CN113488178A - Information generation method and device, storage medium and electronic equipment - Google Patents
Information generation method and device, storage medium and electronic equipment Download PDFInfo
- Publication number
- CN113488178A CN113488178A CN202110817163.9A CN202110817163A CN113488178A CN 113488178 A CN113488178 A CN 113488178A CN 202110817163 A CN202110817163 A CN 202110817163A CN 113488178 A CN113488178 A CN 113488178A
- Authority
- CN
- China
- Prior art keywords
- information
- executed
- generation model
- execution
- information generation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
Abstract
The embodiment of the invention discloses an information generation method, an information generation device, a storage medium and electronic equipment. The method comprises the steps of obtaining multi-dimensional characteristic information of a target object, inputting the multi-dimensional characteristic information into a current information generation model, and obtaining information to be executed of the target object; acquiring sign information and execution data of a target object in the execution process of the information to be executed, and performing evaluation processing on the sign information and the execution data to obtain an evaluation result of the information to be executed; and performing optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model, and updating the current information generation model based on the new information generation model. The multidimensional characteristic information is processed through the information generation model, the information generation efficiency is improved, and the dependence on manpower and manual experience is reduced. Meanwhile, the information generation model is in a cyclic updating process, so that the precision of the information generation model and the quality of the generated information to be executed are ensured.
Description
Technical Field
The embodiment of the invention relates to the field of deep learning, in particular to an information generation method, an information generation device, a storage medium and electronic equipment.
Background
With the continuous development of society, physical health becomes a social concern.
Rehabilitation therapy is a way of performing health therapy on a user through exercise, physical therapy and the like, the current information to be executed is generally set manually by a doctor, the quality of the information to be executed depends on the experience of the doctor, and meanwhile, a large amount of manual labor is consumed.
Disclosure of Invention
The embodiment of the invention provides an information generation method, an information generation device, a storage medium and electronic equipment, and aims to achieve the accuracy and high efficiency of information generation.
In a first aspect, an embodiment of the present invention provides an information generating method, including:
obtaining multi-dimensional characteristic information of a target object, inputting the multi-dimensional characteristic information into a current information generation model, and obtaining to-be-executed information of the target object, wherein the to-be-executed information comprises execution starting time, execution frequency and execution items of each frequency;
acquiring sign information and execution data of the target object in the execution process of the information to be executed, and performing evaluation processing on the sign information and the execution data to obtain an evaluation result of the information to be executed;
and performing optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model, and updating the current information generation model based on the new information generation model, wherein the updated current information generation model is used for generating the next information to be executed.
In a second aspect, an embodiment of the present invention further provides an information generating apparatus, including:
the system comprises an information to be executed acquisition module, a data processing module and a data processing module, wherein the information to be executed acquisition module is used for acquiring multi-dimensional characteristic information of a target object, inputting the multi-dimensional characteristic information into a current information generation model and acquiring information to be executed of the target object, and the information to be executed comprises execution starting time, execution frequency and execution items of each frequency;
the scheme evaluation module is used for acquiring the sign information and the execution data of the target object in the execution process of the information to be executed, and evaluating the sign information and the execution data to obtain an evaluation result of the information to be executed;
and the model optimization module is used for carrying out optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model, and updating the current information generation model based on the new information generation model, wherein the updated current information generation model is used for generating the next information to be executed.
In a third aspect, an embodiment of the present invention further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor executes the computer program to implement the information generating method according to any embodiment of the present invention.
In a fourth aspect, the embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the information generating method provided in any embodiment of the present invention.
According to the technical scheme provided by the embodiment, the comprehensiveness of the characteristic information is improved by collecting the multi-dimensional characteristic information of the target object, and the synchronous analysis of a plurality of rehabilitation parts or a plurality of symptoms of the target object is facilitated. The information generation model is set, the multi-dimensional characteristic information is processed, the information to be executed including the execution starting time, the execution frequency and the execution items of each frequency is generated, the generation efficiency is high, the accuracy is high, the dependence on manpower is reduced, and the dependence on the manual experience is reduced. Meanwhile, the information generation model is in a cyclic updating process, the called information generation model is the updated latest information generation model, and the precision of the information generation model and the quality of the generated information to be executed are guaranteed. Meanwhile, for the generated information to be executed, collecting sign information and execution data of the information to be executed in the execution process, and evaluating the information to be executed to determine the quality of the information to be executed. Further, the current information generation model is subjected to optimization training based on the evaluation result and the information to be executed, an updated information generation model is obtained, and the quality of the next information to be executed is guaranteed.
Drawings
Fig. 1 is a schematic flowchart of an information generating method according to an embodiment of the present invention;
fig. 2 is a schematic flowchart of an information generating method according to an embodiment of the present invention;
FIG. 3 is a schematic structural diagram of an evaluation model provided by an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an information generating apparatus according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of an electronic device according to a fourth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
Example one
Fig. 1 is a schematic flow diagram of an information generating method according to an embodiment of the present invention, where this embodiment is applicable to a case of automatically generating information to be executed of a target object, and the method may be executed by an information generating apparatus according to an embodiment of the present invention, where the information generating apparatus may be implemented by software and/or hardware, and the information generating apparatus may be configured on an electronic computing device, and specifically includes the following steps:
s110, obtaining multi-dimensional characteristic information of a target object, inputting the multi-dimensional characteristic information into a current information generation model, and obtaining information to be executed of the target object, wherein the information to be executed comprises execution starting time, execution frequency and execution items of each frequency.
