CN117494692A - Credential generation method, computer device, and computer storage medium - Google Patents

Credential generation method, computer device, and computer storage medium Download PDF

Info

Publication number
CN117494692A
CN117494692A CN202311762520.1A CN202311762520A CN117494692A CN 117494692 A CN117494692 A CN 117494692A CN 202311762520 A CN202311762520 A CN 202311762520A CN 117494692 A CN117494692 A CN 117494692A
Authority
CN
China
Prior art keywords
credential
data
target
generation model
document
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202311762520.1A
Other languages
Chinese (zh)
Inventor
杨芳
蒋宇翔
张炜
谭青
李慧超
王紫晗
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Kingdee Software China Co Ltd
Original Assignee
Kingdee Software China Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Kingdee Software China Co Ltd filed Critical Kingdee Software China Co Ltd
Priority to CN202311762520.1A priority Critical patent/CN117494692A/en
Publication of CN117494692A publication Critical patent/CN117494692A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/10—Text processing
    • G06F40/166—Editing, e.g. inserting or deleting
    • G06F40/186—Templates
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/082—Learning methods modifying the architecture, e.g. adding, deleting or silencing nodes or connections
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/12—Accounting

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Business, Economics & Management (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Biomedical Technology (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • Technology Law (AREA)
  • General Business, Economics & Management (AREA)
  • Medical Informatics (AREA)
  • Development Economics (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The embodiment of the application discloses a credential generation method, computer equipment and a computer storage medium. The computer equipment can acquire a target voucher generation model, the target voucher generation model is trained by a machine learning algorithm on a plurality of groups of training samples, each group of training samples comprises business scene data and voucher data generated based on the business scene data, and the target voucher data is input into the target voucher generation model to acquire target voucher data output by the target voucher generation model. Therefore, when the voucher is generated based on the receipt, a plurality of processes such as setting fields of the voucher template and analyzing and extracting number conditions by a computer are avoided, only the business scene data to be processed and the target voucher generation model are prepared, the business scene data is directly input into the model, the corresponding voucher data can be quickly obtained, and the voucher generation efficiency is greatly improved.

