Form design method and system and electronic equipment
The form design method, which is trained using a large language model and optimized based on user feedback, solves the problems of low efficiency and poor flexibility in traditional medical form design, and realizes the automation and intelligence of forms, thereby improving design efficiency and user experience.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-12
- Publication Date
- 2026-04-07
AI Technical Summary
Traditional medical form design is inefficient and lacks flexibility, failing to meet the high requirements and efficiency needs of modern business operations.
A large language model training method is adopted. By acquiring training samples, preprocessing, optimizing the loss function and evaluation metrics, the target form is generated. The model is updated based on user feedback, thereby realizing the automated and intelligent design of the form.
It improves the efficiency and automation of form design, reduces manual operations, enhances user experience, and enables rapid response to business changes.
Smart Images
Figure CN121809427A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of medical informatics, and in particular to a form design method, system, and electronic device. Background Technology
[0002] In today's information age, forms play a crucial role in the medical field as a vital tool for data collection and processing. Traditional form design typically requires manual coding or configuration, which is not only time-consuming but also lacks flexibility. Furthermore, with the diversification of data types and the increasing complexity of business needs, traditional form design can no longer meet the high demands and efficiency requirements of modern business operations. Summary of the Invention
[0003] The technical problem to be solved by this disclosure is to overcome the shortcomings of low efficiency and poor flexibility in form processing in the medical field in the prior art, and to provide a form design method, system and electronic device.
[0004] This disclosure solves the above-mentioned technical problems through the following technical solution:
[0005] This disclosure provides a form design method for application in the medical field, the form design method comprising:
[0006] Obtain the form information to be generated; wherein, the form information to be generated includes configuration form rules and form data;
[0007] The form information to be generated is input into the large language model to output the target form.
[0008] Preferably, the large language model is trained through the following steps:
[0009] Obtain training samples; the training samples include form information from medical corpora;
[0010] The training samples are used as input to the large language model, and the actual forms are used as output to train the large language model, thus obtaining a trained large language model.
[0011] Preferably, the large language model is trained through the following steps:
[0012] The training samples are preprocessed, and the preprocessing operation includes at least one of removing redundant data, removing erroneous data, removing inconsistent data, and parsing data.
[0013] Preferably, the large language model is trained through the following steps:
[0014] The loss function is obtained based on the training samples and the actual form;
[0015] Optimize the large language model based on the loss function;
[0016] And / or,
[0017] The large language model is optimized based on an optimization algorithm;
[0018] And / or,
[0019] The performance of the large language model is evaluated based on evaluation metrics, wherein the evaluation metrics include at least one of accuracy and F1 score.
[0020] Based on the evaluation results, the performance of the large language model is improved through hyperparameter optimization methods; wherein, the hyperparameter optimization methods include at least one of network search, random search, and Bayesian optimization.
[0021] Preferably, the form design method further includes:
[0022] Update the large language model based on user feedback or update information from the form to be generated;
[0023] The user feedback information is used to represent the user's optimization suggestions for the large language model.
[0024] Preferably, the large language model is trained using one of the following model architectures: Transformer, BERT, or GPT.
[0025] This disclosure also provides a form design system for use in the medical field, the form design system comprising:
[0026] The acquisition module is used to acquire the form information to be generated; wherein, the form information to be generated includes configuration form rules and form data;
[0027] The form module is used to input the form information to be generated into the large language model in order to output the target form.
[0028] This disclosure also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and for running on the processor, wherein the processor executes the computer program to implement the form design method described above.
[0029] This disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described form design method.
[0030] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described form design method.
[0031] Based on common knowledge in the field, the above-mentioned preferred conditions can be combined arbitrarily to obtain various preferred embodiments of this disclosure.