And S120, acquiring sign information and execution data of the target object in the execution process of the information to be executed, and performing evaluation processing on the sign information and the execution data to obtain an evaluation result of the information to be executed.
S130, performing optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model, and updating the current information generation model based on the new information generation model, wherein the updated current information generation model is used for generating the next information to be executed.
In the embodiment, the information generation model is arranged, has an information prediction function, can output the information to be executed which is suitable for the characteristic information of the target object after processing the characteristic information of the target object, replaces the technical scheme of manually setting the information to be executed, reduces the experience dependence on operators, and ensures the accuracy of the information to be executed.
The information generation model in this embodiment may be, for example, but not limited to, a neural network module, a deep network module, and the like, which is not limited in this respect. In some embodiments, the information to be performed may be a rehabilitation protocol applicable to the target subject. Accordingly, the information generation model may be a schema generation module. It should be noted that the information generation model is not only applicable to an information generation model of a single rehabilitation site or a single disease, but may be adapted to generate corresponding information to be executed for at least one rehabilitation site (or at least one disease), in particular two or more rehabilitation sites (two or more diseases). In the case where two or more kinds of rehabilitation sites (or disorders) coexist in the target subject, execution information corresponding to any one rehabilitation site (or disorder) independently may be executed, and there may be an influence on other rehabilitation sites. According to the method and the device, the information to be executed suitable for the current state of the target object is obtained by comprehensively processing the current state of the target object.
Specifically, the input information for converting the multidimensional characteristic information into the information generation model may be input information of a matrix structure formed by converting the multidimensional characteristic information into corresponding word vectors or sentence vectors respectively and adding the word vectors or sentence vectors to corresponding positions of an input matrix of the information generation model. By inputting the multi-dimensional characteristic information, the information generation model can conveniently carry out comprehensive analysis on the characteristic information of multiple dimensions, and high-precision information to be executed suitable for the target object is obtained.
Optionally, the multi-dimensional feature information includes feature information of physical sign dimensions, feature information of treatment dimensions, and feature information of basic registration. The characteristic information of the physical dimension is various types of physical information of the target object, including but not limited to blood sugar information, blood pressure information, blood fat information, body temperature, electrocardiogram information, electroencephalogram information, and physical information associated with the at least one rehabilitation part. By collecting the sign information associated with the plurality of rehabilitation parts, the current state of the target object can be reflected through the collected sign information.
The characteristic information of the treatment dimension includes historical treatment information of the target object, and exemplarily, the characteristic information of the treatment dimension includes one or more of, but is not limited to, surgical treatment information, drug treatment information, and physical treatment information, wherein the surgical treatment information may include, but is not limited to, a surgical type, a surgical time, a surgical result, and the like, the drug treatment information may include, but is not limited to, a drug type, a drug multiplexing dose, a drug treatment time, and the like, and the physical treatment information may include, but is not limited to, a physical treatment type, a physical treatment frequency, a physical treatment time, and the like.
The characteristic information of the basic registration is to include registration information for authorization, wherein the characteristic information of the basic registration may include but is not limited to sex, age, height, weight, and the like.
The information generation model obtains information to be executed by processing the input information converted by the multi-dimensional characteristic information, wherein the information to be executed may include execution starting time, execution frequency and execution items of each frequency. Taking the information to be executed as the rehabilitation scheme as an example, the execution starting time is the first execution time of the rehabilitation scheme, for example, after n days from the current time, the execution starting time is set to optimize the execution time, so that adverse effects caused by unscientific execution of the information to be executed are avoided. The execution frequency is the execution frequency and the execution times, and the execution items of each frequency are the execution contents, such as lifting the leg N times, walking M meters, and the like. In some embodiments, the interval time of the adjacent times in the execution frequency may be fixed, and may also be variable, for example, the interval time of the adjacent times may be gradually shortened as the execution frequency increases, or the interval time of the adjacent times may be gradually lengthened as the execution frequency increases. Meanwhile, the execution items corresponding to different frequencies may be the same or different, for example, the execution items may include different execution contents and execution parameters of the execution contents, for example, the execution contents in the execution items of different frequencies may be different, and the execution contents in the execution items of different frequencies are the same, but the execution parameters of the execution contents are different.
In the embodiment, the information generation model is in a cyclic optimization process, when the multidimensional characteristic information of the target object is obtained, the current information generation model is called, the current information generation model is the updated latest information generation model, and the multidimensional characteristic information is processed based on the current information generation model, so that the accuracy of the generated information to be executed is improved.
In some embodiments, when obtaining the multi-dimensional feature information of the target object, if the current information generation model is updated, the updated information production model is invoked, if the current information production model is in the process of updating, the previous non-updated information production model is invoked, wherein, the information generation model can be updated, and the information generation model which is not updated is kept, so that in the updating process, the acquired multi-dimensional characteristic information is processed, the generation of the information to be executed is prevented from being influenced by the updating process of the information generation model, correspondingly, after the updated information generation model is obtained, the non-updated information generation model is deleted, and the subsequently acquired multi-dimensional characteristic information is processed on the basis of the updated information generation model, so that the information generation model is updated insensibly.