Description

Credential generation method, computer device, and computer storage medium
Technical Field
The embodiment of the application relates to the field of business processing, in particular to a credential generation method, computer equipment and a computer storage medium.
Background
When generating the voucher according to the business document, the voucher template can be used for generating the voucher, the voucher template is a template for generating the voucher based on the business document according to the accounting book, and a manual configuration mode is adopted currently, and a corresponding relation is set between the fields of the business document and the fields of the voucher, so that the requirement of rapidly and accurately generating the financial voucher by the business document is met.
When the voucher template is set, the setting interface of the voucher template comprises the fields of the subject value condition, the accounting dimension, the sum of money, the entry generation condition and the like, and the template is required to be adjusted at any time along with the service change, so that the learning cost is high, the implementation cost is high and the maintenance cost is high. Meanwhile, the voucher is generated through the voucher template, the fetch condition analysis process takes a long time, and the performance of generating the voucher is slow.
Disclosure of Invention
The embodiment of the application provides a credential generation method, computer equipment and a computer storage medium, corresponding credential data is rapidly generated based on business scene data and a pre-trained credential generation model, and the credential generation efficiency is greatly improved.
A first aspect of an embodiment of the present application provides a credential generating method, including:
acquiring target service scene data related to user service;
obtaining a target credential generation model, wherein the target credential generation model is obtained by training a plurality of groups of training samples through a machine learning algorithm, and each group of training samples comprises business scene data and credential data generated based on the business scene data;
and inputting the target business scene data into the target credential generation model to obtain target credential data output by the target credential generation model.
A second aspect of embodiments of the present application provides a computer device, the computer device comprising:
the first acquisition unit is used for acquiring target service scene data related to user service;
the second acquisition unit is used for acquiring a target credential generation model, the target credential generation model is obtained by training a plurality of groups of training samples through a machine learning algorithm, and each group of training samples comprises business scene data and credential data generated based on the business scene data;
and the generating unit is used for inputting the target business scene data into the target credential generating model so as to obtain target credential data output by the target credential generating model.
A third aspect of the embodiments of the present application provides a computer device comprising a memory storing a computer program and a processor implementing the method of the first aspect when the processor executes the computer program.
A fourth aspect of the embodiments provides a computer storage medium having stored therein instructions which, when executed on a computer, cause the computer to perform the method of the first aspect described above.
A fifth aspect of the embodiments of the present application provides a computer program product which, when run on a computer device, causes the computer device to perform the method of the first aspect described above.
From the above technical solutions, the embodiments of the present application have the following advantages:
the computer equipment can acquire a target voucher generation model, the target voucher generation model is trained by a machine learning algorithm on a plurality of groups of training samples, each group of training samples comprises business scene data and voucher data generated based on the business scene data, and the target voucher data is input into the target voucher generation model to acquire target voucher data output by the target voucher generation model. Therefore, when the voucher is generated based on the receipt, a plurality of processes such as setting fields of the voucher template and analyzing and extracting conditions of a computer by a user are omitted, corresponding voucher data can be obtained quickly by only preparing business scene data to be processed and deploying a target voucher generation model and directly inputting the business scene data into the model, the voucher generation efficiency is greatly improved, the user does not need to configure fields in the voucher template, user operation can be reduced, and the user experience of the user on a voucher generation function is improved.
Drawings
FIG. 1 is a schematic diagram of a network framework in an embodiment of the present application;
FIG. 2 is a schematic flow chart of a credential generation method according to an embodiment of the present application;
FIG. 3 is a schematic diagram of another flow chart of a credential generation method according to an embodiment of the present application;
FIG. 4 is a schematic flow chart of model training of a credential generation model in an embodiment of the present application;
FIG. 5 is an exemplary diagram illustrating a correspondence between documents and vouchers in an embodiment of the present application;
FIG. 6 is a schematic diagram of one manner in which field values of credential fields are generated based on business scenario data in an embodiment of the present application;
FIG. 7 is a schematic diagram of a computer device in an embodiment of the present application;
fig. 8 is a schematic diagram of another structure of a computer device in an embodiment of the present application.
Detailed Description
The embodiment of the application provides a credential generation method, computer equipment and a computer storage medium, corresponding credential data is rapidly generated based on business scene data and a pre-trained credential generation model, and the credential generation efficiency is greatly improved.
Referring to fig. 1, in an embodiment of the present application, a network framework includes:
a service server 100 and a terminal cluster; the terminal cluster may include: terminal devices 200a, 200b, 200c, … …, 200n, and the like.
The service server 100 may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server that provides a cloud database, a cloud service, cloud computing, a cloud function, cloud storage, a network service, cloud communication, a middleware service, a domain name service, a security service, a CDN, and basic cloud computing services such as a big data and an artificial intelligence platform. The terminal devices (including terminal device 200a, terminal device 200b, terminal devices 200c, … …, and terminal device 200 n) may be smart phones, tablet computers, notebook computers, desktop computers, palm computers, mobile internet devices (mobile internet device, MID), wearable devices (e.g., smart watches, smart bracelets, etc.), smart computers, smart vehicles, and other smart terminals.
