CN115509485A - Filling-in method and device of business form, electronic equipment and storage medium - Google Patents

Abstract

The embodiment of the invention provides a method and a device for filling a service form, electronic equipment and a storage medium, wherein the method comprises the following steps: receiving a first voice instruction, and determining a first target intention corresponding to the first voice instruction by adopting an intention model; when the first action instruction model corresponding to the first target intention is a business form display model, displaying a target business form model corresponding to the first target intention; receiving a second voice instruction in the process of displaying the target business form model, and determining a second target intention corresponding to the second voice instruction by adopting an intention model; and when a second action instruction model corresponding to the second target intention is related to the filling target business form model, filling operation is carried out on the target business form model according to the second action instruction model and the second voice instruction. By the embodiment of the invention, the automatic filling of the business form is realized based on the voice recognition without manual input of the user.

Description

A business form filling method, device, electronic equipment and storage medium

technical field

The invention relates to the technical field of data processing, in particular to a method, device, electronic equipment and storage medium for filling out a business form.

Background technique

According to statistics, at present, about 20% of villagers in China still use non-smart phones; and even if they use smart phones, many older villagers have difficulty completing business forms on smart phones due to their low educational level Fill in; and even if it takes a lot of time to fill in, there may be a lot of filling errors.

Contents of the invention

In view of the above problems, a business form filling method, device, electronic device and storage medium for overcoming the above problems or at least partially solving the above problems are proposed, including:

A method for filling out a business form, the method comprising:

Create a business form model for different businesses, an intent model for identifying the intent of the business form model, and a plurality of action instruction models for performing filling-related operations for the business form model; in the intent model There is a corresponding relationship between the intention and the business form model and the action instruction model;

receiving a first voice instruction, and using the intention model to determine a first target intention corresponding to the first voice instruction;

When the first action instruction model corresponding to the first target intent is displaying a business form model, displaying a target business form model corresponding to the first target intent;

In the process of displaying the target business form model, receiving a second voice instruction, and using the intent model to determine a second target intent corresponding to the second voice instruction;

When the second action instruction model corresponding to the second target intention is related to filling the target business form model, perform a filling operation on the target business form model according to the second action instruction model and the second voice instruction .

Optionally, the displaying a target business form model corresponding to the first target intent includes:

determining the target device type of the terminal device used to display the target business form model;

When the target device type is the first type, displaying the target business form model through voice interaction;

When the type of the target device is the second type, the target business form model is displayed on the terminal device.

Optionally, the method also includes:

recording first exception log data when the first target intent cannot be determined and/or the second target intent cannot be determined;

updating the intent model according to the first exception log data;

Optionally, the method also includes:

When the filling operation of the target business form model according to the second action instruction model and the second voice instruction is abnormal, generating second abnormal log data;

The multiple action instruction models are updated according to the second exception log data, and the intention model is updated.

Optionally, the target business form model includes a plurality of fields, and the filling operation of the target business form model according to the second action instruction model and the second voice instruction includes:

determining the target field targeted by the second voice instruction, and extracting the first text information for the target field from the second voice instruction;

Filling the first text information into the target field according to the target algorithm in the second action instruction model.

Optionally, the method also includes:

Detect whether there is any abnormality in the filled target field;

When there is an abnormality in the filled target field, a prompt is given for the target field;

When there is no abnormal filling content in the filled target field, the data filled in the target field is saved.

Optionally, the determining the first target intention corresponding to the first voice instruction by using the intention model includes:

converting the first voice instruction into corresponding second text information;

extracting target keywords for intent recognition from the second text information by using natural language processing;

The first target intent corresponding to the target keyword is determined by using the intent model.

The embodiment of the present invention also provides a device for filling in a business form, the device comprising:

Create a module for creating business form models for different businesses, an intent model for identifying intentions for the business form model, and multiple action instruction models for performing filling-related operations for the business form model; There is a corresponding relationship between the intents in the intent model and the business form model and action instruction model respectively;

A first intention determination module, configured to receive a first voice instruction, and use the intention model to determine a first target intention corresponding to the first voice instruction;

A display module, configured to display a target business form model corresponding to the first target intent when the first action instruction model corresponding to the first target intent is a display business form model;

The second intention determination module is configured to receive a second voice instruction during the process of displaying the target business form model, and determine a second target intention corresponding to the second voice instruction by using the intention model;

A filling module, configured to, when the second action instruction model corresponding to the second target intention is related to filling the target business form model, complete the target business according to the second action instruction model and the second voice instruction The form model performs filling operations.

Optionally, the display module includes:

A target device type determining submodule, configured to determine the target device type of the terminal device used to display the target business form model;

The voice interaction display submodule is used to display the target business form model through voice interaction when the target device type is the first type;

A display submodule, configured to display the target business form model in the terminal device when the target device type is the second type.