In some embodiments, the current information generation model is in an update process, and may also be configured to cache the acquired multidimensional feature information, and call the updated information generation model after the current information generation model is updated.
In this embodiment, by acquiring feature information in the execution process of the information to be executed, evaluating the information to be executed based on the feature information, and updating the current information generation model based on the evaluation result and the information to be executed, the accuracy of the information generation model is further improved, so that the updated information generation model generates high-accuracy information to be executed when being called next time.
During the execution of the execution items of the target subject in each frequency, sign information and execution data of the target subject are acquired, wherein the sign information may be a preset sign type, and may be determined based on the execution items, and different execution items may correspond to different feature information, and may include, but are not limited to, body temperature, blood pressure, heartbeat number, and the like, for example, which is not limited thereto. Similarly, the execution data may be data generated in the process of executing the item, for example, video data or image data of the execution item executed by the target object may be acquired by a camera, or the number of times of operations performed by the target object may be acquired by a counter, and the like, and the execution item may be, for example, 100 times of leg raising, and the number of times of leg raising of the target object may be recorded by a counter, or whether a leg raising criterion is satisfied may be detected by a height monitoring part provided at a leg portion of the target object, and the like. It should be noted that the execution data is determined based on the execution item, different execution items may correspond to different execution data, and correspondingly, the execution data is acquired by different data acquisition devices, which is not limited in this embodiment, and the electronic device that executes the information generation method in this embodiment is electrically connected or communicatively connected to each data acquisition device, and receives the execution data acquired by the data acquisition device.
In some embodiments, evaluating the to-be-executed information through the sign information and the execution data of the target subject may include: and determining whether the target object completes the execution items of each frequency based on the execution data, if so, comparing the sign information with sign information before the execution of the execution items of the frequency, and determining an evaluation result based on the comparison result of the sign information of each frequency. Specifically, if the execution data is video data, identifying an execution operation corresponding to an execution item in the video data, and determining attribute information of the execution operation, where the attribute information may be an execution parameter in the execution item, for example, the execution parameter may include, but is not limited to, an execution number, an execution duration, and the like, comparing the execution operation identified and the corresponding attribute information with the execution item, if the comparison is successful, determining that the corresponding execution item is completed, and if the comparison is failed, determining that the corresponding execution item is not completed. If the execution data is array data, the value includes an identifier of the execution operation and attribute information, and the attribute information may include, but is not limited to, execution times, execution duration, and the like, for example [ raise leg, 100 times, walk, 50m ], and whether the execution item is completed is determined by comparing the array data with the execution item. For a target object with an incomplete execution item, the change of the sign parameter of the target object cannot represent the quality of the information to be executed, i.e. the information to be executed cannot be evaluated.
Optionally, the sign information may be compared with the sign information before the execution of the execution item of the frequency, a variation and a variation trend of the sign information and the sign information before the execution of the execution item of the frequency may be determined, and a comparison result may be determined based on the variation and the variation trend, where the variation trend may include benign and malignant, for example, benign may be represented by "+", malignant may be represented by "-", and the sign and the variation are used as the comparison result, or a variation rate corresponding to the sign and the variation may also be used as the comparison result.
It should be noted that the sign information before the execution of the execution item of the frequency may be the sign information acquired after the execution of the execution item of the previous frequency.
Optionally, determining the evaluation result based on the comparison result of the sign information of each frequency may be: and accumulating the comparison results of each frequency, determining an accumulated comparison result, and taking the accumulated comparison result as an evaluation result, or determining an evaluation result based on the accumulated comparison result and a standard result, wherein the standard result can be standard characteristic information corresponding to the target, the standard characteristic information of the target classified by attributes such as each age group, sex, height, weight and the like is preset, and the standard characteristic information can be matched based on the attribute information of the target to determine the corresponding standard characteristic information. Optionally, determining the evaluation result based on the comparison result of the sign information of each frequency may further be: determining the variation trend of the comparison result of each frequency, and determining the evaluation result based on the variation trend.
In this embodiment, the evaluation result may be a level result, for example, a high level, a medium level, and a low level, where the higher the level is, the higher the quality of the corresponding to-be-executed information is; the evaluation result may also be a numerical result, which is in the range of 0-1, the larger the numerical value, the higher the quality of the corresponding information to be executed.
The information to be executed is evaluated, the quality of the information to be executed is represented based on the evaluation result, and the current information generation model is optimized according to the evaluation result, so that the precision of the current information generation model is improved, and the quality of the next information to be executed is further improved.