The service server 100 may establish communication connection with each terminal device in the terminal cluster, and may also establish communication connection between each terminal device in the terminal cluster. In other words, the service server 100 may establish a communication connection with each of the terminal apparatuses 200a, 200b, 200c, … …, 200n, for example, a communication connection may be established between the terminal apparatus 200a and the service server 100. A communication connection may be established between terminal device 200a and terminal device 200b, and a communication connection may also be established between terminal device 200a and terminal device 200 c. The communication connection is not limited to a connection manner, and may be directly or indirectly connected through a wired communication manner, or may be directly or indirectly connected through a wireless communication manner, and the like, and may be specifically determined according to an actual application scenario, which is not limited herein.
It should be understood that each terminal device in the terminal cluster shown in fig. 1 may be provided with an application client, and when the application client runs in each terminal device, the application client may interact with the service server 100 respectively, so that the service server 100 may receive service data from each terminal device (such as financial management data uploaded by a user through the terminal device). The application client may be an application client having a function of displaying data information such as text, image, audio and video, such as a financial management application, an enterprise transaction management application, a browser application, a social application, an instant messaging application, a live broadcast application, a game application, a short video application, a music application, a shopping application, a novel application, a payment application, etc., and may specifically be determined according to actual application scene requirements, without limitation. The application client may be an independent client, or may be an embedded sub-client integrated in a certain client (such as a financial management client, an enterprise transaction management client, etc.), which may be specifically determined according to an actual application scenario, and is not limited herein.
The credential generation method in the embodiment of the present application will be described below with reference to the network framework of fig. 1:
referring to fig. 2, an embodiment of a credential generation method in an embodiment of the present application includes:
201. acquiring target service scene data related to user service;
the method of the present embodiment is applicable to a computer device, which may be the service server 100 or each terminal device in the network framework shown in fig. 1. The target service scenario data may be any data related to the service of the target user, for example, include user information of the target user, bill data of the target user, bill attribute information of a service bill, and the like, where the bill data may be a bill related to the service of the user, such as a purchase order, a sales order, a payment order, a receipt, and the like, and may also be each bill field and its field value in the service bill, and is not limited herein specifically.
202. Obtaining a target credential generation model, wherein the target credential generation model is obtained by training a plurality of groups of training samples through a machine learning algorithm, and each group of training samples comprises business scene data and credential data generated based on the business scene data;
the target credential generation model may be trained on the computer device such that the computer device directly obtains the target credential generation model; the target credential generation model may also be deployed to a computer device after model training is completed on other devices, not limited herein.
When the model is trained, the target credential generation model learns data characteristics of service scene data and data characteristics of credential data corresponding to the service scene data according to an input training sample, and fits association relations and mapping relations between the data characteristics of the service scene data and the data characteristics of the credential data, so that the capability of generating the credential data according to the service scene data is obtained, and when any service scene data is input, the corresponding credential data can be generated based on the mapping relations between the service scene data and the credential data learned before.
203. Inputting the target business scene data into the target credential generation model to obtain target credential data output by the target credential generation model;
the computer equipment can input the target business scene data into the target credential generation model, the target credential generation model extracts the data characteristics of the target business scene data, determines the data characteristics of the corresponding credential data based on the mapping relation between the business scene data and the credential data learned before, and further generates the target credential data corresponding to the target business scene data according to the determined data characteristics of the credential data, so that the target credential data output by the model can be obtained.
In this embodiment, the computer device may obtain a target credential generation model, where the target credential generation model is obtained by training multiple sets of training samples by a machine learning algorithm, and each set of training samples includes service scene data and credential data generated based on the service scene data, and input the target document data to the target credential generation model to obtain target credential data output by the target credential generation model. Therefore, when the voucher is generated based on the receipt, a plurality of processes such as setting fields of the voucher template and analyzing and extracting conditions of a computer by a user are omitted, corresponding voucher data can be obtained quickly by only preparing business scene data to be processed and deploying a target voucher generation model and directly inputting the business scene data into the model, the voucher generation efficiency is greatly improved, the user does not need to configure fields in the voucher template, user operation can be reduced, and the user experience of the user on a voucher generation function is improved.
Embodiments of the present application will be described in further detail below on the basis of the foregoing embodiment shown in fig. 2. Referring to fig. 3, another embodiment of a credential generation method in an embodiment of the present application includes:
301. acquiring target service scene data related to user service;
the operation performed in this step is similar to the operation performed in step 201 in the embodiment shown in fig. 2, and will not be described here again.
302. Obtaining a target credential generation model, wherein the target credential generation model is obtained by training a plurality of groups of training samples through a machine learning algorithm, and each group of training samples comprises business scene data and credential data generated based on the business scene data;