[0032] The positive and progressive effects of this disclosure are as follows:
[0033] This disclosure utilizes large-scale model artificial intelligence technology, leveraging its generative and knowledge-reasoning capabilities, to automatically identify the characteristics of input data and intelligently generate form controls, attributes, and layouts. This achieves automated and intelligent form design to meet the needs of diverse medical forms. It improves the efficiency and automation of form design, enables rapid response to business changes, reduces manual operations, and enhances user experience. Attached Figure Description
[0034] Figure 1 A flowchart of a form design method provided in Embodiment 1 of this disclosure;
[0035] Figure 2 A flowchart illustrating a specific example of a form design method provided in Embodiment 1 of this disclosure;
[0036] Figure 3 This is a schematic diagram of the structure of a form design system provided in Embodiment 2 of this disclosure;
[0037] Figure 4 This is a schematic diagram of the structure of an electronic device provided in Embodiment 3 of this disclosure. Detailed Implementation
[0038] The present disclosure is further illustrated below by way of embodiments, but the present disclosure is not limited to the scope of the embodiments described herein.
[0039] In this embodiment of the disclosure, the collection, storage, use, processing, transmission, provision, and disclosure of user personal information comply with relevant laws and regulations and do not violate public order and good morals.
[0040] Example 1
[0041] In recent years, large-scale model technology has developed rapidly, achieving significant breakthroughs in algorithms, computing power, and data. It possesses the ability to process multi-modal data, such as text, images, and speech, and its powerful generation and generalization capabilities provide new ideas for form design. This embodiment, by introducing a large language model, can achieve intelligent transformation and implementation of form requirements.
[0042] This embodiment is applied in the medical field; see [link / reference]. Figure 1 Form design methods include:
[0043] S1. Obtain the form information to be generated.
[0044] The form information to be generated includes configuration form rules and form data.
[0045] Form data is used to represent the data required to generate the form content, and form configuration rules are used to represent the relationships between form data.
[0046] S2. Input the form information to be generated into the large language model to output the target form.
[0047] In one alternative implementation, the large language model is trained using one of the following model architectures: Transformer (a neural network architecture), BERT (a pre-trained model based on a Transformer encoder), or GPT (a pre-trained model based on a Transformer decoder).
[0048] This embodiment utilizes large-scale model artificial intelligence technology, leveraging its generative and knowledge-reasoning capabilities to automatically identify the characteristics of input data and intelligently generate form controls, attributes, and layouts. This automates and intelligently designs forms to meet the needs of diverse medical applications. It improves the efficiency and automation of form design, enabling rapid response to business changes, reducing manual operations, and enhancing user experience.
[0049] In one optional implementation, the large language model is trained through the following steps:
[0050] S3. Obtain training samples.
[0051] The training samples include form information from medical corpora.
[0052] S4. Use training samples as input to the large language model and actual forms as output to train the large language model and obtain the trained large language model.
[0053] In an optional implementation, the training steps of the large language model further include:
[0054] S31. Perform preprocessing operations on the training samples.
[0055] The preprocessing operations include at least one of the following: removing redundant data, removing erroneous data, removing inconsistent data, and parsing data.
[0056] In addition, preprocessing operations may include annotating the form data, including the layout, control types, and attributes of the form to be generated.
[0057] In this embodiment, preprocessing ensures higher quality and accuracy of the obtained form information to be generated, thereby ensuring that the generated form better meets the design requirements.
[0058] In an optional implementation, the training steps of the large language model further include:
[0059] S32. Obtain the loss function based on the training samples and the actual form.
[0060] S33. Optimize large language models based on loss functions.
[0061] In an optional implementation, the training steps of the large language model further include:
[0062] S34. Optimize large language models based on optimization algorithms.
[0063] Among them, the optimization algorithms include Adam (Adaptive Moment Estimation, an adaptive learning rate algorithm) and SGD (Stochastic Gradient Descent).
[0064] In an optional implementation, the training steps of the large language model further include:
[0065] S35. Evaluate the performance of large language models based on evaluation metrics.
[0066] The evaluation metrics include at least one of accuracy and F1 score (which is the harmonic mean of precision and recall, mainly used to measure the overall performance of binary classification models on imbalanced datasets).
[0067] S36. Based on the evaluation results, improve the performance of large language models through hyperparameter optimization methods.