In some optional embodiments, performing optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model, including: and counting the information to be executed after the evaluation is completed within the preset time interval according to the preset time interval, and performing optimization training on the current information generation model based on the information to be executed after the evaluation is completed and the corresponding evaluation result to obtain a new information generation model. In this embodiment, the preset time interval may be one day, statistics is performed on information to be executed after evaluation in one day, the information is used as current sample data, and the current information generation model is optimally trained based on the current sample data. By setting the preset time interval, sample data in the preset time interval are accumulated, and the problem that the generation of the information to be executed is influenced by immediately performing optimization training on the information generation model when the information to be executed which is evaluated is obtained is avoided. Optionally, an idle period of the information generation model is determined, and the information generation model is optimally trained based on sample data formed in a preset time interval in the idle period. The idle time period can be preset, for example, 0:00-2:00 in the morning; the idle time may also be determined based on historical usage status of the information generation model, with the period of lowest usage determined as the idle period.
In some optional embodiments, performing optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model, including: matching the evaluation result with a threshold condition; and if the evaluation result does not meet the threshold condition, performing optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model. The threshold condition is a judgment condition for judging whether the information to be executed is used as a training sample, the threshold condition is the same as the data type of the evaluation result, for example, the threshold condition may be a level threshold or a data threshold, and for example, the threshold condition may be that the evaluation result is one level, or that the evaluation value is more than 90%, and may be set as required, which is not limited. If the evaluation result of the information to be executed meets the threshold condition, the quality of the information to be executed generated by the information generation model is high, the optimization degree of the information generation model is small, the information to be executed model belongs to an invalid sample, and if the evaluation result of the information to be executed does not meet the threshold condition, the quality of the information to be executed generated by the information generation model is low, and the optimization degree of the information generation model is large. And forming sample data from the information to be executed which does not meet the threshold condition, and performing optimization training on the current information generation model, so that the training process caused by invalid training can be reduced, and the efficiency and effectiveness of model training are improved.
On the basis of the above embodiment, the performing optimization training on the current information generation model based on the evaluation result and the to-be-executed information to obtain a new information generation model, including: matching corresponding standard information in a standard information base based on the multi-dimensional characteristic information of the target object; forming a loss function based on the information to be executed, the standard information, the evaluation result and the acquired sample data, performing parameter adjustment on a current information generation model based on the loss function, and generating a model by the new information.
In this embodiment, the current information generation model is optimally trained in a supervised manner, and corresponding standard information is determined for any sample data (i.e., information to be executed). The standard information base stores a plurality of standard information, the multi-dimensional characteristic information of the target object is matched in the standard information base, and the standard information is determined, for example, the standard information base stores characteristic information corresponding to the standard information in association with the standard information, and combinations of characteristic information of different age groups, sexes, heights, weights and the like respectively correspond to different standard information.
And calculating the similarity between the multi-dimensional characteristic information of the target object and each group of characteristic information in the standard information base, and determining the standard information corresponding to the characteristic information with the maximum similarity as the standard information corresponding to the information to be executed of the target object. Optionally, if the maximum similarity is smaller than the similarity threshold, it is prompted that no standard information exists, the externally input standard information is received, and the standard information is updated to the standard information base, so that the accuracy of the standard information is improved.
The acquired sample data comprises positive sample data and negative sample data, and the positive sample data and the negative sample data are respectively stored in the positive sample data set and the negative sample data set. The positive sample data or the negative sample data comprise multi-dimensional characteristic information of the historical user and information to be executed generated by the generation model based on the multi-dimensional characteristic information. The sample data formed by the multi-dimensional characteristic information of different historical users and the information to be executed can be positive sample data or negative sample data, and can be determined based on the evaluation result corresponding to the information to be executed. Specifically, if the evaluation result is a numerical result, the positive sample data and the negative sample data may be divided based on the evaluation threshold, for example, the sample data whose evaluation result is greater than or equal to the evaluation threshold is determined as the positive sample data, and the sample data whose evaluation result is less than the evaluation threshold is determined as the negative and positive sample data. If the evaluation result is a level result, the sample data of the first preset level may be determined as positive sample data, and the sample data of the second preset level may be determined as negative sample data. The first preset level may be one or more levels, the second preset level may be one or more levels, and there is no overlap between the first preset level and the second preset level.
And forming a loss function based on the information to be executed, the standard information, the evaluation result, the acquired sample data and a preset loss function generation rule. In this embodiment, the specific form of the loss function generation rule may not be limited, and the loss function generation rule may be set according to a user requirement.
Optionally, forming a loss function based on the information to be executed, the standard information, the evaluation result, and the obtained sample data, including: acquiring positive sample data and negative sample data, respectively traversing the positive sample data and the negative sample data to obtain a first sample distance and a second sample distance, and determining a first loss item based on the first sample distance and the second sample distance and corresponding multi-dimensional characteristic information; determining a second loss term based on the evaluation result; and performing weighted calculation based on the first loss term and the second loss term to form the loss function, wherein the weights of the first loss term and the second loss term are determined based on the information to be executed and the standard information. Specifically, generating functions of a first loss term and a second loss term are called, a first sample distance, a second sample distance and multi-dimensional feature information are input into the generating function of the first loss term to obtain the first loss term, and an evaluation result is input into the generating function of the second loss term to obtain the second loss term. The generating function of the second loss term may be a preset coefficient function of the evaluation result, and the generating function of the first loss term may be determined based on the first sample distance, a similarity function of a variation envelope function of the second sample distance, and the multi-dimensional feature information.
The first loss term and the second loss term respectively correspond to preset weights, and the first loss term and the second loss term are weighted and calculated based on the weights to form the loss function. In some optional embodiments, the information to be executed and the standard information may be processed through a preset formula, and the weight of the first loss term and the weight of the second loss term are obtained through calculation.
In some embodiments, the evaluation result further includes evaluation information input by the user side and evaluation information input by the monitoring side (e.g., a doctor side), which may be a mean value of the evaluation information input by the determination user side, the evaluation information input by the monitoring side, and the evaluation result determined by the sign information and the execution data in this embodiment, as a final evaluation result, and a first loss term is determined from the final evaluation result, so as to improve the comprehensiveness of the evaluation result.
For example, the loss function may be determined by:
b：＝f(Ai，j)
wherein A isi，jInformation of the target object, wherein i is information to be executed, j is standard information, and b is weight of the second loss itemThe value, 1-b is the weighted value of the first loss term, P is the positive sample data set, N is the negative sample data set, m, N are respectively the positive sample data, k, l are respectively the negative sample data set, Dm，kIs a first sample distance, Dn，lF (×) is an envelope similarity function, α is a margin parameter of the multidimensional characteristic information corresponding to the positive sample data, S is an evaluation result, C is the total number of samples in the positive sample data set P,
the loss function for the samples m, n,
as a final loss function.
The above-mentioned loss function is used to perform parameter adjustment on the current information generation model to obtain a new information generation model, where the parameter adjustment may be performed on the information generation model by a gradient descent method, and the parameters to be adjusted include, but are not limited to, a weight and an offset value.
And generating the updated current information generation model for generating the next to-be-executed information of the target object or for generating the to-be-executed information of the next execution object. And the next information to be executed is generated through the updated current information generation model, so that the quality of the next information to be executed is improved.
According to the technical scheme, the comprehensive characteristic information is improved by collecting the multi-dimensional characteristic information of the target object, and the synchronous analysis of multiple rehabilitation parts or multiple diseases of the target object is facilitated. The information generation model is set, the multi-dimensional characteristic information is processed, the information to be executed including the execution starting time, the execution frequency and the execution items of each frequency is generated, the generation efficiency is high, the accuracy is high, the dependence on manpower is reduced, and the dependence on the manual experience is reduced. Meanwhile, the information generation model is in a cyclic updating process, the called information generation model is the updated latest information generation model, and the precision of the information generation model and the quality of the generated information to be executed are guaranteed. Meanwhile, for the generated information to be executed, collecting sign information and execution data of the information to be executed in the execution process, and evaluating the information to be executed to determine the quality of the information to be executed. Further, the current information generation model is subjected to optimization training based on the evaluation result and the information to be executed, an updated information generation model is obtained, and the quality of the next information to be executed is guaranteed.
Example two
Fig. 2 is a schematic flow chart of an information generating method according to an embodiment of the present invention, and on the basis of the above embodiment, optionally, the evaluation processing is performed on the sign information and the execution data to obtain an evaluation result of the information to be executed, where the evaluation result includes: and evaluating the sign information and the execution data in the execution process of the execution items of each frequency based on an evaluation model to obtain an evaluation result of the information to be executed. The method comprises the following steps:
s210, obtaining multi-dimensional characteristic information of a target object, inputting the multi-dimensional characteristic information into a current information generation model, and obtaining information to be executed of the target object, wherein the information to be executed comprises execution starting time, execution frequency and execution items of each frequency.
S220, acquiring sign information and execution data of the target object in the execution process of the information to be executed.
And S230, evaluating the sign information and the execution data in the execution process of the execution items of each frequency based on the evaluation model to obtain the evaluation result of the information to be executed.
S240, performing optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model, and updating the current information generation model based on the new information generation model, wherein the updated current information generation model is used for generating the next information to be executed.
In this embodiment, an evaluation model is preset, and the evaluation model may be a lifting tree model, a logistic regression model, or a neural network module, which is not limited in this respect. The evaluation model evaluates the sign information and the execution data in the execution process of the execution items of each frequency, so that the evaluation process of the information to be executed is simplified, and the evaluation efficiency is improved.
In some embodiments, the evaluation model includes a first evaluation module and a second evaluation module, where the first evaluation module is configured to perform evaluation processing on the sign information and the execution data of the target object during execution of a single frequency to obtain an initial evaluation result, and the second evaluation module performs comprehensive evaluation processing on the initial evaluation results of each frequency to obtain an evaluation result of the information to be executed. By arranging the first evaluation module and the second evaluation module, evaluation processing is respectively carried out on sign information and execution data of a single frequency, and comprehensive evaluation processing is carried out on initial evaluation results of each frequency, so that evaluation precision is improved.
For example, referring to fig. 3, fig. 3 is a schematic structural diagram of an evaluation model provided in an embodiment of the present invention, where the evaluation model in fig. 3 is only an example, and in other embodiments, the number of the first evaluation modules may be one or more, and is not limited to the number of the first evaluation modules. In some embodiments, the first evaluation module has a single number, and performs evaluation processing on the sign information and the execution data corresponding to each frequency, and sends each obtained initial evaluation result to the second evaluation module, and the second evaluation module performs comprehensive evaluation processing on the initial evaluation results received cumulatively. In some embodiments, the first evaluation modules with the same number of frequencies as the information to be executed can be called, and the plurality of first evaluation modules perform synchronous evaluation processing on the sign information and the execution data corresponding to each frequency, and send each obtained initial evaluation result to the second evaluation module, so that the evaluation efficiency is improved.
According to the technical scheme provided by the embodiment, the information to be executed is evaluated through the preset evaluation model, the evaluation process is simplified, the evaluation efficiency is improved, and further the training efficiency of the information generation model is improved.
EXAMPLE III
Fig. 4 is a schematic structural diagram of an information generating apparatus according to an embodiment of the present invention, where the apparatus includes: an information to be executed acquisition module 310, a scheme evaluation module 320, and a model optimization module 330;
the to-be-executed information obtaining module 310 is configured to obtain multi-dimensional feature information of a target object, input the multi-dimensional feature information to a current information generation model, and obtain to-be-executed information of the target object, where the to-be-executed information includes an execution start time, an execution frequency, and execution items of each frequency;
the scheme evaluation module 320 is configured to obtain sign information and execution data of the target object during execution of the to-be-executed information, and perform evaluation processing on the sign information and the execution data to obtain an evaluation result of the to-be-executed information;
a model optimization module 330, configured to perform optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model, and update the current information generation model based on the new information generation model, where the updated current information generation model is used to generate the next information to be executed.
On the basis of the above embodiment, the multi-dimensional feature information includes feature information of a sign dimension, feature information of a treatment dimension, and feature information of a basic registration.
On the basis of the above embodiment, the scheme evaluation module 320 is configured to:
and evaluating the sign information and the execution data in the execution process of the execution items of each frequency based on an evaluation model to obtain an evaluation result of the information to be executed.
On the basis of the above embodiment, the evaluation model includes a first evaluation module and a second evaluation module, where the first evaluation module is configured to perform evaluation processing on the sign information and the execution data of the target object in the execution process of a single frequency to obtain an initial evaluation result, and the second evaluation module performs comprehensive evaluation processing on the initial evaluation results of each frequency to obtain an evaluation result of the information to be executed.
On the basis of the above embodiment, the model optimization module 330 is configured to:
and counting the information to be executed after the evaluation is completed within the preset time interval according to the preset time interval, and performing optimization training on the current information generation model based on the information to be executed after the evaluation is completed and the corresponding evaluation result to obtain a new information generation model.
On the basis of the above embodiment, the model optimization module 330 is configured to:
matching the evaluation result with a threshold condition;
and if the evaluation result does not meet the threshold condition, performing optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model.
On the basis of the above embodiment, the model optimization module 330 includes:
the standard information determining unit is used for matching corresponding standard information in a standard information base based on the multi-dimensional characteristic information of the target object;
and the model optimization unit is used for forming a loss function based on the information to be executed, the standard information, the evaluation result and the acquired sample data, carrying out parameter adjustment on the current information generation model based on the loss function, and generating a new information generation model.
On the basis of the above embodiment, the model optimization unit is configured to:
acquiring positive sample data and negative sample data, respectively traversing the positive sample data and the negative sample data to obtain a first sample distance and a second sample distance, and determining a first loss item based on the first sample distance and the second sample distance and corresponding multi-dimensional characteristic information;
determining a second loss term based on the evaluation result;
and performing weighted calculation based on the first loss term and the second loss term to form the loss function, wherein the weights of the first loss term and the second loss term are determined based on the information to be executed and the standard information.
The information generation device provided by the embodiment of the invention can execute the information generation method provided by any embodiment of the invention, and has corresponding functional modules and beneficial effects of the execution method.
Example four
Fig. 5 is a schematic structural diagram of an electronic device according to a fourth embodiment of the present invention. FIG. 5 illustrates a block diagram of an electronic device 12 suitable for use in implementing embodiments of the present invention. The electronic device 12 shown in fig. 5 is only an example and should not bring any limitation to the function and the scope of use of the embodiment of the present invention. The device 12 is typically an electronic device that undertakes image classification functions.
As shown in FIG. 5, electronic device 12 is embodied in the form of a general purpose computing device. The components of electronic device 12 may include, but are not limited to: one or more processors 16, a memory device 28, and a bus 18 that connects the various system components (including the memory device 28 and the processors 16).
A program 36 having a set (at least one) of program modules 26 may be stored, for example, in storage 28, such program modules 26 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which examples or some combination thereof may include an implementation of a gateway environment. Program modules 26 generally perform the functions and/or methodologies of the described embodiments of the invention.
The processor 16 executes various functional applications and data processing by executing programs stored in the storage device 28, for example, implementing the information generation method provided by the above-described embodiment of the present invention.
EXAMPLE five
Fifth embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the information generating method provided in the fifth embodiment of the present invention.
Of course, the computer program stored on the computer-readable storage medium provided by the embodiment of the present invention is not limited to the method operations described above, and may also execute the information generating method provided by any embodiment of the present invention, where the method includes:
obtaining multi-dimensional characteristic information of a target object, inputting the multi-dimensional characteristic information into a current information generation model, and obtaining to-be-executed information of the target object, wherein the to-be-executed information comprises execution starting time, execution frequency and execution items of each frequency;
acquiring sign information and execution data of the target object in the execution process of the information to be executed, and performing evaluation processing on the sign information and the execution data to obtain an evaluation result of the information to be executed;
and performing optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model, and updating the current information generation model based on the new information generation model, wherein the updated current information generation model is used for generating the next information to be executed.
Computer storage media for embodiments of the invention may employ any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
A computer readable signal medium may include a propagated data signal with computer readable source code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.
Source code embodied on a computer-readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
Computer source code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + +, or the like, as well as conventional procedural programming languages, such as the "C" programming language or similar programming languages. The source code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of gateway, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (11)
1. An information generating method, comprising:
obtaining multi-dimensional characteristic information of a target object, inputting the multi-dimensional characteristic information into a current information generation model, and obtaining to-be-executed information of the target object, wherein the to-be-executed information comprises execution starting time, execution frequency and execution items of each frequency;
acquiring sign information and execution data of the target object in the execution process of the information to be executed, and performing evaluation processing on the sign information and the execution data to obtain an evaluation result of the information to be executed;
and performing optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model, and updating the current information generation model based on the new information generation model, wherein the updated current information generation model is used for generating the next information to be executed.
2. The method of claim 1, wherein the multi-dimensional feature information comprises feature information of sign dimensions, feature information of treatment dimensions, and feature information of basic registration.
3. The method according to claim 1, wherein the performing the evaluation process on the sign information and the execution data to obtain the evaluation result of the information to be executed comprises:
and evaluating the sign information and the execution data in the execution process of the execution items of each frequency based on an evaluation model to obtain an evaluation result of the information to be executed.
4. The method according to claim 3, wherein the evaluation model includes a first evaluation module and a second evaluation module, wherein the first evaluation module is configured to perform evaluation processing on the sign information and the execution data of the target subject during execution of a single frequency to obtain an initial evaluation result, and the second evaluation module performs comprehensive evaluation processing on the initial evaluation results of each frequency to obtain an evaluation result of the information to be executed.
5. The method according to claim 1, wherein the performing optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model comprises:
and counting the information to be executed after the evaluation is completed within the preset time interval according to the preset time interval, and performing optimization training on the current information generation model based on the information to be executed after the evaluation is completed and the corresponding evaluation result to obtain a new information generation model.
6. The method according to claim 1, wherein the performing optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model comprises:
matching the evaluation result with a threshold condition;
and if the evaluation result does not meet the threshold condition, performing optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model.
7. The method according to claim 1, wherein the performing optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model comprises:
matching corresponding standard information in a standard information base based on the multi-dimensional characteristic information of the target object;
forming a loss function based on the information to be executed, the standard information, the evaluation result and the acquired sample data, performing parameter adjustment on a current information generation model based on the loss function, and generating a model by the new information.
8. The method of claim 7, wherein forming a loss function based on the information to be executed, the standard information, the evaluation result, and the acquired sample data comprises:
acquiring positive sample data and negative sample data, respectively traversing the positive sample data and the negative sample data to obtain a first sample distance and a second sample distance, and determining a first loss item based on the first sample distance and the second sample distance and corresponding multi-dimensional characteristic information;
determining a second loss term based on the evaluation result;
and performing weighted calculation based on the first loss term and the second loss term to form the loss function, wherein the weights of the first loss term and the second loss term are determined based on the information to be executed and the standard information.
9. An information generating apparatus, characterized by comprising:
the system comprises an information to be executed acquisition module, a data processing module and a data processing module, wherein the information to be executed acquisition module is used for acquiring multi-dimensional characteristic information of a target object, inputting the multi-dimensional characteristic information into a current information generation model and acquiring information to be executed of the target object, and the information to be executed comprises execution starting time, execution frequency and execution items of each frequency;
the scheme evaluation module is used for acquiring the sign information and the execution data of the target object in the execution process of the information to be executed, and evaluating the sign information and the execution data to obtain an evaluation result of the information to be executed;
and the model optimization module is used for carrying out optimization training on the current information generation model based on the evaluation result and the information to be executed to obtain a new information generation model, and updating the current information generation model based on the new information generation model, wherein the updated current information generation model is used for generating the next information to be executed.
10. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the information generating method according to any of claims 1-8 when executing the program.
11. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the information generating method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110817163.9A CN113488178B (en) | 2021-07-20 | 2021-07-20 | Information generation method and device, storage medium and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110817163.9A CN113488178B (en) | 2021-07-20 | 2021-07-20 | Information generation method and device, storage medium and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113488178A true CN113488178A (en) | 2021-10-08 |
| CN113488178B CN113488178B (en) | 2022-07-12 |
Family
ID=77941523
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110817163.9A Active CN113488178B (en) | 2021-07-20 | 2021-07-20 | Information generation method and device, storage medium and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113488178B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111428021A (en) * | 2020-06-05 | 2020-07-17 | 平安国际智慧城市科技股份有限公司 | Text processing method and device based on machine learning, computer equipment and medium |
| CN111639766A (en) * | 2020-05-26 | 2020-09-08 | 上海极链网络科技有限公司 | Sample data generation method and device |
| CN111724879A (en) * | 2020-06-29 | 2020-09-29 | 中金育能教育科技集团有限公司 | Rehabilitation training evaluation processing method, device and equipment |
| CN111816309A (en) * | 2020-07-13 | 2020-10-23 | 国家康复辅具研究中心 | Rehabilitation training prescription self-adaptive recommendation method and system based on deep reinforcement learning |
| CN112232506A (en) * | 2020-09-10 | 2021-01-15 | 北京迈格威科技有限公司 | Network model training method, image target recognition method, device and electronic equipment |
| CN112837803A (en) * | 2021-01-21 | 2021-05-25 | 镇江大医金缘健康管理有限公司 | Ultrasonic detection and remote pressing diagnosis system and method based on 5G signal transmission |
| JP2021086451A (en) * | 2019-11-28 | 2021-06-03 | 株式会社日立製作所 | Generation device, data analysis system, generation method, and generation program |
| CN112951449A (en) * | 2021-03-30 | 2021-06-11 | 江苏贝泰福医疗科技有限公司 | Cloud AI (artificial intelligence) regulation diagnosis and treatment system and method for neurological dysfunction diseases |
-
2021
- 2021-07-20 CN CN202110817163.9A patent/CN113488178B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2021086451A (en) * | 2019-11-28 | 2021-06-03 | 株式会社日立製作所 | Generation device, data analysis system, generation method, and generation program |
| CN111639766A (en) * | 2020-05-26 | 2020-09-08 | 上海极链网络科技有限公司 | Sample data generation method and device |
| CN111428021A (en) * | 2020-06-05 | 2020-07-17 | 平安国际智慧城市科技股份有限公司 | Text processing method and device based on machine learning, computer equipment and medium |
| CN111724879A (en) * | 2020-06-29 | 2020-09-29 | 中金育能教育科技集团有限公司 | Rehabilitation training evaluation processing method, device and equipment |
| CN111816309A (en) * | 2020-07-13 | 2020-10-23 | 国家康复辅具研究中心 | Rehabilitation training prescription self-adaptive recommendation method and system based on deep reinforcement learning |
| CN112232506A (en) * | 2020-09-10 | 2021-01-15 | 北京迈格威科技有限公司 | Network model training method, image target recognition method, device and electronic equipment |
| CN112837803A (en) * | 2021-01-21 | 2021-05-25 | 镇江大医金缘健康管理有限公司 | Ultrasonic detection and remote pressing diagnosis system and method based on 5G signal transmission |
| CN112951449A (en) * | 2021-03-30 | 2021-06-11 | 江苏贝泰福医疗科技有限公司 | Cloud AI (artificial intelligence) regulation diagnosis and treatment system and method for neurological dysfunction diseases |
Non-Patent Citations (1)
| Title |
|---|
| 张等: "基于深度信念网络的个人健康评估模型", 《软件导刊》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113488178B (en) | 2022-07-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11210292B2 (en) | Search method and apparatus | |
| CN108461151B (en) | Logic enhancement method and device of knowledge graph | |
| WO2020181805A1 (en) | Diabetes prediction method and apparatus, storage medium, and computer device | |
| US10395178B2 (en) | Risk assessment system and data processing method | |
| CA2465706A1 (en) | Patient data mining | |
| CN111564223B (en) | Infectious disease survival probability prediction method, and prediction model training method and device | |
| CN111382275A (en) | Construction method, device and medium of medical knowledge graph and electronic equipment | |
| CN116364299B (en) | Disease diagnosis and treatment path clustering method and system based on heterogeneous information network | |
| KR20160000522A (en) | System and method of emergency telepsychiatry using emergency psychiatric mental state prediction model | |
| CN110634566A (en) | Traditional Chinese medicine clinical diagnosis data processing system and method and information data processing terminal | |
| US20210192365A1 (en) | Computer device, system, readable storage medium and medical data analysis method | |
| US20230401466A1 (en) | Method for temporal knowledge graph reasoning based on distributed attention | |
| CN112420192A (en) | Disease typing method fusing multi-dimensional diagnosis and treatment information and related equipment | |
| CN113724815A (en) | Information pushing method and device based on decision grouping model | |
| CN110575178B (en) | Diagnosis and monitoring integrated medical system for judging motion state and judging method thereof | |
| CN109871866B (en) | Model training method, device, equipment and medium for hospital infection prediction | |
| CN113223734A (en) | Disease diagnosis and big health management platform based on algorithm, medical image and big data | |
| CN111081325B (en) | Medical data processing method and device | |
| CN116719926B (en) | Congenital heart disease report data screening method and system based on intelligent medical treatment | |
| CN113488178B (en) | Information generation method and device, storage medium and electronic equipment | |
| US20220318626A1 (en) | Meta-training framework on dual-channel combiner network system for dialysis event prediction | |
| CN116383413A (en) | Knowledge graph updating method and system based on medical data extraction | |
| CN111640517B (en) | Medical record coding method and device, storage medium and electronic equipment | |
| CN114580915A (en) | Intelligent evaluation method and system for hair-planting effect of novel microneedle technology | |
| US20220391760A1 (en) | Combining model outputs into a combined model output |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |