CN110444263B - Disease data processing method, device, equipment and medium based on federal learning - Google Patents
Disease data processing method, device, equipment and medium based on federal learning Download PDFInfo
- Publication number
- CN110444263B CN110444263B CN201910774155.3A CN201910774155A CN110444263B CN 110444263 B CN110444263 B CN 110444263B CN 201910774155 A CN201910774155 A CN 201910774155A CN 110444263 B CN110444263 B CN 110444263B
- Authority
- CN
- China
- Prior art keywords
- disease
- model
- patient
- federal learning
- trained
- 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.)
- Active
Links
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 title claims abstract description 195
- 201000010099 disease Diseases 0.000 title claims abstract description 194
- 238000003672 processing method Methods 0.000 title claims abstract description 27
- 238000012549 training Methods 0.000 claims abstract description 96
- 230000036541 health Effects 0.000 claims abstract description 73
- 239000013598 vector Substances 0.000 claims abstract description 45
- 238000003745 diagnosis Methods 0.000 claims abstract description 31
- 238000000034 method Methods 0.000 claims abstract description 26
- 230000008569 process Effects 0.000 claims abstract description 15
- 238000012545 processing Methods 0.000 claims description 38
- 238000013527 convolutional neural network Methods 0.000 claims description 17
- 238000000605 extraction Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 3
- 206010012601 diabetes mellitus Diseases 0.000 description 11
- 238000004891 communication Methods 0.000 description 6
- 230000002776 aggregation Effects 0.000 description 4
- 238000004220 aggregation Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000036772 blood pressure Effects 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000035487 diastolic blood pressure Effects 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000013526 transfer learning Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
-
- 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
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- Theoretical Computer Science (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Pathology (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a disease data processing method, device, equipment and medium based on federal learning, wherein the method comprises the following steps: acquiring an electronic health record of a patient diagnosed in a local database and disease data; extracting the characteristics of the electronic health record to obtain the disease characteristic vector of each patient; constructing a local training sample set according to the disease feature vector and the disease data of each patient; and participating in federal learning of data ends of all hospitals based on the local training sample set to obtain a disease prediction model. According to the invention, the data of all hospital ends are combined for federal training, and a high-quality disease prediction model can be trained on the basis of not revealing the privacy of patients of the hospital ends, so that a positive auxiliary effect is exerted in the diagnosis process of doctors.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a disease data processing method, device, equipment and medium based on federal learning.
Background
With the development of the fields of machine learning and artificial intelligence, the application of machine learning in the field of medical health is one current exploration direction. However, machine learning requires a large amount of data as a learning basis, and patient data of each hospital is limited, and it is difficult to train a good-quality machine model by only relying on patient data of one hospital. If the data of each hospital are combined for training, the privacy of the patient of each hospital can be revealed. Therefore, how to realize the data combination of each hospital in a safe and reliable way to obtain a high-quality machine model is a problem to be solved urgently.
Disclosure of Invention
The invention mainly aims to provide a disease data processing method, device, equipment and medium based on federal learning, and aims to solve the technical problem of how to realize data combination of various hospitals in a safe and reliable mode so as to obtain a high-quality machine model.
In order to achieve the above object, the present invention provides a disease data processing method based on federal learning, the disease data processing method based on federal learning including:
acquiring an electronic health record of a patient diagnosed in a local database and disease data;
Extracting the characteristics of the electronic health record to obtain the disease characteristic vector of each patient;
constructing a local training sample set according to the disease feature vector and the disease data of each patient;
and participating in federal learning of data ends of all hospitals based on the local training sample set to obtain a disease prediction model.
Optionally, the electronic health record includes recorded values of multiple pieces of physiological information at multiple time points, and the step of extracting features of the electronic health record to obtain a disease feature vector of each patient includes:
for the electronic health record of each patient, carrying out preset statistical processing on the recorded values of the multiple pieces of physiological information at multiple time points to obtain multidimensional statistics;
and carrying out vectorization processing on the multidimensional statistics to obtain a patient characteristic vector of the patient.
Optionally, the step of performing a preset statistical process on the recorded values of the plurality of pieces of physiological information at a plurality of time points includes:
Grouping the recorded values of each item of physiological information at a plurality of time points according to a preset time interval to obtain a recorded value group of each item of physiological information;
And respectively counting the maximum value, the minimum value, the variance and the average value of each group of recorded value groups.
Optionally, the step of obtaining the disease prediction model based on the local training sample set participating in federal learning of each hospital data end includes:
obtaining initial model parameters of a model to be trained from a preset server;
Carrying out local training on the model to be trained according to the local training sample set and the initial model parameters to obtain model parameter updating;
uploading the model parameter update to the server, so that the server can aggregate the model parameter update uploaded by each hospital data end to obtain an aggregate model parameter, and returning the aggregate model parameter to each hospital data end to continue iterative training when the model to be trained is detected to be in an unconverged state, until the model to be trained is detected to be in a converged state, taking the aggregate model parameter as a final parameter of the model to be trained to obtain a disease prediction model, and issuing the disease prediction model to each hospital data end;
And receiving the disease prediction model issued by the server.
Optionally, after the step of participating in federal learning of each hospital data end based on the local training sample set to obtain the disease prediction model, the method further includes:
Acquiring an electronic health record and a case text of a target user;
Inputting the electronic health record of the target user into the disease prediction model to obtain the disease probability of the target user;
And inputting the case text, the illness probability and the electronic health record of the target user into a convolutional neural network which is trained in advance to obtain a recommended diagnosis scheme for a doctor to use as a diagnosis reference.
Optionally, the convolutional neural network comprises one convolutional layer and three fully-connected layers.