Optionally, the device also includes:

The first update module is used to record the first abnormal log data when the first target intent cannot be determined and/or the second target intent cannot be determined;

updating the intent model according to the first exception log data;

Optionally, the device also includes:

The second update module is used to generate second abnormal log data when filling in the target business form model according to the second action instruction model and the second voice instruction is abnormal; The second exception log data updates the plurality of action instruction models, and updates the intent model.

Optionally, the target business form model includes a plurality of fields, and the filling module is configured to determine the target field targeted by the second voice instruction, and extract the target field from the second voice command. The first text information of the field; filling the first text information into the target field according to the target algorithm in the second action instruction model.

Optionally, the device also includes:

Anomaly detection module, used to detect whether there is an abnormality in the filling content of the filled target field; when there is an abnormal filling content in the filled target field, a prompt is given for the target field; when there is no abnormal filling content in the filled target field to save the data populated into the target field.

Optionally, the first intention determination module is configured to convert the first voice instruction into corresponding second text information; and extract the information for intention recognition from the second text information by using natural language processing. target keywords: using the intent model to determine a first target intent corresponding to the target keywords.

An embodiment of the present invention also provides an electronic device, including a processor, a memory, and a computer program stored on the memory and capable of running on the processor. When the computer program is executed by the processor, the above How to fill out business forms.

An embodiment of the present invention also provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the above method for filling in a business form is realized.

Embodiments of the present invention have the following advantages:

In the embodiment of the present invention, it is possible to pre-create business form models for different businesses, an intention model for identifying intentions for the business form models, and multiple action instruction models for performing filling-related operations for the business form models; the intention model There is a corresponding relationship between the intentions in the business form model and the action instruction model; receive the first voice instruction, and use the intention model to determine the first target intention corresponding to the first voice instruction; when the first action instruction corresponding to the first target intention When the model is a display business form model, display the target business form model corresponding to the first target intent; in the process of displaying the target business form model, receive the second voice instruction, and use the intent model to determine the second voice instruction corresponding to the second voice instruction. Target intent: when the second action instruction model corresponding to the second target intent is related to filling the target business form model, fill in the target business form model according to the second action instruction model and the second voice instruction. Through the embodiment of the present invention, it is realized based on speech recognition to help users complete the automatic filling of business forms without manual input by users; it not only improves the efficiency of users filling in business forms, but also reduces the learning cost for users to fill in business forms; and Based on the intention of the voice, the filling of the business form is automatically completed, which improves the correct rate of filling the user's business form.

Description of drawings

In order to illustrate the technical solution of the present invention more clearly, the accompanying drawings that need to be used in the description of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. Ordinary technicians can also obtain other drawings based on these drawings without paying creative labor.

Fig. 1 is a flow chart of the steps of a method for filling in a business form according to an embodiment of the present invention;

Fig. 2 is a flow chart of the steps of another method for filling in a business form according to an embodiment of the present invention;

Fig. 3 is a schematic diagram of a scene according to an embodiment of the present invention;

Fig. 4 is a schematic flow chart of a filling method of a business form according to an embodiment of the present invention;

Fig. 5 is a structural block diagram of a device for filling in a business form according to an embodiment of the present invention.

detailed description

In order to make the above objects, features and advantages of the present invention more comprehensible, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. Apparently, the described embodiments are some, but not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

According to the survey results, the current problems in rural areas are as follows:

According to statistics, at present, about 20% of the villagers in China still use non-smart phones, and the other part of the villagers do not have a high degree of acceptance of smart phones; There are many barriers; for example: difficult to complete the entry of business forms. In view of this, the embodiment of the present invention provides a method for filling in business forms, which can be based on voice recognition to help villagers complete the automatic filling of business forms without manual input by villagers, which not only improves the efficiency of villagers' filling in business forms The efficiency of the form reduces the learning cost for the villagers to fill in the business form; and the filling of the business form is automatically completed based on the voice intention, which improves the accuracy of the villagers' business form filling.

Referring to Fig. 1, it shows a flow chart of the steps of a method for filling in a business form according to an embodiment of the present invention, which may include the following steps:

Step 101, create business form models for different businesses, an intent model for identifying intents for the business form models, and multiple action instruction models for performing filling-related operations for the business form models; the intents in the intent models are respectively related to There is a corresponding relationship between the business form model and the action instruction model.

In practical applications, different businesses may correspond to different business forms; for different businesses, the business form model of the corresponding business form can be created; wherein, the business form model can include the form name, and store the form ID (Identity document, identification number), form category, etc.

The form name can refer to the name of the form in real life, such as "Application Form for Provident Fund Extraction"; the form ID can refer to the identifier used to uniquely identify the form in the database, which can be identified in the form of numbers and letters, such as "A123456"; the form The type can refer to the type corresponding to the form, for example: application type, evaluation type, etc.