[0068] The hyperparameter optimization methods include at least one of network search, random search, and Bayesian optimization.
[0069] In this implementation, the large language model can be continuously optimized through loss functions, optimization algorithms, and evaluation metrics, thereby improving the quality and accuracy of the generated target form.
[0070] In an optional implementation, the form design method further includes:
[0071] S5. Update the large language model based on user feedback or update information from the form to be generated.
[0072] User feedback information is used to represent users' suggestions for optimizing the large language model.
[0073] In this embodiment, user feedback and form update information can be obtained through user surveys and online feedback systems. By organizing and analyzing the collected opinions, the main problems encountered by users in using the form can be identified, and the problems can be solved to update the large language model. As a result, the target form generated by the updated large language model is more practical.
[0074] The following is a specific example, applied to the field of medical information technology. Combined with the application example of the scientific research disease form, it further illustrates the form design method of this embodiment.
[0075] Figure 2 Here is the flowchart for this example:
[0076] S101. Collect the form datasets needed for scientific research on specific diseases from various hospitals, i.e., obtain training samples.
[0077] Preprocessing of training samples includes correcting typos and grammatical errors, standardizing form data formats, manually verifying the accuracy of form control relationships, and constructing a large model training dataset.
[0078] It's important to note that building a large model training dataset—that is, collecting data to construct a dataset for training a large model—is a complex and meticulous process that directly impacts the model's final performance and effectiveness. First, data collection is crucial. This can involve collecting and organizing various forms from different business systems, or manually creating forms based on specific business scenarios. Second, data preprocessing is required. This involves parsing the raw form data and removing redundancy, errors, or inconsistencies. Third, the form data needs to be labeled, including its layout, control types, and attributes. This may require manual intervention to ensure data quality and accuracy. Finally, the dataset needs to be partitioned into training, validation, and test sets to facilitate performance evaluation and optimization during model training.
[0079] S102. Then, the large form model (i.e., the large language model) is trained in a distributed environment.
[0080] Building upon open-source models such as Qwen2.5 (a comprehensive model for understanding thousands of questions), further pre-training is performed using medical corpora to enable the model to learn specific vocabulary, grammar, and semantic information in the medical field. Fine-tuning is then applied to the pre-trained model to give it the ability to recognize and infer form content.
[0081] It's important to note that training a large-scale form model—that is, training a large model using a constructed large-scale form model dataset to enable it to generate form data items, infer field types, and attribute inferences—requires selecting a model architecture suitable for form data processing based on the actual business scenario, such as Transformer, BERT, or GPT. Consideration should be given to model complexity and computational requirements. Model hyperparameters, such as hidden layer size, number of attention heads, and number of layers, should be set according to computational resources and target performance. Training data should be input into the model in multiple batches, performing forward and backward propagation, and continuously adjusting model parameters to minimize the loss function. Optimization algorithms (such as Adam and SGD) should be used to adjust model parameters and control the learning rate to stabilize the training process. Model performance should be evaluated periodically on the validation set to monitor overfitting. Metrics such as accuracy and F1 score should be used. Hyperparameters can be adjusted through network search, random search, or Bayesian optimization to improve model performance.
[0082] S103. Accept and analyze the input form design requirements.
[0083] For example, researchers in the Acute Pancreatitis Specialty Database input the form information they need to collect for their research. This includes basic patient information such as name, age, gender, admission date, discharge date, length of hospital stay, smoking history, alcohol consumption, history of fatty liver disease, and clinical laboratory indicators such as serum albumin, C-reactive protein, and red blood cell distribution width. Upon receiving the form design requirements, the database uses a large-scale form model for data analysis. The analysis divides the form into basic information and clinical laboratory indicators, including complete blood count and biochemical tests, and identifies the type and attributes of each item within each section.
[0084] It's important to explain the data input and analysis process, which involves receiving user form design requirements and performing in-depth analysis using a large-scale model. Users input form design information, including the form's purpose and content, such as text fields, numeric fields, and date fields. The system receives this information and uses large-scale model AI technology to perform in-depth data analysis, identifying the type and attributes of each data item in the form (including control names, formats, whether it's required, default values, normal value ranges, character limits, text size, layout, etc.) and the relationships between the data.