in this embodiment, the computer device may obtain the target credential generation model through model training. In model training, a number of steps may be performed to complete model training based on the following model training procedure:
3021. acquiring a plurality of groups of training samples, wherein each group of training samples comprises the receipt data and credential data generated based on the receipt data;
in this embodiment, the credential type of the credential that the user wishes to generate can be identified by the user's text/voice and other natural language through the multimodal function of the large model. After determining the credential type of the credential that the user needs to generate, a plurality of target credential fields may be determined according to the credential type, a field value corresponding to each target credential field in the history credential data may be obtained, and history document data for generating a field value of each target credential field may be obtained in the history service document, where the history document data, the target credential field and its field value form a set of training samples.
As shown in FIG. 5, a plurality of historical document data and a plurality of financial document data may be obtained, wherein the document entry corresponds to the document entry one-to-one. For example, the historical receipt data includes receipt fields such as receipt type, date of service, purchase organization, payment amount, and collection amount, and the field values of the voucher fields such as "borrower subject", "borrowing direction", "accounting date", "voucher word" in the voucher can be generated based on the field values of the plurality of fields in the receipt.
After determining the credential type of the credential required to be generated by the user, each credential field can be traversed in turn according to the credential type, and if traversing is performed to a "lending direction" field, a "lending direction" field and a field value thereof contained in the history credential are acquired, and a plurality of document fields and field values of document fields for generating the "lending direction" field and field values thereof are acquired in the history business document to form history document data, so that the "lending direction" field and field values thereof, and history document data for generating the "lending direction" field and field values thereof form a set of training samples.
3022. Acquiring an initial credential generation model;
3023. inputting a plurality of groups of training samples into the initial credential generation model so that the initial credential generation model adjusts model parameters of the initial credential generation model based on fitting results of bill data and credential data in the training samples, and stopping training when convergence conditions of model training are met, so as to obtain the target credential generation model;
after the training samples and the initial credential generation model are obtained, a plurality of groups of training samples can be input into the initial credential generation model, so that the initial credential generation model continuously fits the mapping relation between the bill data and the credential data in the training samples, the mapping relation can be a functional relation, model parameters of the initial credential generation model are adjusted based on the fitting result of the bill data and the credential data in the training samples, and the mapping relation between the bill data and the credential data in the training samples can be fitted again based on the adjusted model parameters until the mapping relation obtained by model fitting can represent the real mapping from the bill data to the credential data in each group of training samples, namely, the convergence condition of model training is met, and at the moment, training can be stopped, so that the target credential generation model is obtained.
When the model is trained, according to the scale of the training set, whether a decision tree or a random forest model is used can be judged, and the maximum information gain is used as a standard for dividing branches of the decision tree. In addition, the model can be optimized by post-CCP pruning (Cost ComplexityPruning, cost-complexity pruning) to prevent overfitting.
In a preferred embodiment, after the target credential generation model is obtained, the target credential generation model may be further trimmed according to the actual service condition of the user, so that the credential generated by the target credential generation model more accords with the actual service condition of the user.
Specifically, the service scene data in the training sample further comprises user information of the user and service data of the user. The document data to be processed of the user can be input into the target document generation model to obtain the predicted document data output by the target document generation model. Whether the user changes the predicted credential data output by the model or not can be judged through the embedded point, and when the user modifying operation of the predicted credential data is detected, the modified credential data of the predicted credential data is obtained. The method comprises the steps of inputting predicted credential data, modified credential data, user information of a user and service data of the user into a target credential generation model, so that the target credential generation model adjusts model parameters of the target credential generation model based on fitting results of the predicted credential data, the modified credential data, the user information of the user and the service data of the user, and training is stopped when convergence conditions of model training are met, and the target credential generation model is obtained, so that fine tuning training of the target credential generation model based on real service conditions of the user is achieved.
The user information of the user may be information related to the user, such as a service scale of a service implemented by the user, industry information, and the like. For example, if the user is an enterprise user, the user information may be industry information of an industry in which the enterprise is located, enterprise information of an enterprise scale, and the like. The service data of the user may be data related to the service of the user, such as service bill, service flowing, service progress status, service change and other service related information.
For example, the modification action input by the user can be used as a prediction target, the enterprise information and the business data information are used as characteristics to be transmitted into the target credential generation model for performing model fine adjustment training, and the target credential generation model for completing the model fine adjustment training can predict what modification decision the user can make on the prediction credential data output by the model and generate the credential conforming to the actual business situation of the user according to the modification decision.
Therefore, by means of the model fine tuning training based on the actual service condition of the user, the generated certificates can be changed according to the use preference of the user in the use process, for example, when the user changes the generated certificates under what condition, the generated certificates are used as characteristics to be transmitted into the target certificate generation model for fine tuning training, so that the intelligent certificate generation model is customized for each user, the certificate generation model generates corresponding certificates according to the actual service condition of the user, and the thousands of people and thousands of sides are achieved.
When the model outputs the predicted voucher data based on the to-be-processed receipt data, the to-be-processed receipt data can be input into the target voucher generation model, so that the target voucher generation model generates corresponding voucher fields and field values thereof based on the to-be-processed receipt data respectively, and generates the predicted voucher data based on the plurality of voucher fields and field values thereof.
For example, the type of the voucher required to be generated by the user may be determined, and a plurality of corresponding voucher fields may be determined based on the type of the voucher, as shown in fig. 6, where it is assumed that the plurality of voucher fields include "borrowing direction", "accounting date", "voucher word", "subject", "amount", and so on, and further field values of the corresponding voucher fields are generated according to a plurality of bill attribute information in the to-be-processed bill data, such as fields and field values of "accounting date", "borrowing direction", "voucher word" and so on are generated according to the bill attribute information such as "date", "organization", where "bill" is illustrated to refer to a bill attribute such as "date" refers to a bill service date "and" organization "refers to an organization to which the bill belongs. Therefore, the field values of the voucher fields such as the lending direction, the billing date, the voucher word and the like can be sequentially generated, the plurality of voucher fields and the field values thereof form the forecast voucher data, and the user can modify the field values in the forecast voucher data according to the actual business condition.
When generating the field value of the credential field based on the document data to be processed, a corresponding weight can be given to each document attribute information, the weight characterizes the influence degree of the document attribute information on the field value generation of the credential field, and the larger the weight is, the larger the influence degree is. As shown in fig. 6, when the "accounting date" field and its field value are generated, the document attribute information such as "document", "date", "organization" may be respectively assigned with weight (i.e. illustrated influence coefficient) x1, x2, x3, etc. weight values, where the weight values may be determined according to the actual service situation, so that the field value of the credential field is generated to more conform to the actual service situation.
303. Inputting the target business scene data into the target credential generation model to obtain target credential data output by the target credential generation model;
in this embodiment, the target credential generation model includes a plurality of neural network structures, each of which is used to generate a field value of one credential field of the predicted credential data, respectively. Thus, after model training is completed to obtain the target credential generation model, field values for multiple credential fields in the target credential data may be generated based on the same multiple neural network structures.
For example, a neural network structure for generating a field value of the credential field "lending direction" may be used to generate a field value of the credential field "lending direction" in the target credential data after its training is completed. And so on, field values for a plurality of credential fields of the target credential data are generated using a plurality of pre-trained neural network structures.
To sum up, the embodiment uses the machine learning algorithm model to skip the voucher template and directly generate the financial voucher from the business document. The technical core of the intelligent accounting machine learning algorithm model is that a target voucher generation model is obtained by training a large amount of business scene data such as client information (such as industries to which clients belong, company sizes, product purchasing modules and the like), receipt attribute information (such as receipt types, receipt business dates, organizations to which receipts belong and the like), receipt data and the like and accounting vouchers corresponding to the business scene data as sample data.
In addition, what modification is made to the accounting document generated by the algorithm under what situation is recognized by the embedded point, and the accounting document is saved together with the enterprise information (the industry to which the client belongs, the geographic position and the like) and the business data of the enterprise. Secondly, modeling the multi-layer feedforward neural network for each credential field. The main model structure has three parts, namely an input layer, a hidden layer and an output layer. The input layer receives the information and takes the information as a characteristic, then establishes a full connection layer as a hidden layer, receives the information transmitted by the input layer, and recombines and extracts important characteristics, the cross entropy is taken as a loss function, a gradient descent method is adopted to update model parameters for multiple times to obtain minimum loss so as to achieve a model convergence condition, the probability of each behavior (including not modifying the credential field or modifying the credential field into a certain specific result) of a user is predicted by taking the minimum loss as a standard, and finally the output is completed through the output layer. After completion, the neural network corresponding to each credential field is assembled into a target credential generation model, and target credential data corresponding to the business scene data of the client is predicted through the neural network, so that the target credential data accords with the real business scene of the client.
Therefore, in this embodiment, the algorithm model may be used to intelligently generate accounting vouchers from document data, and the document data may be obtained by directly inputting the service scene data to the model without carding the service scene or configuring the voucher template.
The credential generation method in the embodiment of the present application is described above, and the computer device in the embodiment of the present application is described below, referring to fig. 7, one embodiment of the computer device in the embodiment of the present application includes:
a first obtaining unit 701, configured to obtain target service scenario data related to a user service;
a second obtaining unit 702, configured to obtain a target credential generating model, where the target credential generating model is obtained by training multiple sets of training samples by a machine learning algorithm, and each set of training samples includes service scene data and credential data generated based on the service scene data;