For different business form models, form field information can also be created and stored in the database; form field information can include: form field ID, form field identification, data type and field classification; among them, the form field ID can refer to the The identifier that uniquely identifies the form field in the database can be identified in the form of numbers and letters; the form field identifier can refer to the identifier used to let developers know the information corresponding to the form field, for example: "ID card number"; the data type can include The type of data in the recharge form field, such as: time type, numerical type, text type, etc.; field classification can refer to the type corresponding to the form field, such as: express delivery order number class, ID number class, time class, etc., the present invention The embodiment does not limit this.

For each field classification, a mapping relationship with the corresponding keyword identifier can be set; thus, when the voice instruction input by the user includes keywords, the form field to be filled can be determined from the business form based on the corresponding keyword identifier.

At the same time, a corresponding intent model can also be established for the business form model; specifically, the intent model can include all intents corresponding to filling in the business form, for example: open a business form model, fill in a form field, modify content, submit the form , specify a link address, pop up a form, close the current page, etc. For different intents, you can upload the corresponding intent corpus to train the intent model. For example, the intent model training may be performed by using a convolutional neural network or a recurrent neural network algorithm, which is not limited in this embodiment of the present invention.

When filling out a business form, you may need to perform different operations, for example: fill the text information corresponding to the voice command into the form field, convert the text information corresponding to the voice command and fill it into the form field, close the business form model, open the business form model , open a link, enlarge a window, etc.; for this, one or more action instruction models for performing filling-related operations on the business form model can be created in advance.

Specifically, the action instruction calling method (for example: fill in, process, return) and the required parameter information (that is, the information to be filled in) can be saved in the database in the form of an action instruction model; the action instruction model can include: action instruction ID , method name, return value type, parameter list, etc.

Wherein, the action instruction ID can refer to the identifier used to uniquely identify the action instruction model in the database, and can be identified in the form of numbers and letters; the method name can include the name of the algorithm used to process the voice instruction; the return value type can refer to The next step after the operation is completed; the parameter table can include parameters of multiple steps, so as to ensure that after the action instruction model is executed, the next step can be automatically executed.

In practical applications, the business form model, intent model, and action instruction model can be created in advance, and the correspondence between the intent in the intent model and the business form model and action instruction model can be established; so that when the voice instruction input by the user is received Afterwards, the business form model to be filled in and the filling operation to be performed on the business form model can be determined based on the corresponding intent of the voice command.

Step 102: Receive a first voice instruction, and use an intention model to determine a first target intention corresponding to the first voice instruction.

When the user needs to fill in a business form, he can first perform the operation of generating the first voice command on the terminal device; for example: say "I want to handle * * business" to the terminal device; the terminal device according to what the user said , the corresponding first voice command can be generated and uploaded to the server.

After the server receives the first voice command, it can use the pre-created intention model to recognize the first voice command, so as to determine the user's intention corresponding to the first voice command; in order to facilitate the distinction, this intention is called the first target intention.

As an example, the first target intention may be to open a business form model, which is not limited in this embodiment of the present invention.

Step 103, when the first action instruction model corresponding to the first target intent is a display business form model, display the target business form model corresponding to the first target intent.

In practical applications, the intent model can include not only intents for business forms, but also other intents; for example: asking questions, comparing, suggesting, thanking, and so on.

After determining the first target intent corresponding to the first voice instruction, the target business form model and the first action instruction model corresponding to the first target intent may be determined based on the correspondence between the intent and the action instruction model and business form model.

If the operation corresponding to the first action instruction model is to display a business form model, the target business form model can be displayed on the user's terminal device, so that the user can fill in the business form based on the target business form displayed on the terminal device.

Step 104: During the process of displaying the target business form model, receive the second voice instruction, and use the intent model to determine the second target intent corresponding to the second voice instruction.

In the process of displaying the target business form model on the terminal device, the user can perform the operation of generating a second voice command for the content in the target business form model, for example: "My ID number is ****"; the terminal device Based on what the user said "my ID card number is ****", a corresponding second voice command can be generated.

After the terminal device generates the second voice instruction, it can upload the second voice instruction to the server, so that the server can use the intent model to determine the second target intent corresponding to the second voice instruction.

As an example, the second target intent may be a filling intent for the target business form model, or a closing intent, submission intent, etc. for the target business form model, which is not limited in this embodiment of the present invention.

Step 105 , when the second action instruction model corresponding to the second target intent is related to filling the target business form model, perform filling operation on the target business form model according to the second action instruction model and the second voice instruction.

After the second target intention is determined, a second action instruction model corresponding to the second target intention can be determined; if the second action instruction model is related to filling a business form model, it can be determined according to the second action instruction model and the second voice instruction Fill in the target business form model to complete the filling of the target business form model.

After completing the filling of the target business form model, the target business form model may be uploaded to a corresponding server for storage, which is not limited in this embodiment of the present invention.

In the embodiment of the present invention, it is possible to pre-create business form models for different businesses, an intention model for identifying intentions for the business form models, and multiple action instruction models for performing filling-related operations for the business form models; the intention model There is a corresponding relationship between the intentions in the business form model and the action instruction model; receive the first voice instruction, and use the intention model to determine the first target intention corresponding to the first voice instruction; when the first action instruction corresponding to the first target intention When the model is a display business form model, display the target business form model corresponding to the first target intent; in the process of displaying the target business form model, receive the second voice instruction, and use the intent model to determine the second voice instruction corresponding to the second voice instruction. Target intent: when the second action instruction model corresponding to the second target intent is related to filling the target business form model, fill in the target business form model according to the second action instruction model and the second voice instruction. Through the embodiment of the present invention, it is realized based on speech recognition to help users complete the automatic filling of business forms without manual input by users; it not only improves the efficiency of users filling in business forms, but also reduces the learning cost for users to fill in business forms; and Based on the intention of the voice, the filling of the business form is automatically completed, which improves the correct rate of filling the user's business form.

Referring to FIG. 2 , it shows a flow chart of the steps of another method for filling in a business form according to an embodiment of the present invention, which may include the following steps:

Step 201, creating business form models for different businesses, an intent model for identifying intentions for the business form models, and multiple action instruction models for performing filling-related operations for the business form models.

In practical applications, different businesses may correspond to different business forms; for different businesses, the business form model of the corresponding business form can be created; at the same time, a corresponding intent model can also be established for the business form model, and can be created for One or more action instruction models that perform fill-in related operations on the business form model

Step 202: Receive a first voice instruction, and convert the first voice instruction into corresponding second text information.

After receiving the first voice command, the server may recognize the first voice command into the second text information by calling a voice recognition interface using ASR (Automatic Speech Recognition, automatic speech recognition technology) technology.

Step 203, using natural language processing to extract target keywords for intention recognition from the second text information.

Then, the server may use NLP (Natural Language Processing, natural language processing) technology to extract target keywords that can be used for intent recognition from the second text information.

Step 204, using the intent model to determine the first target intent corresponding to the target keyword.

After acquiring the target keyword, the server may use an intent model to identify the intent corresponding to the target keyword, so as to determine the first target intent corresponding to the target keyword.

As an example, intent recognition may be performed based on word segmentation, part-of-speech tagging, named entity recognition, dependency syntax analysis, semantic analysis, or rule analysis.

For example: to identify intent by means of rule parsing:

The second text information is "{I} want to apply for the minimum living allowance", and the rule template can be used to obtain the first target intention after extraction: [user]: [handling] [the minimum living allowance].

In an example, the action instruction model associated with [handling] can be the action instruction model of [pop-up window], and the business form model associated with [subsistence allowances] can be the business form model of [subsistence allowance handling form]. Examples are not limited to this.

Step 205: When the first action instruction model corresponding to the first target intention is a display business form model, determine the target device type of the terminal device used to display the target business form model.

After determining the first target intent corresponding to the first voice instruction, the target business form model and the first action instruction model corresponding to the first target intent may be determined based on the correspondence between the intent and the action instruction model and business form model.

If the operation corresponding to the first action instruction model is to display a business form model, the target business form model can be displayed on the user's terminal device, so that the user can fill in the business form based on the target business form displayed on the terminal device.

In practical applications, due to the low popularity of smart devices in rural areas, some user terminal devices may not have a human-computer interaction interface, and it may be difficult to display the target business form template; therefore, the embodiment of the present invention displays the target business form model Previously, it was possible to determine the target device type of the end device displaying the target business form model.

As an example, the target device type may include a first type without a human-computer interaction interface, and a second type with a human-computer interaction interface.

Step 206, when the target device type is the first type, display the target business form model through voice interaction.

If the target device type of the terminal device is the first type, it may indicate that the terminal device does not have a human-computer interaction interface; at this time, the target business form model may be displayed through voice interaction.

For example: through TTS (Text To Speech, from text to speech) technology, according to the information in the target business form model, the corresponding voice prompt tone can be synthesized, and played to the user through the terminal device to prompt the user to fill in the content and format .

Step 207, when the target device type is the second type, display the target business form model on the terminal device.

If the target device type of the terminal device is the second type, it can indicate that the terminal device has a human-computer interaction interface; at this time, the target business form model can be directly displayed on the terminal device, so that the user can perform operations based on the displayed target business form model. fill in.

Step 208: During the process of displaying the target business form model, receive the second voice instruction, and use the intent model to determine the second target intent corresponding to the second voice instruction.

During the process of displaying the target business form model by the terminal device, the user may perform an operation of generating a second voice command for the content in the target business form model; the terminal device may generate a corresponding second voice command based on what the user said.

After the terminal device generates the second voice instruction, it can upload the second voice instruction to the server, so that the server can use the intent model to determine the second target intent corresponding to the second voice instruction.

As an example, when the target business form model is displayed through voice interaction, the voice instruction received after each output voice may be used as the second voice instruction.

As another example, when the target business form model is presented through display, the voice instruction received during the display process may be used as the second voice instruction.

Step 209: The target business form model includes a plurality of fields. When the second action instruction model corresponding to the second target intention is a fill-in business form model, determine the target field targeted by the second voice command, and from the second voice command, First text information for the target field is extracted.

In practical applications, the target business form model may include multiple fields to be filled in; therefore, after receiving the second voice command, the field that the user currently needs to fill in can be determined based on the second voice command; for example: target business form The model includes the following fields: "name", "age", "sex", "address", "ID number"; when the second voice command is "My ID number is 123456789", then "ID number The "Number" field is used as the target field, which is not limited in this embodiment of the present invention.

After the target field is determined, the first text information for the target field can also be extracted from the second voice instruction; following the above example, "123456789" can be used as the first text information.

Step 210: Fill the first text information into the target field according to the target algorithm in the second action instruction model.

After determining the first text information, the server may fill the first text information into the target field according to the target algorithm in the second action instruction model. For example: the first text information can be directly filled into the target field; or, when the target field is selected for filling, a target option is determined from the options corresponding to the target field according to the first text information to be filled. No limit.

After all the fields in the target business form model are filled, the target business form model can be uploaded to a corresponding server for storage, which is not limited in this embodiment of the present invention.

In an embodiment of the present invention, after filling the target field, the following steps may also be included:

Detect whether there is an abnormality in the filling content of the filled target field; when there is an abnormal filling content in the filled target field, prompt for the target field; to save.

In order to avoid submitting a business form filled with wrong information, it is possible to detect whether there is an abnormality in the filled target field. For example: overlapping time, typos, wrongly formatted sentences, and sentences inconsistent with questions and answers, etc., which are not limited in this embodiment of the present invention.

If there is an abnormal filling content, a prompt can be given for the target field to remind the user that the target field is filled in incorrectly and needs to be filled in again; when prompting, a prompt can be given based on the type of error to inform the user how to modify it.

If there is no exception in the filling content, the data filled into the target field can be saved; for example: when the data filled into the target field is the first text information, the first text information can be included under the target field; When the data of the field is a target option determined from the options corresponding to the target field according to the first text information, the target option may be included under the target field, which is not limited in this embodiment of the present invention.

In an embodiment of the present invention, the following steps may also be included:

When the first target intent cannot be determined and/or the second target intent cannot be determined, record the first abnormal log data; update the intent model according to the first abnormal log data;

If the intent model cannot determine the first target intent based on the first voice command, and/or, cannot determine the second target intent based on the second voice command, it may indicate that an unrecognizable voice (such as a dialect) appears at this time, a failed intent Analysis; at this time, the first abnormal log data can be recorded for the abnormal process and saved in the database.

When the intent model needs to be updated, the cause of the exception can be analyzed based on the first exception log data; then the intent model can be updated based on the cause of the exception; for example: adding a dialect voice database, adjusting the corpus corresponding to the intent, etc. In order to further improve the correct rate of intent recognition and form filling.

In an embodiment of the present invention, the following steps may also be included:

When the target business form model is filled in according to the second action instruction model and the second voice instruction, when the filling operation is abnormal, the second abnormal log data is generated; multiple action instruction models are updated according to the second abnormal log data, and the intention The model is updated.

Of course, there may be errors in the process of filling in the business form model, for example: the user said what he wanted to fill in in a dialect, but there is no voice database corresponding to the dialect in the current database; or, when matching the action instruction model and /or an exception occurs when the action instruction model is executed; at this time, the second exception log data can be recorded for the abnormal process and stored in the database.

When the intention model or the action instruction model needs to be updated, the cause of the exception can be analyzed based on the second exception log data; then the intention model and the action instruction model can be updated based on the cause of the exception.

As an example, when updating the action instruction model, it is possible to adjust the correspondence between the intention in the intention model and the action instruction model, or adjust the calling method and required parameter information in the action instruction model. The embodiment of the present invention There is no limit to this.

Figure 3 shows a schematic diagram of a scenario of an embodiment of the present invention; wherein, the terminal device can perform voice interaction with the user to fill in the business form model; the server can perform recognition based on the voice instruction uploaded by the terminal device, and Save the completed business form model.

FIG. 4 shows a schematic flow chart of a filling method of a business form according to an embodiment of the present invention;

S1. Dynamically design the business form model, intent model, and action command model that need to be filled in, and establish an intent model library and an intent sentence library based on rural business. And associate the business form model and the action instruction model with the intent model respectively.

Intent model design:

1) Create corresponding intent models based on business forms, such as: affirmation, negation, questioning, providing information, handling, comparing, suggesting, thanking, etc.;

2) After the intent model is created, upload the corresponding intent corpus, which is used for intent recognition based on the text;

3) After uploading the corpus, the intention model training is carried out through algorithms such as convolutional neural network and recurrent neural network, and the algorithm parameters can be fine-tuned according to the characteristics of business data.