In addition, in order to achieve the above object, the present invention also provides a disease data processing apparatus based on federal learning, the disease data processing apparatus based on federal learning including:
the acquisition module is used for acquiring the electronic health record of the diagnosed patient and the data of the disease in the local database;
The extraction module is used for extracting the characteristics of the electronic health record to obtain the disease characteristic vector of each patient;
the construction module is used for constructing a local training sample set according to the disease characteristic vector of each patient and the disease data;
and the training module is used for participating in federal learning of the data end of each hospital based on the local training sample set to obtain a disease prediction model.
Optionally, the extracting module includes:
the statistics unit is used for carrying out preset statistics processing on the recorded values of the plurality of physiological information at a plurality of time points for the electronic health record of each patient;
and the vectorization processing unit is used for vectorizing the multidimensional statistics to obtain the patient characteristic vector of the patient.
Optionally, the statistics unit includes:
a grouping subunit, configured to group the recorded values of each item of physiological information at a plurality of time points according to a preset time interval, so as to obtain a group of recorded values of each item of physiological information;
And the statistics subunit is used for respectively counting the maximum value, the minimum value, the variance and the average value of each group of recorded value groups.
Optionally, the training module includes:
the acquisition unit is used for acquiring initial model parameters of the model to be trained from a preset server;
The local training unit is used for carrying out local training on the model to be trained according to the local training sample set and the initial model parameters to obtain model parameter updating;
The uploading unit is used for uploading the model parameter update to the server side so that the server side can aggregate the model parameter update uploaded by each hospital data side to obtain an aggregate model parameter, and returning the aggregate model parameter to each hospital data side to continue iterative training when the model to be trained is detected to be in an unconverged state until the model to be trained is detected to be in a converged state, taking the aggregate model parameter as a final parameter of the model to be trained to obtain a disease prediction model, and issuing the disease prediction model to each hospital data side;
And the receiving unit is used for receiving the disease prediction model issued by the server.
Optionally, the obtaining module is further configured to obtain an electronic health record and a case text of the target user after obtaining the disease prediction model;
The disease data processing device based on federal learning further comprises:
The input module is used for inputting the electronic health record of the target user into the illness prediction model to obtain the illness probability of the target user; and inputting the case text, the illness probability and the electronic health record of the target user into a convolutional neural network which is trained in advance to obtain a recommended diagnosis scheme for a doctor to use as a diagnosis reference.
In addition, to achieve the above object, the present invention also provides a disease data processing apparatus based on federal learning, which includes a memory, a processor, and a disease data processing program based on federal learning stored on the memory and executable on the processor, the disease data processing program based on federal learning implementing the steps of the disease data processing method based on federal learning as described above when executed by the processor.
In addition, in order to achieve the above object, the present invention also provides a computer-readable storage medium having stored thereon a disease data processing program based on federal learning, which when executed by a processor, implements the steps of the disease data processing method based on federal learning as described above.
In the invention, the electronic health record of the diagnosed patient and the data of the disease are obtained through each hospital end; extracting the characteristics of the electronic health record to obtain the disease characteristic vector of each patient; constructing a local training sample set according to the disease feature vector of each patient and the disease; and participating in federal learning of each hospital data end based on the local training sample set to obtain a disease prediction model. By combining the data of all hospital ends, the federal training is carried out, and a high-quality disease prediction model can be trained on the basis of not revealing the privacy of patients of the hospital ends, so that a positive auxiliary effect is exerted in the diagnosis process of doctors.
Drawings
FIG. 1 is a schematic diagram of a hardware operating environment according to an embodiment of the present invention;
FIG. 2 is a flow chart of a first embodiment of a federal learning-based disease data processing method according to the present invention;
FIG. 3 is a block diagram of a preferred embodiment of a federally learned disease data processing apparatus according to the present invention.
The achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
The invention provides disease data processing equipment based on federal learning, and referring to fig. 1, fig. 1 is a schematic structural diagram of a hardware operation environment related to an embodiment of the invention.
It should be noted that fig. 1 may be a schematic structural diagram of a hardware operating environment of a disease data processing apparatus based on federal learning. The disease data processing device based on federal learning in the embodiment of the invention can be a PC, or can be a terminal device with a display function, such as a smart phone, a smart television, a tablet personal computer, a portable computer and the like.
As shown in fig. 1, the federal learning-based disease data processing apparatus may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, a communication bus 1002. Wherein the communication bus 1002 is used to enable connected communication between these components. The user interface 1003 may include a Display, an input unit such as a Keyboard (Keyboard), and the optional user interface 1003 may further include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface). The memory 1005 may be a high-speed RAM memory or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also optionally be a storage device separate from the processor 1001 described above.
Optionally, the federal learning-based disease data processing apparatus may further include a camera, RF (Radio Frequency) circuitry, sensors, audio circuitry, wiFi modules, and the like. It will be appreciated by those skilled in the art that the federal learning-based disease data processing apparatus structure illustrated in fig. 1 is not limiting of the federal learning-based disease data processing apparatus, and may include more or fewer components than illustrated, or certain components may be combined, or a different arrangement of components.
As shown in fig. 1, an operating system, a network communication module, a user interface module, and a federal learning-based disease data processing program may be included in a memory 1005, which is a computer storage medium.
In the disease data processing device based on federal learning shown in fig. 1, the network interface 1004 is mainly used for connecting to a background server, and performing data communication with the background server; the user interface 1003 is mainly used for connecting a client (user side) and performing data communication with the client; and the processor 1001 may be configured to invoke the federal learning-based disease data processing program stored in the memory 1005 and perform the following operations:
acquiring an electronic health record of a patient diagnosed in a local database and disease data;
Extracting the characteristics of the electronic health record to obtain the disease characteristic vector of each patient;
constructing a local training sample set according to the disease feature vector and the disease data of each patient;
and participating in federal learning of data ends of all hospitals based on the local training sample set to obtain a disease prediction model.