[0085] S104. Based on the analysis results, automatically generate form data that constitutes the form data items and construct the content of the current form view.
[0086] This includes the form's layout, text controls for name and medical history, numerical controls for age, and normal ranges and units for various test indicators.
[0087] It's important to explain the generation of the form's basic data, which involves automatically constructing the form view based on the data analysis structure. The basic form content is built according to the analysis results of the large model. This includes generating form data items and constructing the form view content, such as the form's layout, form control properties, and control order. This step fully utilizes machine learning algorithms to optimize the ease of form interaction with users.
[0088] S105. Based on the analysis of the data relationship in the form in step S103, generate configuration form rules such as the calculation rules for calculating the number of days of hospitalization based on the admission time and discharge time.
[0089] It should be noted that form rule settings involve further configuring form rules based on the relationships between data. This automatically configures display / hide rules, data validation rules, and calculation rules within the form based on the logical relationships between form controls. These logical relationships can be based on system default rules or manually defined rules.
[0090] S106. Form optimization.
[0091] For example, update the model regularly based on user feedback and new data, and optimize the form design, such as adjusting the layout of fields and modifying the interaction methods.
[0092] It's important to note that form optimization involves refining and adjusting the design of forms generated from large-scale models based on user feedback and business changes, thereby improving the usability and adaptability of the forms. This includes collecting user feedback on the forms through user surveys and online feedback systems. The collected feedback is then organized and analyzed to identify and address the main problems users encounter when using the forms.
[0093] Using the form design method in this example can improve the efficiency of form design, reduce manual workload, and enhance the automation and intelligence of data processing.
[0094] Example 2
[0095] Corresponding to the foregoing embodiments of form design methods, this disclosure also provides embodiments of a form design system. This embodiment is applied in the medical field; see [link to relevant documentation]. Figure 3 The form design system includes:
[0096] Module 1 is used to obtain the form information to be generated.
[0097] The form information to be generated includes configuration form rules and form data.
[0098] Form module 2 is used to input the form information to be generated into the large language model in order to output the target form.
[0099] In one alternative implementation, the large language model is trained using one of the following model architectures: Transformer, BERT, or GPT.
[0100] In one alternative implementation, see Figure 3 The form design system also includes:
[0101] Training module 3 is used to obtain training samples, which include form information from medical corpora. It is also used to train the large language model by using the training samples as input and the actual forms as output, thus obtaining a trained large language model.
[0102] In an optional implementation, the training module 3 is further configured to perform preprocessing operations on the training samples, the preprocessing operations including at least one of removing redundant data, removing erroneous data, removing inconsistent data, and parsing data.
[0103] In an optional implementation, training module 3 is also used to obtain a loss function based on training samples and actual forms; and to optimize the large language model based on the loss function.
[0104] In an optional implementation, training module 3 is also used to optimize the large language model based on an optimization algorithm.
[0105] In an optional implementation, training module 3 is further configured to evaluate the performance of the large language model based on evaluation metrics, wherein the evaluation metrics include at least one of accuracy and F1 score; and based on the evaluation results, improve the performance of the large language model through hyperparameter optimization methods, wherein the hyperparameter optimization methods include at least one of web search, random search, and Bayesian optimization.
[0106] In one alternative implementation, see Figure 3 The form design system also includes:
[0107] Optimization module 4 is used to update the large language model based on user feedback or update information from the form to be generated; the user feedback is used to represent the user's optimization suggestions for the large language model.
[0108] For the system embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The system embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs.
[0109] Example 3
[0110] Figure 4 This is a schematic diagram of the structure of an electronic device according to an example embodiment of the present disclosure. The electronic device includes a memory, a processor, and a computer program stored in the memory and used to run on the processor. When the processor executes the computer program, it implements the form design method of any of the above embodiments. Figure 4 The electronic device 40 shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments disclosed herein.