Intent sentence library design:

1) Establish the intention sentence data structure, including: sentence pattern ID, sentence pattern name, sentence pattern description, sentence pattern formula, associated intention model, keyword set, and create intention sentence pattern database management functions according to the data structure, including: new , edit, delete functions;

2) Basic data entry, according to the actual business to enter information into the intention sentence database. For example: {I} want to handle {provident fund}{extraction} business; {I} want to handle {social security}{query} business.

Dynamic business form model design:

1) Create a business form model in a visual way. The business form model includes: form name, form ID stored in the database, and form category;

2) Create a form field and store it in the database. The form field information includes: form field ID, form field identification, data type, field classification;

3) Establish a form field classification type and keyword identification mapping;

4) Establish a mapping relationship between the form ID and the intent model, and the relationship is many-to-many.

Action instruction model design:

Action instructions are methods in programming languages, which can control the current operation of the application, such as jumping to a specified link address, popping up a window, closing the current page, submitting a form, and so on. Save information such as method calling method and required parameters to the database, including: action command ID, method name, return value type, parameter table;

Associate the action instruction ID with the intent model.

S2. The user wakes up the device terminal (for example, a smart terminal device) by voice, and the terminal uploads the voice command to the server.

S3. After receiving the voice command, the server uses the ASR technology to recognize the voice command into text by calling the voice recognition interface, and extracts keywords through NLP technology to identify the intent of the current keyword.

Intent recognition:

Intent recognition based on word segmentation, part-of-speech tagging, named entity recognition, dependency syntax analysis, semantic analysis or rule analysis.

Among them, the intent of the query can be obtained through rule analysis.

For example: "{I} want to apply for the minimum living allowance", you can use the rule template to extract the result: [user]: [handling] [the minimum living allowance]. The action command associated with [handling] is [pop-up form], and the form associated with [dibao] is [form for subsistence allowance].

Then, according to the type of device used by the current user, corresponding operations can be performed to determine whether to enable multiple rounds of interactive responses for form filling.

S4. After the intent recognition is completed, the corresponding action instruction model and business form model are matched, that is, the operation of opening the business form model related to the intent, jumping to the corresponding link, and interactive session is performed according to the action instruction model corresponding to the intent. Fill in the related business form model, record the filling process, and save it in the form of a log.

Match business form model:

Through intent recognition, analyze whether the intent belongs to the form operation intent. If so, query the form ID associated with the intent according to the intent data, and then judge whether to jump to the form operation interface according to the type of smart device. If the smart device has a drone interface, such as a smart speaker, the system will use multiple rounds of conversational interaction to fill in the form content.

Fill out forms by voice

When filling out the form, it will first confirm whether the smart device used by the user has a man-machine interface. If it has a man-machine interface, the system will enter the corresponding form interface to fill in, and the user can perform voice input according to the content of the form.

If there is no man-machine interface or if it is difficult to use the man-machine interface, the system will enable multiple rounds of conversational interaction to complete the form content. The system will set the number of sessions according to the number of fields in the form, and interact one by one. The specific method is: use TTS technology to synthesize voice according to the field name and prompt the user to fill in the content and format. After the user completes the input, the system will verify and save it.

After all the form data is filled in, the form data is analyzed, and prompts are given for filling in the problematic data. After the user completes the modification, the form is automatically sent back to the background; the analysis scope of the problem data includes overlapping time, typos, and wrongly formatted sentences Statements inconsistent with the question and answer.

S5. The server background analyzes the difficult-to-recognize voice and the failed intent analysis process during the form filling process, and updates the business form model, intent model and action instruction model to improve the correct rate of intent recognition and form filling.

In the embodiment of the present invention, it is possible to pre-create business form models for different businesses, an intent model for identifying intentions for the business form models, and multiple action instruction models for performing filling-related operations for the business form models; receiving the first A voice command, converting the first voice command into corresponding second text information; using natural language processing to extract target keywords for intent recognition from the second text information; using an intent model to determine the target keyword corresponding A target intent; when the first action instruction model corresponding to the first target intent is a display business form model, determine the target device type of the terminal device used to display the target business form model; when the target device type is the first type, Display the target business form model by way of voice interaction; when the target device type is the second type, display the target business form model in the terminal device; receive the second voice instruction and adopt the intent during the process of displaying the target business form model The model determines the second target intent corresponding to the second voice instruction; when the second action instruction model corresponding to the second target intent is to fill in a business form model, determine the target field targeted by the second voice command, and select from the second voice command , extract the first text information for the target field; fill the first text information into the target field according to the target algorithm in the second action instruction model. Through the embodiment of the present invention, it is realized based on speech recognition to help users complete the automatic filling of business forms without manual input by users; it not only improves the efficiency of users filling in business forms, but also reduces the learning cost for users to fill in business forms; and Based on the intention of the voice, the filling of the business form is automatically completed, which improves the correct rate of filling the user's business form.

In addition, the way of displaying the business form model is determined based on the type of the terminal device, so that the terminal device without a human-computer interaction interface can also fill in the business form based on voice interaction.

It should be noted that, for the method embodiment, for the sake of simple description, it is expressed as a series of action combinations, but those skilled in the art should know that the embodiment of the present invention is not limited by the described action sequence, because According to the embodiment of the present invention, certain steps may be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification belong to preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

Referring to Fig. 5, it shows a schematic structural diagram of a device for filling in a business form according to an embodiment of the present invention, which may include the following modules:

The creation module 501 is used to create business form models for different businesses, an intention model for identifying intentions for the business form models, and multiple action instruction models for performing filling-related operations for the business form models; in the intention model There is a corresponding relationship between the intent and the business form model and the action instruction model;

The first intention determination module 502 is configured to receive the first voice instruction, and use the intention model to determine the first target intention corresponding to the first voice instruction;

A display module 503, configured to display a target business form model corresponding to the first target intent when the first action instruction model corresponding to the first target intent is a display business form model;

The second intention determination module 504 is configured to receive the second voice instruction during the process of displaying the target business form model, and determine the second target intention corresponding to the second voice instruction by using the intention model;

The filling module 505 is configured to fill in the target business form model according to the second action command model and the second voice command when the second action instruction model corresponding to the second target intent is related to filling the target business form model.

In an optional embodiment of the present invention, the display module 503 includes:

The target device type determination submodule is used to determine the target device type of the terminal device used to display the target business form model;

The voice interaction display submodule is used to display the target business form model through voice interaction when the target device type is the first type;

The display submodule is used to display the target business form model in the terminal device when the target device type is the second type.

In an optional embodiment of the present invention, the device also includes:

The first update module is used to record the first abnormal log data when the first target intent cannot be determined and/or the second target intent cannot be determined;

updating the intent model according to the first exception log data;

In an optional embodiment of the present invention, the device also includes:

The second line-changing module is used to generate second abnormal log data when the target business form model is filled in according to the second action instruction model and the second voice instruction when the filling operation is abnormal; according to the second abnormal log data, a plurality of actions The instruction model is updated, and the intent model is updated.

In an optional embodiment of the present invention, the target business form model includes multiple fields, and the filling module 505 is configured to determine the target field for the second voice instruction, and extract the first field for the target field from the second voice instruction. A text information; fill the first text information into the target field according to the target algorithm in the second action instruction model.

In an optional embodiment of the present invention, the device also includes:

The abnormality detection module is used to detect whether there is an abnormality in the filling content of the filled target field; when there is an abnormal filling content in the filled target field, a prompt is given for the target field; when there is no abnormal filling content in the filled target field, the The data filled into the target field is saved.

In an optional embodiment of the present invention, the first intent determination module 502 is configured to convert the first voice instruction into corresponding second text information; use natural language processing to extract the second text information for intent recognition target keyword; using an intent model to determine the first target intent corresponding to the target keyword.

In the embodiment of the present invention, it is possible to pre-create business form models for different businesses, an intention model for identifying intentions for the business form models, and multiple action instruction models for performing filling-related operations for the business form models; the intention model There is a corresponding relationship between the intentions in the business form model and the action instruction model; receive the first voice instruction, and use the intention model to determine the first target intention corresponding to the first voice instruction; when the first action instruction corresponding to the first target intention When the model is a display business form model, display the target business form model corresponding to the first target intent; in the process of displaying the target business form model, receive the second voice instruction, and use the intent model to determine the second voice instruction corresponding to the second voice instruction. Target intent: when the second action instruction model corresponding to the second target intent is related to filling the target business form model, fill in the target business form model according to the second action instruction model and the second voice instruction. Through the embodiment of the present invention, it is realized based on speech recognition to help users complete the automatic filling of business forms without manual input by users; it not only improves the efficiency of users filling in business forms, but also reduces the learning cost for users to fill in business forms; and Based on the intention of the voice, the filling of the business form is automatically completed, which improves the correct rate of filling the user's business form.

The embodiment of the present invention also provides an electronic device, including a processor, a memory, and a computer program stored on the memory and capable of running on the processor. When the computer program is executed by the processor, the above method for filling in the business form is realized.

The embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above method for filling in a business form is realized.

As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and for related parts, please refer to the part of the description of the method embodiment.

Each embodiment in this specification is described in a progressive manner, each embodiment focuses on the difference from other embodiments, and the same and similar parts of each embodiment can be referred to each other.

Those skilled in the art should understand that the embodiments of the present invention may be provided as methods, devices, or computer program products. Accordingly, embodiments of the invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, embodiments of the invention may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

Embodiments of the present invention are described with reference to flowcharts and/or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and a combination of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor or processor of other programmable data processing terminal equipment to produce a machine such that instructions executed by the computer or processor of other programmable data processing terminal equipment Produce means for realizing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the The instruction means implements the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions can also be loaded into a computer or other programmable data processing terminal equipment, so that a series of operational steps are performed on the computer or other programmable terminal equipment to produce computer-implemented processing, thereby The instructions executed above provide steps for implementing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

Having described preferred embodiments of embodiments of the present invention, additional changes and modifications can be made to these embodiments by those skilled in the art once the basic inventive concept is appreciated. Therefore, the appended claims are intended to be construed to cover the preferred embodiment and all changes and modifications which fall within the scope of the embodiments of the present invention.

Finally, it should also be noted that in this text, relational terms such as first and second etc. are only used to distinguish one entity or operation from another, and do not necessarily require or imply that these entities or operations, any such actual relationship or order exists. Furthermore, the term "comprises", "comprises" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article, or terminal equipment comprising a set of elements includes not only those elements, but also includes elements not expressly listed. other elements identified, or also include elements inherent in such a process, method, article, or end-equipment. Without further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article or terminal device comprising said element.

The above provides a detailed introduction to the filling method, device, electronic equipment and storage medium of a business form provided. In this paper, specific examples are used to illustrate the principle and implementation of the present invention. The description of the above embodiments is only It is used to help understand the method of the present invention and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation and scope of application. In summary, this The content of the description should not be construed as limiting the present invention.

Claims (10)

1. A method for filling out a business form, characterized in that the method comprises: Create a business form model for different businesses, an intent model for identifying the intent of the business form model, and a plurality of action instruction models for performing filling-related operations for the business form model; in the intent model There is a corresponding relationship between the intention and the business form model and the action instruction model; receiving a first voice instruction, and using the intention model to determine a first target intention corresponding to the first voice instruction; When the first action instruction model corresponding to the first target intent is displaying a business form model, displaying a target business form model corresponding to the first target intent; In the process of displaying the target business form model, receiving a second voice instruction, and using the intent model to determine a second target intent corresponding to the second voice instruction; When the second action instruction model corresponding to the second target intention is related to filling the target business form model, perform a filling operation on the target business form model according to the second action instruction model and the second voice instruction . 2. The method according to claim 1, wherein the displaying the target business form model corresponding to the first target intent comprises: determining the target device type of the terminal device used to display the target business form model; When the target device type is the first type, displaying the target business form model through voice interaction; When the type of the target device is the second type, the target business form model is displayed on the terminal device. 3. The method according to claim 1, characterized in that the method further comprises: recording first exception log data when the first target intent cannot be determined and/or the second target intent cannot be determined; The intent model is updated according to the first exception log data. 4. The method according to claim 1, wherein the method further comprises: When the filling operation of the target business form model according to the second action instruction model and the second voice instruction is abnormal, generating second abnormal log data; The multiple action instruction models are updated according to the second exception log data, and the intention model is updated. 5. The method according to claim 1, wherein the target business form model comprises a plurality of fields, and the target business form model is set according to the second action instruction model and the second voice instruction Perform filling operations, including: determining the target field targeted by the second voice instruction, and extracting the first text information for the target field from the second voice instruction; Filling the first text information into the target field according to the target algorithm in the second action instruction model. 6. The method according to claim 5, further comprising: Detect whether there is any abnormality in the filled target field; When there is an abnormality in the filled target field, a prompt is given for the target field; When there is no abnormal filling content in the filled target field, the data filled in the target field is saved. 7. The method according to any one of claims 1-6, wherein the determining the first target intention corresponding to the first voice instruction by using the intention model comprises: converting the first voice instruction into corresponding second text information; extracting target keywords for intent recognition from the second text information by using natural language processing; The first target intent corresponding to the target keyword is determined by using the intent model. 8. A device for filling out a business form, characterized in that the device comprises: Create a module for creating business form models for different businesses, an intent model for identifying intentions for the business form model, and multiple action instruction models for performing filling-related operations for the business form model; There is a corresponding relationship between the intents in the intent model and the business form model and action instruction model respectively; A first intention determination module, configured to receive a first voice instruction, and use the intention model to determine a first target intention corresponding to the first voice instruction; A display module, configured to display a target business form model corresponding to the first target intent when the first action instruction model corresponding to the first target intent is a display business form model; The second intention determination module is configured to receive a second voice instruction during the process of displaying the target business form model, and determine a second target intention corresponding to the second voice instruction by using the intention model; A filling module, configured to, when the second action instruction model corresponding to the second target intention is related to filling the target business form model, complete the target business according to the second action instruction model and the second voice instruction The form model performs filling operations. 9. An electronic device, characterized by comprising a processor, a memory, and a computer program stored on the memory and capable of running on the processor, when the computer program is executed by the processor, the The method of filling out the business form mentioned in any one of requirements 1 to 7. 10. A computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the business form as described in any one of claims 1 to 7 is realized How to fill in.

Images