Further, the electronic health record includes recorded values of multiple pieces of physiological information at multiple time points, and the step of extracting features of the electronic health record to obtain a disease feature vector of each patient includes:
for the electronic health record of each patient, carrying out preset statistical processing on the recorded values of the multiple pieces of physiological information at multiple time points to obtain multidimensional statistics;
and carrying out vectorization processing on the multidimensional statistics to obtain a patient characteristic vector of the patient.
Further, the step of performing a preset statistical process on the recorded values of the plurality of pieces of physiological information at a plurality of time points includes:
Grouping the recorded values of each item of physiological information at a plurality of time points according to a preset time interval to obtain a recorded value group of each item of physiological information;
And respectively counting the maximum value, the minimum value, the variance and the average value of each group of recorded value groups.
Further, the step of obtaining the disease prediction model based on the local training sample set participating in federal learning of each hospital data end includes:
obtaining initial model parameters of a model to be trained from a preset server;
Carrying out local training on the model to be trained according to the local training sample set and the initial model parameters to obtain model parameter updating;
uploading the model parameter update to the server, so that the server can aggregate the model parameter update uploaded by each hospital data end to obtain an aggregate model parameter, and returning the aggregate model parameter to each hospital data end to continue iterative training when the model to be trained is detected to be in an unconverged state, until the model to be trained is detected to be in a converged state, taking the aggregate model parameter as a final parameter of the model to be trained to obtain a disease prediction model, and issuing the disease prediction model to each hospital data end;
And receiving the disease prediction model issued by the server.
Further, after the step of participating in federal learning at each hospital data end based on the local training sample set to obtain a disease prediction model and predicting the disease probability of the target user in combination with the electronic health record of the target user, the processor 1001 may be configured to invoke a disease data processing program based on federal learning stored in the memory 1005, and further perform the following operations:
Acquiring an electronic health record and a case text of a target user;
Inputting the electronic health record of the target user into the disease prediction model to obtain the disease probability of the target user;
And inputting the case text, the illness probability and the electronic health record of the target user into a convolutional neural network which is trained in advance to obtain a recommended diagnosis scheme for a doctor to use as a diagnosis reference.
Further, the convolutional neural network comprises a convolutional layer and three full-connection layers.
Based on the hardware structure, various embodiments of the disease data processing method based on federal learning are provided.
Referring to fig. 2, a first embodiment of a federal learning-based disease data processing method of the present invention provides a federal learning-based disease data processing method, and it should be noted that although a logic sequence is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than that shown or described herein. The execution subject of each embodiment of the disease data processing method based on federal learning of the present invention may be a terminal device such as a PC, a smart phone, a smart television, a tablet computer, and a portable computer, and for convenience of description, in the following embodiments, a hospital end (i.e., a hospital data end) is used as the execution subject. The disease data processing method based on federal learning comprises the following steps:
step S10, acquiring an electronic health record of a patient diagnosed in a local database and disease data;
each hospital end stores patient data of each patient in a local database thereof, such as an electronic health record, wherein each piece of physiological information of the patient, such as heart rate, PH, blood pressure, diastolic blood pressure and the like, is recorded in the electronic health record, and the collection of the electronic health record can be obtained by measuring hospital equipment or can be reported by the patient. The hospital end can develop federal learning cooperation, combine patient data of the hospital end, perform machine model training, such as combine patient data of diabetes patients in the hospital end database, perform diabetes prediction model training, be used for predicting diabetes patients before diagnosing the patients by doctors, so that the doctors can refer to the patients when diagnosing the patients, if the prediction results in extremely high diabetes disease probability, the doctors can further diagnose the patients to determine whether the patients really suffer from diabetes, if the disease probability is extremely low, the doctors can basically exclude the possibility that the patients suffer from diabetes, and further diagnose the patients in other directions to determine the real focus of the patients.
The hospital side obtains the electronic health record of the diagnosed patient and the corresponding recorded disease data of the patient in the local database. The doctor can make diagnosis and determine the disease of the doctor, and the local database can store the electronic health record of the patient and the disease data of the disease. The disease data may include the name or label of the disease, or include the severity level of the disease.
Step S20, extracting the characteristics of the electronic health record to obtain the disease characteristic vector of each patient;
And the hospital end performs feature extraction on the acquired electronic health records to obtain the patient feature vectors of all patients. The data of each item of physiological information in the electronic health record comprises a plurality of medical relations, a hospital end can process each item of physiological information according to the medical relations, extract medical features of each item of physiological information, and vector the medical features to obtain patient feature vectors of each patient. Specifically, for an electronic health record of a patient, medical features are extracted from the electronic health record, and each medical feature is taken as a vector element to form a disease feature vector of the patient.
Step S30, constructing a local training sample set according to the disease feature vector of each patient and the disease data;
The hospital end constructs a local training sample set according to the disease characteristic vector of each patient and the disease data. Specifically, the local training sample set includes a plurality of training samples, each training sample includes an input value and an output value, the hospital end uses the disease feature vector of the patient as the input value, uses the disease data of the patient as the output value to form a training sample, and the training samples of all patients form the local training sample set.
And step S40, participating in federal learning of the data end of each hospital based on the local training sample set, and obtaining a disease prediction model.
After the hospital end obtains the local training sample set, the local training sample set is adopted to participate in federal learning. Specifically, in each iterative training of federal learning, the hospital end adopts a local training sample set to perform local training on a model to be trained, wherein the model to be trained can be a neural network model or a deep learning model. And taking a final model obtained by federal learning as a disease prediction model.
It should be noted that, each hospital end can select different federal learning mode modeling according to specific tasks and data conditions, and the modeling modes include horizontal, longitudinal, migration and mixed models. Specifically, the characteristic dimensions of the patient data at each hospital end may be different, and if the patient data at each hospital end has highly overlapped characteristic dimensions, but the patient overlap is smaller, a horizontal federal learning mode may be adopted. If the patient data of each hospital end are overlapped with the patient height, but the feature dimension overlap is small, a longitudinal federal learning mode can be adopted. If neither the sample nor the feature dimensions of the dataset overlap sufficiently, a federal transfer learning mode may be employed.
Further, step S40 includes:
step S401, obtaining initial model parameters of a model to be trained from a preset server;
The server side transmits initial model parameters of the model to be trained to each hospital side participating in federal learning, wherein the model parameters can be weight parameters connected between nodes of the neural network. The hospital side obtains initial model parameters of the model to be trained from the server side.
Step S402, carrying out local training on the model to be trained according to the local training sample set and the initial model parameters to obtain model parameter updating;
After acquiring the initial model parameters, the hospital end carries out local training on the model to be trained according to the initial model parameters and the local training sample set to obtain model parameter updating. Specifically, the local training process may be that the hospital end substitutes the initial model parameters into the model to be trained, takes a diseased feature vector in a local training sample as input, takes diseased data as output, calculates a gradient value by using a local training sample set, updates the initial model parameters according to the gradient value, and obtains updated model parameters, thereby obtaining updated model parameters.
Step S403, uploading the model parameter update to the server side, so that the server side can aggregate the model parameter update uploaded by each hospital data side to obtain an aggregate model parameter, and returning the aggregate model parameter to each hospital data side to continue iterative training when detecting that the model to be trained is in an unconverged state, until detecting that the model to be trained is in a converged state, taking the aggregate model parameter as a final parameter of the model to be trained to obtain a disease prediction model, and issuing the disease prediction model to each hospital data side;
And the hospital side updates and uploads the model parameters obtained by the local training to the server side. The server receives the model parameter updates uploaded by the hospital terminals, and aggregates the model parameter updates to obtain aggregated model parameters, wherein the aggregation can be carried out by weighted average on the model parameter updates; after obtaining the aggregate model parameters, the server detects whether the model to be trained is in a convergence state, and if the model to be trained is detected to be in an unconverged state, the aggregate model parameters are issued to the hospital terminals, and the hospital terminals continue to perform local training according to the aggregate model parameters and the local training sample set; the method comprises the steps of circulating until a server detects that a model to be trained is in a convergence state, ending training, and taking the latest obtained aggregate model parameters as final parameters of the model to be trained, so that a final disease prediction model is obtained; the server side issues the obtained disease prediction model to each hospital side, specifically, may issue the final parameters to each hospital side, or issue the final disease prediction model file to each hospital side.
The server side detects whether the model to be trained is in a convergence state or not, wherein the method can be used for calculating a difference value between the latest aggregation model parameter and the last aggregation model parameter, if the difference value is smaller than a preset value, the model to be trained is determined to be in the convergence state, and if the difference value is not smaller than the preset value, the model to be trained is determined to be in an unconverged state; or judging whether the number of iterative training reaches the preset number, if so, determining that the model to be trained is in a convergence state; or judging whether the training time length is longer than the preset time length, if so, determining that the model to be trained is in a convergence state. The preset value, the preset times and the preset time length can be set according to the needs.
Furthermore, the hospital end can encrypt the model parameter update obtained by local training according to a preset encryption algorithm, upload the model parameter update after encryption to the server end, and the server end carries out aggregation processing on the model parameter update after encryption. The preset encryption algorithm may be a homomorphic encryption algorithm (Homomorphic Encryption).
Step S404, receiving the disease prediction model issued by the server.
The hospital side receives the disease prediction model issued by the server, namely, each hospital side obtains the disease prediction model obtained through federal learning. The hospital end can predict the disease probability of the patient according to the disease prediction model.
In this embodiment, the hospital end obtains the electronic health record of the diagnosed patient and the disease data; extracting the characteristics of the electronic health record to obtain the disease characteristic vector of each patient; constructing a local training sample set according to the disease feature vector of each patient and the disease; and participating in federal learning of each hospital data end based on the local training sample set to obtain a disease prediction model. By combining the data of all hospital ends, the federal training is carried out, and a high-quality disease prediction model can be trained on the basis of not revealing the privacy of patients of the hospital ends, so that a positive auxiliary effect is exerted in the diagnosis process of doctors.
Further, based on the first embodiment, a second embodiment of the disease data processing method based on federal learning according to the present invention provides a disease data processing method based on federal learning. In this embodiment, the electronic health record includes recorded values of a plurality of physiological information at a plurality of time points, and the step S20 includes:
step S201, for the electronic health record of each patient, carrying out preset statistical processing on the record values of the plurality of items of physiological information at a plurality of time points to obtain multidimensional statistics;
The electronic health record includes recorded values of multiple physiological information of the patient at multiple time points, i.e. each physiological information corresponds to multiple recorded values, and the recorded values are recorded at different time points, for example, the physiological information corresponding to blood pressure of the patient includes 24 blood pressure values recorded every other hour in 24 hours a day.
And the hospital end performs preset statistical treatment on the recorded values of the multiple pieces of physiological information in the electronic health records of each patient. The preset statistical processing may be a preset processing manner, for example, for a plurality of recorded values of each item of physiological information, an average value of the plurality of recorded values is calculated, and the statistic of the plurality of items of physiological information is used as the statistic of the item of physiological information, that is, the multidimensional statistic is formed.
Step S202, vectorizing the multidimensional statistics to obtain the patient characteristic vector of the patient.
The hospital end carries out vectorization processing on the multidimensional statistics to obtain the characteristic vector of the patient. The vectorization process is to take each dimension of most statistics as one element of the vector to form a disease characteristic vector with multiple characteristic dimensions.
Further, in order to better characterize the patient data, the step S201 includes:
Step S2011, grouping the recorded values of each item of physiological information at a plurality of time points according to a preset time interval to obtain a recorded value group of each item of physiological information;
The hospital end groups the recorded values of each item of physiological information at a plurality of time points according to a preset time interval to obtain a recorded value group of each item of physiological information. The preset time intervals may be preset as required, for example, 6 time intervals may be set, which are respectively the first 10%, the first 20%, the first 50%, the second 20% and the second 10% of 24 hours, and a record value group is established for each time interval, where the record value group includes record values of each record time point in the time interval. The recorded values of the physiological information recorded in 24 hours are subdivided according to time intervals, so that the change characteristics of the physiological information along with time can be embodied, and the trained model can be used for predicting the illness probability of a patient more accurately.
Step S2012, respectively counting the maximum value, the minimum value, the variance and the average value of each group of the record value groups.
The hospital side counts the maximum value, the minimum value, the variance and the square difference of each group of record value groups respectively. If n record values exist in a group of record value groups, the hospital end compares the n record values to obtain the largest record value and the smallest record value, calculates the variance and the square variance of the n record values, and then obtains four statistics. For example, m pieces of physiological information are included in an electronic health record of a patient, the hospital end divides each piece of physiological information into 6 groups according to a preset time interval to obtain 6m groups, and four statistics of maximum value, minimum value, variance and mean square error are calculated for each group to obtain 24 m-dimensional statistics.
In this embodiment, the multidimensional statistics term is obtained by carrying out preset statistical processing on multiple items of physiological information of the patient, and the multidimensional statistics is vectorized to obtain a disease feature vector of the patient, so that the disease feature vector obtained by extraction can embody disease features of the patient, and the disease prediction model obtained by training can predict the disease probability of the patient more accurately, thereby playing a more positive auxiliary role in the diagnosis process of doctors.
Further, based on the first or second embodiment, a third embodiment of the disease data processing method based on federal learning according to the present invention provides a disease data processing method based on federal learning. In this embodiment, after the step S40, the method further includes:
step S50, acquiring an electronic health record and a case text of a target user;
for a target user to be diagnosed by a doctor, the hospital end can acquire an electronic health record and a case text of the target user, wherein the case text can be handwriting or an electronic case text uploaded by the doctor, and a preliminary diagnosis result of the doctor on the target user can be recorded.
Step S60, inputting the electronic health record of the target user into the disease prediction model to obtain the disease probability of the target user;
The hospital end can extract the characteristics of the electronic health record of the target user to obtain the characteristic vector of the target user, input the characteristic vector into a disease prediction model, and output the disease probability of the target user by the disease prediction model. For example, when the hospital end trains the obtained disease prediction model according to the disease data of the diabetes patient, the disease prediction model can be used for predicting the probability of the diabetes of the patient so as to assist doctors in diagnosing the diabetes of the patient.
And step S70, inputting the case text, the illness probability and the electronic health record of the target user into a convolutional neural network which is obtained by training in advance to obtain a recommended diagnosis scheme for a doctor to use as a diagnosis reference.
In this embodiment, after obtaining the disease probability of the target user, if the disease probability of the target user exceeds a certain threshold, if it is predicted that the disease probability of diabetes of the target user is higher than a certain threshold, it is indicated that the target user is likely to have diabetes.
And the hospital end inputs the case text, the illness probability of the target user output by the illness prediction model and the electronic health record of the target user into a convolutional neural network obtained by pre-training to obtain a recommended diagnosis scheme. The convolutional neural network can output the number of the recommended diagnosis scheme, and the hospital end searches the corresponding recommended diagnosis scheme in the database according to the output number. One possible treatment method for the disease can be recorded in the recommended diagnosis scheme, for example, what medicine is taken, and a doctor can make a specific treatment scheme for a target user according to the recommended diagnosis scheme output by a hospital end.
The training process of the convolutional neural network may be: the convolutional neural network consists of a convolutional layer and three full-connection layers; the hospital end obtains a large number of case texts, illness probabilities and electronic health records of the diagnosed patients, performs feature extraction on the case texts and the electronic health records, vectorizes the extracted features and illness probabilities to be used as input vectors, uses a manually-marked recommended diagnosis scheme as output for calculating loss values, performs repeated iterative training, and obtains a trained convolutional neural network after knowing that the convergence of the convolutional neural network is detected.
In this embodiment, after the patient probability of the target user is obtained, the case text, the patient probability and the electronic health record of the target user are input into the convolutional neural network trained in advance, so as to obtain a recommended diagnosis scheme, so as to assist a doctor in making a specific treatment scheme for the target user, thereby helping the doctor to improve diagnosis efficiency.
In addition, an embodiment of the present invention further provides a disease data processing device based on federal learning, referring to fig. 3, where the disease data processing device based on federal learning includes:
an acquisition module 10, configured to acquire an electronic health record of a patient diagnosed and disease data of the patient in a local database;
the extracting module 20 is configured to perform feature extraction on the electronic health record to obtain a disease feature vector of each patient;
A construction module 30, configured to construct a local training sample set according to the disease feature vector and the disease data of each patient;
the training module 40 is configured to participate in federal learning of each hospital data end based on the local training sample set to obtain a disease prediction model, and combine with electronic health record prediction of a target user to obtain a disease probability of the target user, so as to provide a doctor as a diagnosis reference.
Further, the electronic health record includes recorded values of a plurality of physiological information at a plurality of time points, and the extracting module 20 includes:
the statistics unit is used for carrying out preset statistics processing on the recorded values of the plurality of physiological information at a plurality of time points for the electronic health record of each patient;
and the vectorization processing unit is used for vectorizing the multidimensional statistics to obtain the patient characteristic vector of the patient.
Further, the statistics unit includes:
a grouping subunit, configured to group the recorded values of each item of physiological information at a plurality of time points according to a preset time interval, so as to obtain a group of recorded values of each item of physiological information;
And the statistics subunit is used for respectively counting the maximum value, the minimum value, the variance and the average value of each group of recorded value groups.