the generating unit 703 is configured to input the target service scenario data to the target credential generation model, so as to obtain target credential data output by the target credential generation model.
In a preferred implementation manner of this embodiment, the service scenario data includes document data; the second acquisition unit 702 is specifically configured to:
acquiring a plurality of groups of training samples, wherein each group of training samples comprises the receipt data and credential data generated based on the receipt data;
acquiring an initial credential generation model;
and inputting a plurality of groups of training samples into the initial credential generation model so that the initial credential generation model adjusts model parameters of the initial credential generation model based on fitting results of bill data and credential data in the training samples, and stopping training when convergence conditions of model training are met, so as to obtain the target credential generation model.
In a preferred implementation manner of this embodiment, the service scenario data further includes user information of a user and service data of the user; the second acquisition unit 702 is further configured to:
inputting bill data to be processed into the target credential generation model to obtain predicted credential data output by the target credential generation model;
when the modification operation of the user on the predicted credential data is detected, obtaining the modified credential data of the predicted credential data;
inputting the predicted credential data, the modified credential data, the user information of the user and the service data of the user into the target credential generation model, so that the target credential generation model adjusts model parameters of the target credential generation model based on fitting results of the predicted credential data, the modified credential data, the user information of the user and the service data of the user, and stops training when convergence conditions of model training are met, thereby obtaining the target credential generation model.
In a preferred implementation manner of this embodiment, the second obtaining unit 702 is specifically configured to:
determining a credential type of a credential that the user needs to generate, and determining a plurality of target credential fields according to the credential type;
acquiring a field value corresponding to each target credential field in the history credential data, and acquiring history document data for generating a field value of each target credential field in a history service document, wherein the history document data, the target credential field and the field value thereof form a group of training samples.
In a preferred implementation manner of this embodiment, the second obtaining unit 702 is specifically configured to:
inputting the document data to be processed into the target document generation model, so that the target document generation model generates corresponding document fields and field values thereof based on the document data to be processed respectively, and generates the prediction document data based on a plurality of the document fields and field values thereof.
In a preferred implementation manner of this embodiment, the document data to be processed includes a plurality of document attribute information; the second acquisition unit 702 is specifically configured to:
acquiring an influence coefficient which is pre-endowed to each piece of document attribute information of the document data to be processed, wherein the influence coefficient is used for representing the influence degree of the document attribute information on the generation of a field value of a credential field;
and generating a corresponding credential field and a field value thereof based on the plurality of bill attribute information of the bill data to be processed and the corresponding influence coefficient.
In a preferred implementation of this embodiment, the target credential generation model includes a plurality of neural network structures, each of which is configured to generate a field value of a credential field of the predicted credential data.
In this embodiment, the operations performed by the units in the computer device are similar to those described in the embodiments shown in fig. 2 to 3, and are not repeated here.
In the embodiment, when the voucher is generated based on the receipt, a plurality of processes such as setting fields of the voucher template and analyzing conditions of a computer by a user are avoided, the corresponding voucher data can be obtained quickly by only preparing business scene data to be processed and deploying a target voucher generation model and directly inputting the business scene data into the model, the voucher generation efficiency is greatly improved, the user does not need to configure the fields in the voucher template, the user operation can be reduced, and the user experience on the voucher generation function is improved.
Referring to fig. 8, an embodiment of a computer device in the embodiment of the present application includes:
the computer device 800 may include one or more central processing units (central processing units, CPU) 801 and memory 805, with one or more application programs or data stored in the memory 805.
Wherein the memory 805 may be volatile storage or persistent storage. The program stored in the memory 805 may include one or more modules, each of which may include a series of instruction operations in a computer device. Still further, the central processor 801 may be arranged to communicate with a memory 805 to execute a series of instruction operations in the memory 805 on the computer device 800.
The computer device 800 may also include one or more power supplies 802, one or more wired or wireless network interfaces 803, one or more input/output interfaces 804, and/or one or more operating systems, such as Windows ServerTM, mac OS XTM, unixTM, linuxTM, freeBSDTM, etc.
The cpu 801 may perform the operations performed by the computer device in the embodiments shown in fig. 2 to 3, and will not be described in detail herein.
Embodiments of the present application also provide a computer storage medium, where one embodiment includes: the computer storage medium has stored therein instructions which, when executed on a computer, cause the computer to perform the operations performed by the computer device in the embodiments shown in fig. 2 to 3 described above.
Embodiments of the present application also provide a computer program product, one of which includes: the computer program product, when run on a computer device, causes the computer device to perform the operations performed by the computer device in the embodiments shown in fig. 2 to 3 described above.
It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, which are not repeated herein.
In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods may be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of the units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a storage medium, including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a read-only memory (ROM), a random access memory (RAM, random access memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

Claims (10)

1. A method of credential generation, the method comprising:
acquiring target service scene data related to user service;
obtaining a target credential generation model, wherein the target credential generation model is obtained by training a plurality of groups of training samples through a machine learning algorithm, and each group of training samples comprises business scene data and credential data generated based on the business scene data;
and inputting the target business scene data into the target credential generation model to obtain target credential data output by the target credential generation model.
2. The method of claim 1, wherein the business scenario data comprises document data; the obtaining the target credential generation model includes:
acquiring a plurality of groups of training samples, wherein each group of training samples comprises the receipt data and credential data generated based on the receipt data;
acquiring an initial credential generation model;
and inputting a plurality of groups of training samples into the initial credential generation model so that the initial credential generation model adjusts model parameters of the initial credential generation model based on fitting results of bill data and credential data in the training samples, and stopping training when convergence conditions of model training are met, so as to obtain the target credential generation model.
3. The method of claim 2, wherein the business scenario data further comprises user information of a user and business data of the user; the method further comprises the steps of:
inputting bill data to be processed into the target credential generation model to obtain predicted credential data output by the target credential generation model;
when the modification operation of the user on the predicted credential data is detected, obtaining the modified credential data of the predicted credential data;
inputting the predicted credential data, the modified credential data, the user information of the user and the service data of the user into the target credential generation model, so that the target credential generation model adjusts model parameters of the target credential generation model based on fitting results of the predicted credential data, the modified credential data, the user information of the user and the service data of the user, and stops training when convergence conditions of model training are met, thereby obtaining the target credential generation model.
4. A method according to claim 2 or 3, wherein the obtaining a plurality of sets of training samples comprises:
determining a credential type of a credential that the user needs to generate, and determining a plurality of target credential fields according to the credential type;
acquiring a field value corresponding to each target credential field in the history credential data, and acquiring history document data for generating a field value of each target credential field in a history service document, wherein the history document data, the target credential field and the field value thereof form a group of training samples.
5. A method according to claim 3, wherein inputting document data to be processed into the target document generation model to obtain predicted document data output by the target document generation model comprises:
inputting the document data to be processed into the target document generation model, so that the target document generation model generates corresponding document fields and field values thereof based on the document data to be processed respectively, and generates the prediction document data based on a plurality of the document fields and field values thereof.
6. The method of claim 5, wherein the document data to be processed includes a plurality of document attribute information;
the generating the corresponding credential field and the field value thereof based on the document data to be processed comprises the following steps:
acquiring an influence coefficient which is pre-endowed to each piece of document attribute information of the document data to be processed, wherein the influence coefficient is used for representing the influence degree of the document attribute information on the generation of a field value of a credential field;
and generating a corresponding credential field and a field value thereof based on the plurality of bill attribute information of the bill data to be processed and the corresponding influence coefficient.
7. The method of claim 5, wherein the target credential generation model comprises a plurality of neural network structures, each for generating a field value for a credential field of the predicted credential data.
8. A computer device, the computer device comprising:
the first acquisition unit is used for acquiring target service scene data related to user service;
the second acquisition unit is used for acquiring a target credential generation model, the target credential generation model is obtained by training a plurality of groups of training samples through a machine learning algorithm, and each group of training samples comprises business scene data and credential data generated based on the business scene data;
and the generating unit is used for inputting the target business scene data into the target credential generating model so as to obtain target credential data output by the target credential generating model.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the method of any of claims 1 to 7 when executing the computer program.
10. A computer storage medium having instructions stored therein, which when executed on a computer, cause the computer to perform the method of any of claims 1 to 7.
CN202311762520.1A 2023-12-19 2023-12-19 Credential generation method, computer device, and computer storage medium Pending CN117494692A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311762520.1A CN117494692A (en) 2023-12-19 2023-12-19 Credential generation method, computer device, and computer storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311762520.1A CN117494692A (en) 2023-12-19 2023-12-19 Credential generation method, computer device, and computer storage medium