[0111] like Figure 4 As shown, the electronic device 40 can be manifested as a general-purpose computing device, such as a server device. The components of the electronic device 40 may include, but are not limited to: at least one processor 41, at least one memory 42, and a bus 43 connecting different system components (including memory 42 and processor 41).
[0112] Bus 43 includes a data bus, an address bus, and a control bus.
[0113] The memory 42 may include volatile memory, such as random access memory (RAM) 421 and / or cache memory 422, and may further include read-only memory (ROM) 423.
[0114] The memory 42 may also include a program tool 425 (or utility) having a set (at least one) program module 424, such program module 424 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.
[0115] The processor 41 executes various functional applications and data processing by running computer programs stored in the memory 42, such as the form design method provided in any of the above embodiments.
[0116] Electronic device 40 can also communicate with one or more external devices 44 (e.g., keyboard, pointing device, etc.). This communication can be performed via input / output (I / O) interface 45. Furthermore, electronic device 40 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public network, such as the Internet) via network adapter 46. As shown, network adapter 46 communicates with other modules of electronic device 40 via bus 43. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with electronic device 40, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems.
[0117] It should be noted that although several units / modules or sub-units / modules of the electronic device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.
[0118] Example 4
[0119] This disclosure also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the form design method provided in any of the above embodiments.
[0120] The readable storage medium may be more specifically adopted, including but not limited to: portable disk, hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.
[0121] Example 5
[0122] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the form design method described above.
[0123] The program code for executing the computer program product disclosed herein can be written in any combination of one or more programming languages. The program code can be executed entirely on a user device, partially on a user device, as a stand-alone software package, partially on a user device and partially on a remote device, or entirely on a remote device.
[0124] While specific embodiments of this disclosure have been described above, those skilled in the art should understand that these are merely illustrative examples, and the scope of protection of this disclosure is defined by the appended claims. Those skilled in the art can make various changes or modifications to these embodiments without departing from the principles and essence of this disclosure, but all such changes and modifications fall within the scope of protection of this disclosure.
Claims
1. A form design method, characterized in that, Applied to the medical field, the form design method includes: Obtain the form information to be generated; wherein, the form information to be generated includes configuration form rules and form data; The form information to be generated is input into the large language model to output the target form.
2. The form design method as described in claim 1, characterized in that, The large language model is trained through the following steps: Obtain training samples; the training samples include form information from medical corpora; The training samples are used as input to the large language model, and the actual forms are used as output to train the large language model, thus obtaining a trained large language model.
3. The form design method as described in claim 2, characterized in that, The large language model is trained through the following steps: The training samples are preprocessed, and the preprocessing operation includes at least one of removing redundant data, removing erroneous data, removing inconsistent data, and parsing data.
4. The form design method as described in claim 2, characterized in that, The large language model is trained through the following steps: The loss function is obtained based on the training samples and the actual form; Optimize the large language model based on the loss function; And / or, The large language model is optimized based on an optimization algorithm. And / or, The performance of the large language model is evaluated based on evaluation metrics, wherein the evaluation metrics include at least one of accuracy and F1 score. Based on the evaluation results, the performance of the large language model is improved through hyperparameter optimization methods; wherein, the hyperparameter optimization methods include at least one of network search, random search, and Bayesian optimization.
5. The form design method as described in claim 1, characterized in that, The form design method also includes: Update the large language model based on user feedback or update information from the form to be generated; The user feedback information is used to represent the user's optimization suggestions for the large language model.
6. The form design method as described in claim 1, characterized in that, The large language model is trained using one of the following model architectures: Transformer, BERT, or GPT.
7. A form design system, characterized in that, Applied to the medical field, the form design system includes: The acquisition module is used to acquire the form information to be generated; wherein, the form information to be generated includes configuration form rules and form data; The form module is used to input the form information to be generated into the large language model in order to output the target form.
8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and for running on the processor, characterized in that, When the processor executes the computer program, it implements the form design method according to any one of claims 1 to 6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the form design method according to any one of claims 1 to 6.
10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the form design method as described in any one of claims 1-6.