Further, the training module 40 includes:
the acquisition unit is used for acquiring initial model parameters of the model to be trained from a preset server;
The local training unit is used for carrying out local training on the model to be trained according to the local training sample set and the initial model parameters to obtain model parameter updating;
The uploading unit is used for uploading the model parameter update to the server side so that the server side can aggregate the model parameter update uploaded by each hospital data side to obtain an aggregate model parameter, and returning the aggregate model parameter to each hospital data side to continue iterative training when the model to be trained is detected to be in an unconverged state until the model to be trained is detected to be in a converged state, taking the aggregate model parameter as a final parameter of the model to be trained to obtain a disease prediction model, and issuing the disease prediction model to each hospital data side;
And the receiving unit is used for receiving the disease prediction model issued by the server.
Further, the obtaining module 10 is further configured to obtain an electronic health record and a case text of the target user after obtaining the disease prediction model;
The disease data processing device based on federal learning further comprises:
The input module is used for inputting the electronic health record of the target user into the illness prediction model to obtain the illness probability of the target user; and inputting the case text, the illness probability and the electronic health record of the target user into a convolutional neural network which is trained in advance to obtain a recommended diagnosis scheme for a doctor to use as a diagnosis reference.
Further, the convolutional neural network comprises a convolutional layer and three full-connection layers.
The expansion content of the specific implementation mode of the disease data processing device based on federal learning is basically the same as that of each embodiment of the disease data processing method based on federal learning, and is not repeated here.
In addition, the embodiment of the invention also provides a computer readable storage medium, wherein the computer readable storage medium stores a disease data processing program based on federal learning, and the disease data processing program based on federal learning realizes the steps of the disease data processing method based on federal learning when being executed by a processor.
The expansion content of the specific implementation mode of the disease data processing device and the computer readable storage medium based on federal learning is basically the same as that of each embodiment of the disease data processing method based on federal learning, and is not described in detail herein.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.
The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. ROM/RAM, magnetic disk, optical disk) as described above, comprising instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the method according to the embodiments of the present invention.
The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the scope of the invention, but rather is intended to cover any equivalents of the structures or equivalent processes disclosed herein or in the alternative, which may be employed directly or indirectly in other related arts.
Claims (5)
1. A disease data processing method based on federal learning, the disease data processing method based on federal learning comprising:
Acquiring an electronic health record of a diagnosed patient and disease data of the diagnosed patient in a local database, wherein the electronic health record comprises recorded values of multiple items of physiological information at multiple time points;
carrying out preset statistical treatment on the recorded values of the multiple items of physiological information at multiple time points to obtain multi-dimensional statistics, and carrying out vectorization treatment on the multi-dimensional statistics to obtain a diseased feature vector;
constructing a local training sample set according to the disease feature vector of each patient and the disease data of each patient, wherein the disease feature vector of each patient is taken as an input value, the disease data of each patient is taken as an output value to form a training sample, and the training samples of all patients form the local training sample set;
based on the local training sample set, participating in federal learning of each hospital data end, and obtaining a disease prediction model;
Acquiring an electronic health record and a case text of a target user;
Inputting the electronic health record of the target user into the disease prediction model to obtain the disease probability of the target user;
Inputting the case text, the illness probability and the electronic health record of the target user into a convolutional neural network which is trained in advance to obtain a recommended diagnosis scheme for a doctor to use as a diagnosis reference;
The step of obtaining a disease prediction model based on the local training sample set participating in federal learning of each hospital data end comprises the following steps:
obtaining initial model parameters of a model to be trained from a preset server;
Carrying out local training on the model to be trained according to the local training sample set and the initial model parameters to obtain model parameter updating;
uploading the model parameter update to the server, so that the server can aggregate the model parameter update uploaded by each hospital data end to obtain an aggregate model parameter, and returning the aggregate model parameter to each hospital data end to continue iterative training when the model to be trained is detected to be in an unconverged state, until the model to be trained is detected to be in a converged state, taking the aggregate model parameter as a final parameter of the model to be trained to obtain a disease prediction model, and issuing the disease prediction model to each hospital data end;
And receiving the disease prediction model issued by the server.
2. The federally learned disease data processing method according to claim 1, wherein the step of performing a preset statistical process on the recorded values of the plurality of pieces of physiological information at a plurality of time points comprises:
Grouping the recorded values of each item of physiological information at a plurality of time points according to a preset time interval to obtain a recorded value group of each item of physiological information;
And respectively counting the maximum value, the minimum value, the variance and the average value of each group of recorded value groups.
3. A federal learning-based disease data processing apparatus, the federal learning-based disease data processing apparatus comprising:
The system comprises an acquisition module, a diagnosis module and a diagnosis module, wherein the acquisition module is used for acquiring an electronic health record of a diagnosed patient and disease data in a local database, and the electronic health record comprises record values of multiple items of physiological information at multiple time points;
The extraction module is used for carrying out preset statistical processing on the recorded values of the multiple items of physiological information at multiple time points to obtain multi-dimensional statistics, and carrying out vectorization processing on the multi-dimensional statistics to obtain a diseased feature vector;
The construction module is used for constructing a local training sample set according to the disease feature vector of each patient and the disease data of each patient, wherein the disease feature vector of each patient is taken as an input value, the disease data of each patient is taken as an output value to form a training sample, and the training samples of all patients form the local training sample set;
The training module is used for participating in federal learning of the data end of each hospital based on the local training sample set to obtain a disease prediction model;
the disease data processing device based on federal learning is also used for: acquiring an electronic health record and a case text of a target user;
Inputting the electronic health record of the target user into the disease prediction model to obtain the disease probability of the target user;
Inputting the case text, the illness probability and the electronic health record of the target user into a convolutional neural network which is trained in advance to obtain a recommended diagnosis scheme for a doctor to use as a diagnosis reference;
wherein, training module includes:
the acquisition unit is used for acquiring initial model parameters of the model to be trained from a preset server;
The local training unit is used for carrying out local training on the model to be trained according to the local training sample set and the initial model parameters to obtain model parameter updating;
The uploading unit is used for uploading the model parameter update to the server side so that the server side can aggregate the model parameter update uploaded by each hospital data side to obtain an aggregate model parameter, and returning the aggregate model parameter to each hospital data side to continue iterative training when the model to be trained is detected to be in an unconverged state until the model to be trained is detected to be in a converged state, taking the aggregate model parameter as a final parameter of the model to be trained to obtain a disease prediction model, and issuing the disease prediction model to each hospital data side;
And the receiving unit is used for receiving the disease prediction model issued by the server.
4. A federal learning-based disease data processing apparatus, comprising a memory, a processor, and a federal learning-based disease data processing program stored on the memory and executable on the processor, which federal learning-based disease data processing program when executed by the processor implements the steps of the federal learning-based disease data processing method of any one of claims 1 to 2.
5. A computer readable storage medium, characterized in that it has stored thereon a federal learning-based disease data processing program which, when executed by a processor, implements the steps of the federal learning-based disease data processing method according to any one of claims 1 to 2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910774155.3A CN110444263B (en) | 2019-08-21 | 2019-08-21 | Disease data processing method, device, equipment and medium based on federal learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910774155.3A CN110444263B (en) | 2019-08-21 | 2019-08-21 | Disease data processing method, device, equipment and medium based on federal learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110444263A CN110444263A (en) | 2019-11-12 |
| CN110444263B true CN110444263B (en) | 2024-04-26 |
Family
ID=68436898
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910774155.3A Active CN110444263B (en) | 2019-08-21 | 2019-08-21 | Disease data processing method, device, equipment and medium based on federal learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110444263B (en) |
Families Citing this family (34)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110890157A (en) * | 2019-11-18 | 2020-03-17 | 京东方科技集团股份有限公司 | Analytical model establishing method, auxiliary diagnosis device, auxiliary diagnosis equipment and medium |
| CN110942154B (en) * | 2019-11-22 | 2021-07-06 | 深圳前海微众银行股份有限公司 | Data processing method, device, equipment and storage medium based on federal learning |
| CN111310938A (en) * | 2020-02-10 | 2020-06-19 | 深圳前海微众银行股份有限公司 | Semi-supervision-based horizontal federal learning optimization method, equipment and storage medium |
| CN111400442B (en) * | 2020-02-28 | 2024-06-04 | 深圳前海微众银行股份有限公司 | Method, device, equipment and readable storage medium for analyzing resident address |
| CN111382706A (en) * | 2020-03-10 | 2020-07-07 | 深圳前海微众银行股份有限公司 | Prediction method and device based on federal learning, storage medium and remote sensing equipment |
| CN111428620A (en) * | 2020-03-20 | 2020-07-17 | 深圳前海微众银行股份有限公司 | Identity recognition method, device, equipment and medium based on federal in-vivo detection model |
| CN113470806B (en) * | 2020-03-31 | 2024-05-24 | 中移(成都)信息通信科技有限公司 | Method, device, equipment and computer storage medium for determining disease detection model |
| CN111462097A (en) * | 2020-04-03 | 2020-07-28 | 深圳前海微众银行股份有限公司 | Image processing method, device, equipment and storage medium based on federal learning |
| CN111477326B (en) * | 2020-04-13 | 2024-05-10 | 深圳前海微众银行股份有限公司 | Infectious virus carrier detection method, system, device, apparatus and medium |
| CN111477340B (en) * | 2020-04-13 | 2023-01-24 | 深圳前海微众银行股份有限公司 | Infectious disease contact condition detection method, device, equipment and readable storage medium |
| CN111753006B (en) * | 2020-06-10 | 2021-03-16 | 北京智源人工智能研究院 | Prediction system and method based on federal learning |
| CN111739600A (en) * | 2020-06-22 | 2020-10-02 | 平安医疗健康管理股份有限公司 | Information processing method and device, computer equipment and readable storage medium |
| CN111723942B (en) * | 2020-06-29 | 2024-02-02 | 南方电网科学研究院有限责任公司 | Enterprise electricity load prediction method, power grid business subsystem and prediction system |
| CN111724000B (en) * | 2020-06-29 | 2024-02-09 | 南方电网科学研究院有限责任公司 | User electricity charge recycling risk prediction method, device and system |
| CN111767411B (en) * | 2020-07-01 | 2024-09-24 | 深圳前海微众银行股份有限公司 | Knowledge graph representation learning optimization method, device and readable storage medium |
| CN112259238A (en) * | 2020-10-20 | 2021-01-22 | 平安科技（深圳）有限公司 | Electronic device, disease type detection method, apparatus, and medium |
| CN112289448A (en) * | 2020-11-06 | 2021-01-29 | 新智数字科技有限公司 | Health risk prediction method and device based on joint learning |
| CN112435755A (en) * | 2020-11-23 | 2021-03-02 | 平安科技（深圳）有限公司 | Disease analysis method, disease analysis device, electronic device, and storage medium |
| CN112309527A (en) * | 2020-11-24 | 2021-02-02 | 南通市第二人民医院 | Big data based medical record statistical management method and system |
| CN112447298A (en) * | 2020-11-24 | 2021-03-05 | 平安科技（深圳）有限公司 | Neurodegenerative disease modeling device based on federal transfer learning and related equipment |
| CN112435747A (en) * | 2020-11-26 | 2021-03-02 | 新智数字科技有限公司 | Disease risk prediction method, device, equipment and computer readable medium |
| CN112435748A (en) * | 2020-11-26 | 2021-03-02 | 新智数字科技有限公司 | Risk prediction method, device, equipment and computer readable medium |
| CN112447299A (en) * | 2020-12-01 | 2021-03-05 | 平安科技（深圳）有限公司 | Medical care resource prediction model training method, device, equipment and storage medium |
| CN112768056A (en) * | 2021-01-14 | 2021-05-07 | 新智数字科技有限公司 | Disease prediction model establishing method and device based on joint learning framework |
| CN113057587A (en) * | 2021-03-17 | 2021-07-02 | 上海电气集团股份有限公司 | Disease early warning method and device, electronic equipment and storage medium |
| CN113133768A (en) * | 2021-04-21 | 2021-07-20 | 东南大学 | Cardiovascular disease auxiliary diagnosis model and training method based on federal learning |
| CN113421638B (en) * | 2021-06-22 | 2022-07-15 | 平安科技（深圳）有限公司 | Model generation method and device based on transfer learning and computer equipment |
| CN113571203B (en) * | 2021-07-19 | 2024-01-26 | 复旦大学附属华山医院 | Multi-center federal learning-based brain tumor prognosis survival prediction method and system |
| CN113610303B (en) * | 2021-08-09 | 2024-03-19 | 北京邮电大学 | Load prediction method and system |
| CN113986890B (en) * | 2021-12-30 | 2022-03-11 | 四川华迪信息技术有限公司 | Joint hospital data migration method and system based on few-sample model learning |
| TWI796228B (en) | 2022-05-25 | 2023-03-11 | 臺中榮民總醫院 | Acute kidney injury predicting system and method thereof |
| CN116258720B (en) * | 2023-05-16 | 2023-08-18 | 联仁健康医疗大数据科技股份有限公司 | Image recognition method, device, electronic equipment and storage medium |
| CN116936048B (en) * | 2023-07-04 | 2024-03-19 | 吉林大学 | Federal learning hospital selection method, device and storage medium for heterogeneous medical information |
| CN117476217B (en) * | 2023-12-26 | 2024-03-26 | 山东大学齐鲁医院 | Chronic heart disease state of illness trend prediction system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010070489A1 (en) * | 2008-12-17 | 2010-06-24 | Koninklijke Philips Electronics, N.V. | Intelligent query routing for federated pacs |
| CN101836893A (en) * | 2009-03-19 | 2010-09-22 | 朱晓阳 | Remote network medical diagnosis service system |
| CN103237041A (en) * | 2012-11-07 | 2013-08-07 | 无锡成电科大科技发展有限公司 | Wireless medical data transmission method and wireless medical data transmission system |
| CN106778014A (en) * | 2016-12-29 | 2017-05-31 | 浙江大学 | A kind of risk Forecasting Methodology based on Recognition with Recurrent Neural Network |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090112794A1 (en) * | 2007-10-31 | 2009-04-30 | Richard Dean Dettinger | Aliased keys for federated database queries |
-
2019
- 2019-08-21 CN CN201910774155.3A patent/CN110444263B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010070489A1 (en) * | 2008-12-17 | 2010-06-24 | Koninklijke Philips Electronics, N.V. | Intelligent query routing for federated pacs |
| CN101836893A (en) * | 2009-03-19 | 2010-09-22 | 朱晓阳 | Remote network medical diagnosis service system |
| CN103237041A (en) * | 2012-11-07 | 2013-08-07 | 无锡成电科大科技发展有限公司 | Wireless medical data transmission method and wireless medical data transmission system |
| CN106778014A (en) * | 2016-12-29 | 2017-05-31 | 浙江大学 | A kind of risk Forecasting Methodology based on Recognition with Recurrent Neural Network |
Non-Patent Citations (1)
| Title |
|---|
| 联邦学习;面包君;《http://zhuanlan.zhihu.com/p/67513614》;1-8 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110444263A (en) | 2019-11-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110444263B (en) | Disease data processing method, device, equipment and medium based on federal learning | |
| CN111027487B (en) | Behavior recognition system, method, medium and equipment based on multi-convolution kernel residual error network | |
| US20240144105A1 (en) | Computer based object detection within a video or image | |
| EP3809328A1 (en) | Image annotation method and device | |
| CN110020623B (en) | Human body activity recognition system and method based on conditional variation self-encoder | |
| KR101846370B1 (en) | Method and program for computing bone age by deep neural network | |
| US9060714B2 (en) | System for detection of body motion | |
| KR101779800B1 (en) | System and method for evaluating multifaceted growth based on machine learning | |
| US20230394368A1 (en) | Collecting observations for machine learning | |
| CN103690240A (en) | Medical system | |
| CN111696661B (en) | Patient grouping model construction method, patient grouping method and related equipment | |
| CN110852367A (en) | Image classification method, computer device, and storage medium | |
| CN113628726B (en) | Traditional Chinese medicine diagnosis and treatment recommendation system and method based on graph neural network and electronic equipment | |
| CN116110597B (en) | Digital twinning-based intelligent analysis method and device for patient disease categories | |
| CN110909876A (en) | Sign information monitoring method and system based on multiple physiological parameters and CNN | |
| CN118471542B (en) | Medical health management system based on big data | |
| EP4449441A1 (en) | A computer implemented method and a system | |
| Sharma et al. | Data pre-processing using neural processes for modeling personalized vital-sign time-series data | |
| CN110276283B (en) | Picture identification method, target identification model training method and device | |
| WO2020083831A9 (en) | Computer based object detection within a video or image | |
| EP3721445B1 (en) | System to improve clinical workflow | |
| Zhou et al. | An unsupervised deep clustering for Bone X-ray classification and anomaly detection | |
| Magade et al. | Automating Decision Process of Overnight Patient Care Using Hybrig Machine Learning Algorithms | |
| Mukherjee et al. | Fast algorithm for computing weighted projection quantiles and data depth for high-dimensional large data clouds | |
| CN116110584B (en) | Human health risk assessment early warning system |
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 |