setting item, etc.) according to a setting item associated with the control; and / or

[0250] 6023. Detecting a mouse cursor dwelling event at a content item on a management page, and determining a target problem based on the content item; and / or

[0251] 6024. Monitor the store data viewing operation and determine the target problem based on the store data.

[0252] In the above 6021, the user can directly enter the question information in the smart assistant interaction area on the management page.

[0253] In the above 6022, for example, the setting item may be a keyword setting item. The user may set keywords for the product in the keyword setting item. As shown in the figure, a control associated with the setting item is set near the keyword setting item on the management page. After the user touches the control, the target question may be determined to be "How to set keywords?"

[0254] In the above 6023, if the user stays at a certain content item for a long time, it is more likely that the user has questions. Therefore, when it is detected that the mouse cursor stays at a content item for more than a certain time (such as 10s, 30s, etc.), the target question can be determined based on the content item.

[0255] In the above 6024, the user viewed the recent store visits through the management page. The target question can be determined based on the store visits currently seen by the user, such as "Why is the visit volume reduced?".

[0256] Furthermore, the method provided in this embodiment may also include the following steps:

[0257] 604. Display the smart assistant logo on the management page corresponding to the target store;

[0258] 605. In response to the smart assistant wake-up operation, present a smart assistant page including an input box and at least one recommended query; wherein, at least one recommended query is obtained by diagnosing the target store.

[0259] Users can enter their target questions through the input box and interact with the intelligent assistant to obtain required data, problem solutions, and more.

[0260] The methods provided in the above two embodiments are written from the perspective of an intelligent assistant, which is the product form corresponding to the technical solution provided in this application. Based on the technical support provided by the above-mentioned server, the intelligent assistant can help merchants improve the efficiency of analysis, provide effective strategies and seize market opportunities, and help merchants grow better. The intelligent assistant can diagnose the problems of the store, use causal inference technology to analyze the store optimization items that solve or improve the problems, as well as at least one factor of the store optimization items, and then use the big model to output response copy that conforms to human language logic and conforms to the store operation scenario to help merchants grow.

[0261] An embodiment of the present application also provides an information processing device. The information processing device includes: a determination module, a retrieval module, a generation module and a calculation module. Among them, the determination module is used to determine the target problem for the target store. The retrieval module is used to retrieve at least one factor of the target problem in the copy database. The generation module is used to generate a prompt word Prompt based on the target problem and the suggested content associated with the at least one factor. The calculation module is used to output the copy corresponding to the Prompt using the first preset model to help users solve or improve the target problem.

[0262] Among them, the copywriting database stores: at least one factor of the preset problem obtained by attribution analysis of the preset problem using a causal inference algorithm; the first preset model is obtained by adjusting and training the pre-trained model based on the pre-trained model for the intelligent answering task in the store operation scenario.

[0263] Furthermore, the document database stores: at least one factor of the query template obtained by performing attribution analysis on a preset question associated with the query template using a causal inference algorithm. The retrieval module is used to determine a target query according to the target question; retrieve a query template matching the target query in the document database to obtain at least one factor of the matched query template; wherein, when generating the prompt, the suggested content associated with the target query and at least one factor of the matched query template is generated.

[0264] Furthermore, when the determination module determines the target Query according to the target question, it is used to sort multiple Query templates in the Query template library according to the target question; determine the target Query based on the first-ranked Query template, or use the first number of Query templates ranked in front as candidate Queries for user selection, and determine the target Query based on a candidate Query selected by the user.

[0265] Furthermore, the copywriting database stores: store optimization items that can solve or improve the preset problems and at least one factor of the store optimization items, obtained by performing attribution analysis on the preset problems associated with the Query template using a causal inference algorithm. When the retrieval module retrieves the Query template matching the target Query in the copywriting database and obtains at least one factor of the matched Query template, it is specifically used to: retrieve the Query template matching the target Query in the copywriting database; obtain the store optimization items associated with the matched Query template, and at least one factor of the store optimization items; wherein, when generating the Prompt, the suggested content is generated based on the target Query, the store optimization items associated with the matched Query template, and at least one factor of the store optimization items.

[0266] Furthermore, the device provided in this embodiment may also include a copywriting database construction module. The copywriting database construction module may use a causal inference algorithm to perform attribution analysis on the preset problem associated with the query template, and obtain a store optimization item and at least one factor of the store optimization item that solves or improves the preset problem. Specifically, the copywriting database construction module may be specifically used for:

[0267] For the preset problems associated with the Query template, the funnel analysis technology is used to analyze the traffic data and operation data of the stores on the platform to analyze the store optimization items that solve or improve the preset problems; the store optimization items are used as target modeling, and a causal tree model is constructed based on multiple factors; a data set is constructed based on the traffic data and operation data of the stores on the platform; using the data set, the influence of each of the multiple factors on the prediction results of the causal tree model is calculated to obtain the corresponding influence degree of each of the multiple factors; according to the corresponding influence degree of each of the multiple factors, the multiple factors are sorted to obtain the factor ranking of the store optimization item; wherein, at least one factor of the store optimization item is: the factor ranked second in the top in the factor ranking of the store optimization item.

[0268] Furthermore, the copywriting database construction module in the device provided in this embodiment is also used to: group the stores on the platform; when constructing the data set, construct it based on the traffic data and operation data of the stores in a group; and use the factor ranking corresponding to the store optimization items obtained from the data set to rank the factors of the store optimization items corresponding to the group. When obtaining the store optimization items associated with the matched Query template and at least one factor of the store optimization items, determine the group to which the target store belongs; obtain the store optimization items associated with the matched Query template and at least one factor of the store optimization items corresponding to the group to which the target store belongs.

[0269] Furthermore, the device provided in this embodiment also includes a Query template library construction module, a sample generation module, and an adjustment training module. Among them, the Query template library construction module is used to construct a Query template library, and the Query template library includes multiple Query templates. The copy material construction module is used to construct the copy material library based on the Query template library. The sample generation module is used to generate the training sample set based on the Query template library and the copy material library. The adjustment training module is used to use the training sample set to adjust and train the pre-trained model.

[0270] Furthermore, when determining the target problem for the target store, the determination module is specifically used to:

[0271] In response to user input, determining a target question for the target store based on the user input information; and / or

[0272] In response to a request triggered by a user on a management page of a target store for a setting item, determining the target problem for the setting item; and / or

[0273] Diagnose the target store and determine the target problem based on the diagnosis result; and / or

[0274] At least one recommended question recommended by the network side based on big data analysis is obtained, and in response to a selection event, a selected recommended question is used as the target question.

[0275] Among them, when the determination module diagnoses the target store and determines the target problem based on the diagnosis result, it is specifically used to: diagnose the gap between the target store and the reference store, and / or the data difference between the target store in different periods, and / or the gap between the store data of the target store and the set indicators to obtain a diagnosis result; determine the target problem based on the diagnosis result.

[0276] Another embodiment of the present application provides an information processing device including: a Query template library construction module, a copy database construction module, a sample generation module and an adjustment training module. The Query template library construction module is used to construct a query Query template library, and the Query template library includes multiple Query templates, and one Query template is associated with a preset question related to store operations. The copy database construction module is used to construct a copy database based on the Query template library; wherein the copy database stores: at least one factor of the Query template obtained by performing attribution analysis on the preset question associated with the Query template using a causal inference algorithm. The sample generation module is used to generate a training sample set based on the Query template library and the copy database. The adjustment training module is used to use the training sample set to adjust and train the pre-trained model to obtain a first preset model that can handle intelligent response tasks in store operation scenarios.

[0277] In a specific implementation scheme, when constructing the document database, the document database construction module uses a causal inference algorithm to perform attribution analysis on a preset question associated with a Query template to obtain at least one factor of the Query template, and is specifically used to:

[0278] For the preset problems associated with the Query template, the funnel analysis technology is used to analyze the traffic data and operation data of the stores on the platform to analyze the store optimization items that solve or improve the preset problems; the store optimization items are used as target modeling, and a causal tree model is constructed based on multiple factors; a data set is constructed based on the traffic data and operation data of the stores on the platform; using the data set, the influence of each of the multiple factors on the prediction results of the causal tree model is calculated to obtain the corresponding influence degree of each of the multiple factors; according to the corresponding influence degree of each of the multiple factors, the multiple factors are sorted to obtain the factor sorting of the store optimization items.

[0279] Furthermore, the copywriting database construction module is also used to: group the stores on the platform; when constructing the data set, it is constructed based on the traffic data and operation data of the stores in a group; and the factor ranking of the store optimization items obtained by using the data set is the factor ranking of the store optimization items corresponding to the group.

[0280] Furthermore, when constructing a copywriting database based on the Query template library, the copywriting database construction module is specifically used to: obtain and store attribute information of stores on the platform; obtain and store traffic data and operation data of stores on the platform; use a causal inference algorithm to perform attribution analysis on preset problems associated with the Query template in the Query template library to obtain at least one factor of the Query template; determine and store recommended content associated with the factors; and construct a corpus knowledge base of copywriting template information corresponding to different Query templates.

[0281] When generating a training sample set according to the Query template library and the document database, the sample generation module is specifically used to: output multiple training samples based on the Query template library and the document database using a third preset model; wherein the training sample set includes the multiple training samples.

[0282] Furthermore, when constructing the query template library, the Query template library construction module is specifically used to: determine multiple basic Query templates; use a second preset model to expand the multiple basic Query templates to obtain multiple extended Query templates; evaluate the multiple extended Query templates and retain the extended Query templates that pass the evaluation; wherein the Query template library includes multiple basic Query templates and multiple extended Query templates that pass the evaluation.

[0283] It should be noted here that the information processing devices provided in the above two embodiments can implement the technical solutions described in the above corresponding method embodiments. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above method embodiments, which will not be repeated here.

[0284] The store diagnosis device provided in one embodiment of the present application includes a trigger module and an output module. Among them, the trigger module is used to respond to a diagnostic trigger event for a target store, triggering the store diagnosis to present and / or play the diagnosed target problem. The output module is used to display and / or play a copy, and the copy includes suggested content to help users solve or improve the target problem. Among them, the copy is obtained using a first preset model, and the first preset model is based on the pre-trained model, and the pre-trained model is adjusted and trained for the intelligent answering task in the store operation scenario; the suggested content is associated with at least one factor of the target problem, and the at least one factor is obtained by attributing the target problem using a causal inference algorithm.

[0285] It should be noted here that: the store diagnosis device provided in the above embodiment can implement the technical solution described in the above corresponding method embodiment. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above corresponding method embodiment, which will not be repeated here.

[0286] The store smart assistant device provided by one embodiment of the present application includes a display device, a monitoring and capturing module and an output module. Among them, the display module is used to display the management page of the target store. The monitoring and capturing module is used to monitor the user's operations on the management page to capture the user's target problems in store operations. The output module is used to respond to the confirmation instruction triggered by the user for the target problem, and display and / or play the copy, which includes suggested content to help the user solve or improve the target problem.

[0287] Among them, the copy is obtained by using a first preset model, which is obtained by adjusting and training the pre-trained model for the intelligent answering task in the store operation scenario on the basis of the pre-trained model; the suggested content is associated with at least one factor of the target problem, and the at least one factor is obtained by attribution analysis of the target problem using a causal inference algorithm.

[0288] It should be noted here that the above embodiment provides a store smart assistant device that can implement the technical solution described in the above corresponding method embodiment. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above corresponding method embodiment, which will not be repeated here.

[0289] An embodiment of the present application also provides an electronic device. Figure 8 As shown, the electronic device includes a processor 42 and a memory 41. The memory 41 is used to store one or more computer programs; the processor 42 is coupled to the memory 41 and is used for the at least one or more computer programs to implement the steps in the method provided in each embodiment of the present application.

[0290] The above-mentioned memory 41 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0291] Furthermore, the electronic device also includes other components such as a communication component 43, a display 44, a power component 45, and an audio component 46. Only some components are schematically shown here, which does not mean that the electronic device only has these components.

[0292] Accordingly, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a computer, the method steps or functions provided in the above-mentioned embodiments of the present application can be implemented.

[0293] The method in the present application can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. Thus, the present application also provides a computer program product. The computer program product includes a computer program / instruction, and when the computer program / instruction is executed by an electronic component such as a processor, the steps or functions in the method provided in each embodiment of the present application can be executed in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user device, a core network device, an OAM or other programmable device.

[0294] Through the description of the above implementation modes, those skilled in the art can clearly understand that each implementation mode can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on such an understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0295] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.< / mask>

Claims

1. An information processing method, It is characterized in that include: Determine the target problems for the target stores; Retrieving at least one factor of the target problem in the document database; Generate a prompt word Prompt based on the target question and the suggestion content associated with the at least one factor; Outputting the text corresponding to the Prompt using the first preset model to help the user solve or improve the target problem; Among them, the copywriting database stores: at least one factor of the preset problem obtained by attribution analysis of the preset problem using a causal inference algorithm; the first preset model is obtained by adjusting and training the pre-trained model based on the pre-trained model for the intelligent answering task in the store operation scenario.

2. The method according to claim 1, It is characterized in that The document database stores: at least one factor of the Query template obtained by performing attribution analysis on a preset question associated with the Query template using a causal inference algorithm; as well as In the document database, at least one factor of the target problem is retrieved, including: According to the target question, determine the target Query; In the document database, a query template matching the target query is retrieved, and at least one factor of the matching query template is obtained; Wherein, when generating the Prompt, suggestion content is generated based on the target Query and at least one factor association of the matched Query template.

3. The method according to claim 2, It is characterized in that Determining a target Query according to the target question includes: According to the target question, multiple Query templates in the Query template library are sorted; The target query is determined based on the first ranked query template, or the first number of query templates ranked first are used as candidate queries for the user to select, and the target query is determined based on a candidate query selected by the user.

4. The method according to claim 2, It is characterized in that The copywriting database stores: store optimization items and at least one factor of the store optimization items that can solve or improve the preset problems obtained by performing attribution analysis on the preset problems associated with the query template using a causal inference algorithm; as well as In the document database, a query template matching the target query is retrieved, and at least one factor of the matched query template is obtained, including: Retrieving a query template matching the target query in the document database; Acquire a store optimization item associated with the matched Query template and at least one factor of the store optimization item; Wherein, when generating the Prompt, suggestion content is generated based on the target Query, the store optimization item associated with the matched Query template, and at least one factor associated with the store optimization item.

5. The method according to claim 4, It is characterized in that Using a causal inference algorithm to perform attribution analysis on the preset problem associated with the Query template, a store optimization item that solves or improves the preset problem and at least one factor of the store optimization item are obtained, including: For the preset problems associated with the query template, the funnel analysis technology is used to analyze the traffic data and operation data of the stores on the platform to analyze the store optimization items that solve or improve the preset problems; Taking the store optimization item as the target model, a causal tree model is constructed based on multiple factors; Build a data set based on the traffic data and operation data of the stores on the platform; Using the data set, calculating the influence of each of the multiple factors on the prediction result of the causal tree model, and obtaining the influence degree corresponding to each of the multiple factors; Sorting the multiple factors according to the influences corresponding to the multiple factors to obtain the factor ranking of the store optimization item; Among them, at least one factor of the store optimization item is: the factors ranked in the top second number in the factor ranking of the store optimization item.

6. The method according to claim 5, It is characterized in that Also includes: Group the stores on the platform; When constructing the data set, it is constructed based on the traffic data and operation data of the stores in a group; as well as The factor ranking corresponding to the store optimization item obtained by using the data set is the factor ranking of the store optimization item corresponding to the group; Among them, when obtaining the store optimization items associated with the matched Query template and at least one factor of the store optimization items, the group to which the target store belongs is determined; the store optimization items associated with the matched Query template and at least one factor of the store optimization items corresponding to the group to which the target store belongs are obtained.

7. The method according to any one of claims 1 to 6, It is characterized in that Also includes: Building a Query template library, wherein the Query template library includes a plurality of Query templates; Based on the Query template library, construct the copywriting database; Generate the training sample set according to the Query template library and the document database; The pre-training model is adjusted and trained using the training sample set.

8. The method according to claim 7, It is characterized in that The Query template library includes at least one of the following: Query template corresponding to the reason why the data of any store on the platform differs from that of the reference store; Query templates corresponding to the reasons for differences in data at different times for any store on the platform; Query template corresponding to store management issues; Query template corresponding to store optimization items; A query template extended by using the second preset model; The document database stores: Attribute information of stores on the platform; Traffic data and operation data of stores on the platform; Query template associated store optimization items; Factor ranking of store optimization items; Suggested content of factor associations; A corpus knowledge base of copywriting template information corresponding to different query templates.

9. The method according to any one of claims 1 to 6, It is characterized in that For the target stores, determine the target issues, including: In response to user input, determining a target question for the target store based on the user input information; and / or In response to a request triggered by a user for a setting item on a management page of a target store, determining the target problem for the setting item; and / or Diagnose the target store and determine the target problem based on the diagnosis result; and / or At least one recommended question recommended by the network side based on big data analysis is obtained, and in response to a selection event, a selected recommended question is used as the target question.

10. The method according to claim 9, It is characterized in that Diagnosing the target store and determining the target problem based on the diagnosis result includes: Diagnose the gap between the target store and the reference store, and / or the data difference between the target store in different periods, and / or the gap between the store data of the target store and the set indicators, and obtain a diagnosis result; The target problem is determined based on the diagnosis result.

11. An information processing method, It is characterized in that include: Building a query template library, wherein the query template library includes multiple query templates, and each query template is associated with a preset question related to store operations; Based on the Query template library, a document database is constructed; wherein the document database stores: at least one factor of the Query template obtained by performing attribution analysis on a preset question associated with the Query template using a causal inference algorithm; Generate a training sample set according to the Query template library and the document database; The pre-trained model is adjusted and trained using the training sample set to obtain a first preset model that can handle intelligent response tasks in store operation scenarios.

12. The method according to claim 11, It is characterized in that When constructing the document database, a causal inference algorithm is used to perform attribution analysis on the preset questions associated with the Query template to obtain at least one factor of the Query template, including: For the preset problems associated with the query template, the funnel analysis technology is used to analyze the traffic data and operation data of the stores on the platform to analyze the store optimization items that solve or improve the preset problems; Taking the store optimization item as the target model, a causal tree model is constructed based on multiple factors; Build a data set based on the traffic data and operation data of the stores on the platform; Using the data set, calculating the influence of each of the multiple factors on the prediction result of the causal tree model, and obtaining the influence degree corresponding to each of the multiple factors; The multiple factors are sorted according to the influences corresponding to each of the multiple factors to obtain the factor sorting of the store optimization items.

13. The method according to claim 12, It is characterized in that Also includes: Group the stores on the platform; When constructing the data set, it is constructed based on the traffic data and operation data of the stores in a group; as well as The factor ranking of the store optimization items obtained by using the data set is the factor ranking of the store optimization items corresponding to the group.

14. The method according to any one of claims 11 to 13, It is characterized in that Based on the Query template library, a copywriting database is constructed, including: Obtain and store attribute information of stores on the platform; Obtain and store traffic data and operation data of stores on the platform; Performing attribution analysis on preset questions associated with the Query templates in the Query template library using a causal inference algorithm to obtain at least one factor of the Query template; determining and storing suggested content associated with the factor; Construct a corpus knowledge base of copywriting template information corresponding to different query templates.

15. The method according to claim 14, It is characterized in that Generating a training sample set according to the Query template library and the document database includes: Outputting a plurality of training samples based on the Query template library and the document database using a third preset model; The training sample set includes the multiple training samples.

16. The method according to any one of claims 11 to 13, It is characterized in that Build a query template library, including: Determine multiple basic Query templates; Using the second preset model, the multiple basic Query templates are expanded to obtain multiple extended Query templates; Evaluating the multiple extended Query templates and retaining extended Query templates that pass the evaluation; The Query template library includes a plurality of basic Query templates and a plurality of extended Query templates that have passed evaluation.

17. A store diagnosis method, It is characterized in that Applicable to the client, the method includes: In response to a diagnosis trigger event for a target store, triggering store diagnosis to present and / or play a diagnosed target problem; Displaying and / or playing a copy, wherein the copy includes suggestion content to help the user solve or improve the target problem; Among them, the copy is obtained by using a first preset model, which is obtained by adjusting and training the pre-trained model for the intelligent answering task in the store operation scenario on the basis of the pre-trained model; the suggested content is associated with at least one factor of the target problem, and the at least one factor is obtained by attribution analysis of the target problem using a causal inference algorithm.

18. A method for working a store intelligent assistant. It is characterized in that Applicable to the client, the method includes: Display the management page of the target store; Monitor the user's operations on the management page to capture the user's target problems in store operations; In response to a confirmation instruction triggered by the user regarding the target problem, display and / or play a copy, wherein the copy includes suggestion content to help the user solve or improve the target problem; Among them, the copy is obtained by using a first preset model, which is obtained by adjusting and training the pre-trained model for the intelligent answering task in the store operation scenario on the basis of the pre-trained model; the suggested content is associated with at least one factor of the target problem, and the at least one factor is obtained by attribution analysis of the target problem using a causal inference algorithm.

19. A service system, It is characterized in that include: A client, used to implement the method described in claim 17 or 18; The server is used to implement the method described in any one of claims 1 to 16.

20. An electronic device, It is characterized in that comprising a memory and a processor; wherein, The memory is used to store executable instructions; The processor implements the steps in the information processing method described in any one of claims 1 to 16, or the steps in the store diagnosis method described in claim 17, or the steps in the store intelligent assistant working method described in claim 18 by running the executable instructions.

21. A computer-readable storage medium, It is characterized in that The storage medium stores computer instructions which, when executed by a processor, can implement the steps in the information processing method described in any one of claims 1 to 16, or the steps in the store diagnosis method described in claim 17, or the steps in the store intelligent assistant working method described in claim 18.

22. A computer program product, It is characterized in that The computer program product includes a computer program or instructions, which, when executed by a processor, enables the processor to execute the steps in the information processing method described in any one of claims 1 to 16, or the steps in the store diagnosis method described in claim 17, or the steps in the store intelligent assistant working method described in claim 18.

Citation Information

Patent Citations

  • Construction method of mold professional question-answering system based on LLM model

    CN117909458A

  • Marketing combination modeling and prediction method and system based on causal structure learning

    CN118071393A

  • Page content displaying method and apparatus, and electronic device

    WO2020177573A1

Cited By

  • Information processing methods, store diagnosis method, intelligent assistant working method, system and device

    WO2026001385A1