Publications (1)

Publication Number Publication Date
CN117494692A true CN117494692A (en) 2024-02-02

Family

ID=89671191

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311762520.1A Pending CN117494692A (en) 2023-12-19 2023-12-19 Credential generation method, computer device, and computer storage medium

Country Status (1)

Country Link
CN (1) CN117494692A (en)

Similar Documents

Publication Publication Date Title
CN106649890A (en) Data storage method and device
CN112633962B (en) Service recommendation method and device, computer equipment and storage medium
CN113011889B (en) Account anomaly identification method, system, device, equipment and medium
CN108021934A (en) The method and device of more key element identifications
CN113742492B (en) Insurance scheme generation method and device, electronic equipment and storage medium
CN111582341B (en) User abnormal operation prediction method and device
CN110555749B (en) Credit behavior prediction method and device based on neural network
CN111369344B (en) Method and device for dynamically generating early warning rules
CN113282623A (en) Data processing method and device
CN111275485A (en) Power grid customer grade division method and system based on big data analysis, computer equipment and storage medium
CN112950359A (en) User identification method and device
CN113761350A (en) Data recommendation method, related device and data recommendation system
WO2020093817A1 (en) Identity verification method and device
CN116777646A (en) Artificial intelligence-based risk identification method, apparatus, device and storage medium
CN116071150A (en) Data processing method, bank product popularization, wind control system, server and medium
CN111401675A (en) Similarity-based risk identification method, device, equipment and storage medium
CN117494692A (en) Credential generation method, computer device, and computer storage medium
CN115601047A (en) Data processing method and device
CN112329928B (en) Heterogeneous model-based user satisfaction analysis method
CN114742163A (en) After-sales task auditing method and device
US20210166318A1 (en) Systems and methods for client profile-based sales decisions
CN112990311A (en) Method and device for identifying admitted client
CN113673687A (en) Data processing method, device, equipment and computer readable storage medium
CN111340632A (en) Resource control method and device
CN111882339A (en) Prediction model training and response rate prediction method, device, equipment and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination