KR20220102944A - Method for constructing dataset - Google Patents

Abstract

The present invention relates to a method for constructing a dataset for model learning. The method comprises the steps of: constructing a plurality of candidate unlabeled data groups from an unlabeled data pool; constructing an input dataset including one candidate unlabeled data group among the plurality of candidate unlabeled data groups and at least one piece of label data; calculating a representative eigenvalue corresponding to an input data set by inputting the constructed input data set to a first neural network model; and determining the unlabeled data group for label switching from among a plurality of candidate unlabeled data groups based on the calculated plurality of representative eigenvalues, wherein each of the plurality of representative eigenvalues may correspond to one input data set.

Description

{METHOD FOR CONSTRUCTING DATASET}

The present invention relates to a method of constructing a dataset, and more particularly, to a method of constructing a dataset for model learning.

In general, in order to detect or classify data through machine learning, it is necessary to secure learning data for machine learning in advance. For machine learning, at least 1,000 to 10,000 training data must be secured.

In order to learn a large amount of data, it is possible to select more efficient data for learning among unlabeled data and request it to an expert, and configure a dataset for learning with the labeled data according to the request. .

However, among a large amount of data, not only is the amount of unlabeled data much larger, but also a high cost may occur in securing sufficient training data because professional personnel are required for data labeling.

Accordingly, in training a model, there is a need in the art for constructing an effective training dataset.

Korean Patent Laid-Open No. 2016-0012537 discloses a method and apparatus for learning a neural network, and a data processing apparatus.

The present disclosure has been made in response to the above-mentioned background art, and an object of the present disclosure is to provide a method of constructing a dataset for model learning.

A method of constructing a dataset according to an embodiment of the present disclosure for realizing the above-described problem includes configuring a plurality of candidate unlabeled data groups from an unlabeled data pool, the plurality of candidate unlabeled data groups constructing an input dataset including one candidate unlabeled data group and one or more labeled data among the label data groups, inputting the configured input dataset into a first neural network model to the input dataset calculating a representative eigenvalue corresponding to , and determining an unlabeled data group for label switching from among a plurality of candidate unlabeled data groups based on the calculated plurality of representative eigenvalues. can Each of the plurality of representative eigenvalues may correspond to one input data set.

In an alternative embodiment of the data set configuration method, determining a plurality of unlabeled data included in the determined unlabeled data group for label switching as data to be labeled, and the plurality of unlabeled data determined as the data to be labeled. After the labeling is performed, the method may further include the step of configuring the unlabeled data and the one or more label data on which the labeling is performed into a training dataset of a neural network model for performing a classification task.

In an alternative embodiment of the method for constructing a dataset, the first neural network model uses an Empirical Neural Tangent Kernel (Empirical NTK), the minimum, maximum, or It may be a model for maximizing the average value.

In an alternative embodiment of the method for constructing a dataset, the configuring of the plurality of candidate unlabeled data groups from the unlabeled data pool includes: a plurality of candidate unlabeled data groups to be included in each of the plurality of candidate unlabeled data groups in the unlabeled data pool. of the plurality of candidates in a manner of randomly extracting unlabeled data of Unlabeled data groups can be configured.

In an alternative embodiment of the method for constructing a dataset, calculating the representative eigenvalue corresponding to the input dataset includes first eigenvalues corresponding to each of the one or more label data and the candidate unlabeled data group calculating second eigenvalues corresponding to each of the plurality of unlabeled data, and calculating a representative eigenvalue corresponding to the input data set based on the calculated first eigenvalues and second eigenvalues It may further include the step of

In an alternative embodiment of the method for constructing a dataset, the calculating of the representative eigenvalue corresponding to the input dataset includes an eigenvalue having a minimum value or a maximum value among the calculated first eigenvalues and second eigenvalues. may be determined as a representative eigenvalue corresponding to the input data set, or an average value of the calculated first and second eigenvalues may be determined as a representative eigenvalue corresponding to the input data set.

In an alternative embodiment of the method for constructing a dataset, the determining of an unlabeled data group for label switching from among the plurality of candidate unlabeled data groups includes: a plurality of eigenvalues calculated in response to the plurality of input datasets selecting an input dataset having the highest eigenvalue among the plurality of input datasets or selecting a plurality of input datasets having a higher-level eigenvalue from among a plurality of eigenvalues calculated in response to the plurality of input datasets; and The method may include determining a candidate unlabeled data group included in one or more selected input datasets as the unlabeled data group for label conversion.

A computer program stored in a computer-readable storage medium according to an embodiment of the present disclosure for realizing the above-described problems, when the computer program is executed in one or more processors, performs the following steps for configuring a data set wherein the steps include: constructing a plurality of candidate unlabeled data groups from an unlabeled data pool; an input including one candidate unlabeled data group of the plurality of candidate unlabeled data groups and one or more label data constructing a dataset, inputting the configured input dataset into a first neural network model to calculate a representative eigenvalue corresponding to the input dataset, and adding the calculated representative eigenvalues to the The method may include determining an unlabeled data group for label switching from among a plurality of candidate unlabeled data groups based on the plurality of unlabeled data groups. Each of the plurality of representative eigenvalues may correspond to one input data set.

A computing device for providing a data set configuration method according to an embodiment of the present disclosure for realizing the above-described problems includes a processor including one or more cores, and a memory, wherein the processor includes unlabeled data composing a plurality of candidate unlabeled data groups from a pool, composing an input dataset including one candidate unlabeled data group and one or more label data among the plurality of candidate unlabeled data groups; is input to the first neural network model to calculate a representative eigenvalue corresponding to the input dataset, and converts a label among a plurality of candidate unlabeled data groups based on the calculated representative eigenvalues It is possible to determine the unlabeled data group for use. Each of the plurality of representative eigenvalues may correspond to one input data set.

The present invention can provide a method of constructing a dataset for effective training of a model.

BRIEF DESCRIPTION OF THE DRAWINGS So that the above-mentioned features of the present disclosure may be understood in detail, with a more specific description, with reference to the following embodiments, some of the embodiments are shown in the accompanying drawings. Also, like reference numerals in the drawings are intended to refer to the same or similar functions throughout the various aspects. However, it should be noted that the appended drawings show only certain typical embodiments of the present disclosure and are not to be considered as limiting the scope of the present disclosure, and other embodiments having the same effect may be fully appreciated. Take note.
1 is a diagram illustrating a block diagram of a computing device performing an operation for providing a method for configuring a dataset, according to an embodiment of the present disclosure.
2 is a schematic diagram illustrating a network function according to an embodiment of the present disclosure;
3 is a flowchart illustrating a data set configuration process according to an embodiment of the present disclosure.
4A and 4B are tables showing performance results of a data set configuration method and another method according to an embodiment of the present disclosure.
5A and 5B are graphs showing performance results of a data set configuration method and another method according to an embodiment of the present disclosure.
6 is a graph showing a convergence rate prediction performance result of a data set configuration method according to an embodiment of the present disclosure.
7 is a simplified, general schematic diagram of an exemplary computing environment in which embodiments of the present disclosure may be implemented.

Various embodiments are now described with reference to the drawings. In this specification, various descriptions are presented to provide an understanding of the present disclosure. However, it is apparent that these embodiments may be practiced without these specific descriptions.

The terms “component,” “module,” “system,” and the like, as used herein, refer to a computer-related entity, hardware, firmware, software, a combination of software and hardware, or execution of software. For example, a component can be, but is not limited to being, a process running on a processor, a processor, an object, a thread of execution, a program, and/or a computer. For example, both an application running on a server and a server can be components. One or more components may reside within a processor and/or thread of execution. A component may be localized within one computer. A component may be distributed between two or more computers. In addition, these components can execute from various computer readable media having various data structures stored therein. Components may communicate via a network such as the Internet with another system, for example via a signal having one or more data packets (eg, data and/or signals from one component interacting with another component in a local system, distributed system, etc.) may communicate via local and/or remote processes depending on the data being transmitted).

In addition, the term “or” is intended to mean an inclusive “or” rather than an exclusive “or.” That is, unless otherwise specified or clear from context, "X employs A or B" is intended to mean one of the natural implicit substitutions. That is, X employs A; X employs B; or when X employs both A and B, "X employs A or B" may apply to either of these cases. It should also be understood that the term “and/or” as used herein refers to and includes all possible combinations of one or more of the listed related items.

Also, the terms "comprises" and/or "comprising" should be understood to mean that the feature and/or element in question is present. However, it should be understood that the terms "comprises" and/or "comprising" do not exclude the presence or addition of one or more other features, elements and/or groups thereof. Also, unless otherwise specified or unless it is clear from context to refer to a singular form, the singular in the specification and claims should generally be construed to mean “one or more”.

And, the term "at least one of A or B" should be interpreted to mean "when including only A", "when including only B", and "when combined with the configuration of A and B".

Those skilled in the art will further appreciate that the various illustrative logical blocks, configurations, modules, circuits, means, logics, and algorithm steps described in connection with the embodiments disclosed herein may be implemented in electronic hardware, computer software, or combinations of both. It should be recognized that they can be implemented with To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, configurations, means, logics, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application. However, such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.

Descriptions of the presented embodiments are provided to enable those skilled in the art to use or practice the present invention. Various modifications to these embodiments will be apparent to those skilled in the art of the present disclosure. The generic principles defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Thus, the present invention is not limited to the embodiments presented herein. The invention is to be construed in its widest scope consistent with the principles and novel features presented herein.

In an embodiment of the present disclosure, the computing device may include a server and/or user equipment. A computing device in this disclosure may include any type of device with computing capabilities. The computing device is a digital device, and may be a digital device having a computing capability including a memory and a processor, such as a laptop computer, a notebook computer, a desktop computer, a web pad, and a mobile phone. In an embodiment, when the computing device corresponds to a server, the computing device may be a web server that processes a service. The types of computing devices described above are merely examples, and the present disclosure is not limited thereto.

Data used throughout the detailed description and claims of the present invention may refer to data including values or information in any form, such as images or text.

The term "image" or "image data" as used throughout the present description and claims refers to multidimensional data composed of discrete image elements (eg, pixels in a two-dimensional image), in other words, ( For example, a term that refers to a visible object (eg, displayed on a video screen) or a digital representation of the object (eg, a file corresponding to a pixel output of a CT, MRI detector, etc.).

For example, “image” or “image” can be defined as computed tomography (CT), magnetic resonance imaging (MRI), fundus imaging, ultrasound or any other medical imaging known in the art. It may be a medical image of a subject collected by the system. The image does not necessarily have to be provided in a medical context, but may be provided in a non-medical context, for example, there may be X-ray imaging for security screening.

Throughout the detailed description and claims of the present invention, the 'DICOM (Digital Imaging and Communications in Medicine)' standard is a generic term for various standards used for digital image expression and communication in medical devices. The DICOM standard is published by a joint committee formed by the American Society of Radiological Medicine (ACR) and the National Electrical Engineers Association (NEMA).

In addition, throughout the detailed description and claims of the present invention, 'Picture Archiving and Communication System (PACS)' is a term that refers to a system that stores, processes, and transmits in accordance with the DICOM standard, X-ray, CT , medical imaging images acquired using digital medical imaging equipment such as MRI are stored in DICOM format and transmitted to terminals inside and outside the hospital through a network, to which reading results and medical records can be added.

Also, throughout this specification, a neural network, a neural network model, a model, a neural network, or a network function may be used interchangeably. A neural network may be composed of a set of interconnected computational units, which may be generally referred to as “nodes”. These “nodes” may also be referred to as “neurons”. A neural network is configured to include at least two or more nodes. Nodes (or neurons) constituting neural networks may be interconnected by one or more “links”.

1 is a diagram illustrating a block diagram of a computing device that performs an operation for providing a method for configuring a dataset, according to an embodiment of the present disclosure.

The configuration of the computing device 100 shown in FIG. 1 is only a simplified example. In an embodiment of the present disclosure, the computing device 100 may include other components for performing the computing environment of the computing device 100 , and only some of the disclosed components may configure the computing device 100 .

The computing device 100 may include a processor 110 , a memory 130 , and a network unit 150 .

In the present disclosure, the processor 110 relates to a method of constructing a dataset for learning a classification model, and by selecting sample data for label conversion for a query from an unlabeled data pool, eigenvalues without additional training of the classification model Efficient data for learning can be collected from unlabeled data through (eigenvalue) operation.

In general, in order to train a neural network model, the construction of a training dataset for the neural network model should be preceded. In order to learn the neural network model, a training dataset including at least 1,000 to 10,000 or more training data may be required. However, when training a neural network model for a specific field, such as a medical field, training data or label data for training is often insufficient. For example, in the case of a rare disease, since the probability of the occurrence of the rare disease itself is low, it may be difficult to collect learning data on the rare disease.

Furthermore, in teacher learning in the field of artificial intelligence, securing learning data and labeling tasks in which humans write down answers to learning data occupy a large proportion. In particular, in the medical deep learning field, the subject of labeling the diagnosis results on medical images is inevitably a medical professional. Considering these high labor costs of medical professionals, increasing the efficiency in the labeling process of learning data or the construction process of learning data plays a big role in increasing the realism of the business (ie, product price competitiveness).

Therefore, the dataset construction method of the present disclosure is a dataset construction method for learning a classification model, and it is possible to collect efficient data for learning from unlabeled data through eigenvalue operation without additional learning of the classification model, It is possible to efficiently collect and collect learning data necessary for not only common diseases but also rare diseases, and achieve efficiency in the labeling process of learning data or the construction process of learning data.

According to an embodiment of the present disclosure, the processor 110 configures a plurality of candidate unlabeled data groups from an unlabeled data pool, and unlabels one candidate among the plurality of candidate unlabeled data groups. Constructing an input dataset including a data group and one or more labeled data, inputting the configured input dataset to the first neural network model to calculate a representative eigenvalue corresponding to the input dataset, An unlabeled data group for label switching may be determined from among a plurality of candidate unlabeled data groups based on the calculated plurality of representative eigenvalues. Here, each of the plurality of representative eigenvalues may correspond to one input data set.

The unlabeled data pool of the present disclosure is a set obtained by classifying only unlabeled data among data required for model learning, and the label data pool may refer to a set obtained by classifying only label data among data required for learning. . The candidate unlabeled data group of the present disclosure may refer to a group including unlabeled data extracted from the unlabeled data pool. Here, the unlabeled data included in the candidate unlabeled data group may be unlabeled data to be labeled later.

The input dataset of the present disclosure is input data input to the first neural network model in order to determine an unlabeled data group for label conversion, and one input dataset includes a candidate unlabeled data group including unlabeled data; It may consist of one or more label data. The unlabeled data group for label switching according to the present disclosure may refer to a candidate unlabeled data group including unlabeled data to be labeled for model training among candidate unlabeled data groups. For example, the unlabeled data group for label switching may be transmitted to another user terminal or server performing labeling.

The representative eigenvalue of the present disclosure may refer to an eigenvalue representing an input dataset including one candidate unlabeled data group including a plurality of unlabeled data and one or more label data.

According to an embodiment of the present disclosure, the processor 110 determines a plurality of unlabeled data included in the determined unlabeled data group for label switching as data to be labeled, and adds the plurality of unlabeled data determined as data to be labeled. After the labeling is performed, the unlabeled data and one or more label data on which the labeling is performed may be configured as a training dataset of a second neural network model for performing a classification task.

According to an embodiment of the present disclosure, the first neural network model uses an Empirical Neural Tangent Kernel (Empirical NTK) to control a convergence rate of the second neural network model, the first neural network model It may be a model for maximizing a minimum value, a maximum value, or an average value of eigenvalues corresponding to input data of . The eigenvalues corresponding to the input data may be the eigenvalues of the label data constituting the matrix and the eigenvalues of the unlabeled data. The eigenvalues of the label data and the unlabeled data constituting the matrix will be described later.

In an embodiment, the first neural network model may mean the same model as the second neural network model. In another embodiment, in an actual application process, the first neural network model may be different from the second neural network model.

For example, the same neural network model may mean a model in which at least one of a structure of a neural network, a training method, an inference method, an operation method, a variable, or a weight is the same. As another example, different neural network models may refer to models in which at least one of a neural network structure, a training method, an inference method, an operation method, a variable, or a weight is different.

According to an embodiment of the present disclosure, when configuring a plurality of candidate unlabeled data groups from the unlabeled data pool, the processor 110 is configured to include a plurality of candidate unlabeled data groups in each of the plurality of candidate unlabeled data groups in the unlabeled data pool. It can be configured by randomly extracting the unlabeled data of The foregoing is merely an example, and the present disclosure is not limited thereto.

According to an embodiment of the present disclosure, when configuring a plurality of candidate unlabeled data groups from the unlabeled data pool, the processor 110 is configured to include a plurality of candidate unlabeled data groups in each of the plurality of candidate unlabeled data groups in the unlabeled data pool. It can be configured by extracting unlabeled data of , based on a preset number. Here, the preset number may be set based on the total number of unlabeled data included in the unlabeled data pool. For example, the preset number may be less than or equal to the total number of unlabeled data included in the unlabeled data pool. That is, when the total number of unlabeled data included in the unlabeled data pool is about 5000, if the preset number of extracted candidate unlabeled data is set to about 50, 50 candidate unlabeled data from the unlabeled data pool are It is possible to construct one candidate unlabeled data group by extracting it as a set. Accordingly, if the total number of unlabeled data included in the unlabeled data pool is about 5000 and the number of extraction settings of candidate unlabeled data is set to about 50, about 100 or more candidate unlabeled data groups may be configured. Accordingly, the processor 110 of the present disclosure may optimally determine the size of the candidate unlabeled data group in order to improve network performance by efficiently handling the memory problem while minimizing the complexity of the eigenvalue operation. The foregoing is merely an example, and the present disclosure is not limited thereto.

According to an embodiment of the present disclosure, when configuring the candidate unlabeled data group, the processor 110 configures one candidate unlabeled data group including the candidate unlabeled data extracted from the unlabeled data pool, and It is checked whether the total number of unlabeled data groups is the maximum number, and if the total number of candidate unlabeled data groups is the maximum number, the configuration of the candidate unlabeled data groups may be terminated. Here, when the processor 110 determines whether the total number of candidate unlabeled data groups is the maximum number, if the total number of candidate unlabeled data groups is not the maximum number, the processor 110 receives at least one candidate unlabeled data from the unlabeled data pool. A candidate unlabeled data group may be configured based on the extracted and extracted candidate unlabeled data. For example, if the total number of unlabeled data included in the unlabeled data pool is about 5000 and the number of extraction settings for candidate unlabeled data is set to about 50, 50 pieces of candidate unlabeled data extracted from the unlabeled data pool are included. Up to about 100 or more candidate unlabeled data groups can be configured. Therefore, when the maximum number of candidate unlabeled data groups to be configured is optimally preset in accordance with the total number of unlabeled data included in the unlabeled data pool, if the total number of candidate unlabeled data groups is the preset maximum number, the candidate If the configuration of the unlabeled data group is terminated and the total number of the candidate unlabeled data groups is not the preset maximum number, the process of configuring the candidate unlabeled data group may be repeated. The foregoing is merely an example, and the present disclosure is not limited thereto.

The maximum number of candidate unlabeled data groups may be set based on the total number of unlabeled data included in the unlabeled data pool and the number of candidate unlabeled data extracted from the unlabeled data pool. For example, if the number of candidate unlabeled data extracted from the unlabeled data pool is equal to the total number of unlabeled data included in the unlabeled data pool, the processor 110 may determine that the maximum number of candidate unlabeled data groups is one. It can be set as a dog. For example, if the total number of unlabeled data included in the unlabeled data pool is about 50 and the number of extraction settings for candidate unlabeled data is set to about 50, 50 pieces of candidate unlabeled data extracted from the unlabeled data pool are included. A maximum of one candidate unlabeled data group can be configured. The foregoing is merely an example, and the present disclosure is not limited thereto.

In addition, the processor 110, if the number of candidate unlabeled data extracted from the unlabeled data pool is less than the total number of unlabeled data included in the unlabeled data pool, the maximum number of candidate unlabeled data groups is plural. can be set. For example, if the total number of unlabeled data included in the unlabeled data pool is about 500 and the number of extraction settings of candidate unlabeled data is set to about 50, 50 pieces of candidate unlabeled data extracted from the unlabeled data pool are included. A maximum of 10 or more candidate unlabeled data groups can be configured. The foregoing is merely an example, and the present disclosure is not limited thereto.

According to an embodiment of the present disclosure, one candidate unlabeled data group includes at least one candidate unlabeled data extracted from an unlabeled data pool, and candidate unlabeled data included in one candidate unlabeled data group. , may be duplicated included in another candidate unlabeled data group. For example, unlabeled data included in the first configured first candidate unlabeled data group may be included in the second candidate unlabeled data group as well. The foregoing is merely an example, and the present disclosure is not limited thereto.

According to an embodiment of the present disclosure, the processor 110, when calculating the representative eigenvalue corresponding to the input data set, includes the first eigenvalues corresponding to each of the one or more label data and the candidate unlabeled data group. Second eigenvalues corresponding to each of the plurality of unlabeled data may be calculated, and a representative eigenvalue corresponding to the input data set may be calculated based on the calculated first eigenvalues and second eigenvalues. For example, when an input dataset consisting of one candidate unlabeled data group including 50 unlabeled data and 100 labeled data is input to the neural network model, one representative eigenvalue representing the corresponding input dataset is calculated. can be Accordingly, when there are 100 candidate unlabeled data groups, about 100 representative eigenvalues each representing about 100 input datasets may be calculated. Here, the reason for calculating the eigenvalue is to select data that can affect the learning curve convergence of the model in order to minimize the classification error of the classification model. That is, the eigenvalues representing one input dataset including one or more label data and one candidate unlabeled data group can be a factor that makes it possible to know that the convergence of the learning curve varies depending on which data is queried. have. Accordingly, the processor 110 of the present disclosure may effectively control the learning dynamics by querying data capable of effectively controlling the convergence rate. The foregoing is merely an example, and the present disclosure is not limited thereto.

According to an embodiment of the present disclosure, when the processor 110 calculates the representative eigenvalue corresponding to the input data set, the eigenvalue having the minimum value or the maximum value among the calculated first eigenvalues and the second eigenvalues. A value may be determined as a representative eigenvalue corresponding to the input data set, or an average value of the calculated first and second eigenvalues may be determined as a representative eigenvalue corresponding to the input data set. The foregoing is merely an example, and the present disclosure is not limited thereto.

According to an embodiment of the present disclosure, when the processor 110 determines an unlabeled data group for label switching from among a plurality of candidate unlabeled data groups, a plurality of eigenvalues calculated in response to a plurality of input data sets An input dataset having the highest eigenvalue may be selected, and a candidate unlabeled data group included in the selected input dataset may be determined as an unlabeled data group for label conversion. For example, if one candidate unlabeled data group including 50 unlabeled data and one input dataset consisting of 100 labeled data are input to the neural network model, there is one representative unique representative of the corresponding input dataset. A value can be calculated. Therefore, when there are 100 candidate unlabeled data groups, when about 100 representative eigenvalues representing each of about 100 input datasets are calculated, it is included in the input dataset having the highest eigenvalue among about 100 representative eigenvalues. The candidate unlabeled data group may be determined as the unlabeled data group for label switching. The foregoing is merely an example, and the present disclosure is not limited thereto.

In addition, according to an embodiment of the present disclosure, when determining an unlabeled data group for label switching from among a plurality of candidate unlabeled data groups, the processor 110 is configured to generate a plurality of unique values calculated in response to a plurality of input datasets. It is possible to select a plurality of input datasets having a higher-level eigenvalue from among the values, and determine a candidate unlabeled data group included in the plurality of selected input datasets as an unlabeled data group for label conversion. Here, when the processor 110 selects a plurality of input datasets having a high-level eigenvalue from among a plurality of eigenvalues calculated in response to the input dataset, the processor 110 interlocks with the target number of unlabeled data to be labeled. It is possible to select a plurality of input datasets with eigenvalues of the level. For example, if one candidate unlabeled data group including 50 unlabeled data and one input dataset consisting of 100 labeled data are input to the neural network model, there is one representative unique representative of the corresponding input dataset. A value can be calculated. Accordingly, when there are 100 candidate unlabeled data groups, about 100 representative eigenvalues each representing about 100 input datasets may be calculated. Here, when it is necessary to select the target number of 150 unlabeled data to be labeled, among about 100 representative eigenvalues, the highest first-level eigenvalue, the second-highest second-level eigenvalue, and the third-highest By selecting a candidate unlabeled data group included in each input dataset having three levels of eigenvalues, about 150 unlabeled data to be labeled can be selected. The foregoing is merely an example, and the present disclosure is not limited thereto.

In addition, according to an embodiment of the present disclosure, the processor 110 may select unlabeled data that maximizes the minimum value among the n eigenvalues for controlling the learning convergence rate of the model, and the maximum value among the n eigenvalues. It is also possible to select unlabeled data that maximizes , or unlabeled data that maximizes the average value of n eigenvalues. The foregoing is merely an example, and the present disclosure is not limited thereto.

Next, according to an embodiment of the present disclosure, the processor 110 determines a plurality of unlabeled data included in the determined unlabeled data group for label conversion as data to be labeled, and the plurality of unlabeled data determined as data to be labeled. After labeling is performed on data, unlabeled data on which labeling is performed and one or more label data may be configured as a training dataset of a second neural network model for performing a classification task. In one example, the second neural network model here may include the same or different model from the first neural network model.

And, according to an embodiment of the present disclosure, the processor 110 may update the unlabeled data pool by labeling the unlabeled data to be labeled. For example, the label data pool and the unlabeled data pool may be updated by labeling unlabeled data included in the candidate unlabeled data group, removing the label from the unlabeled data pool, and adding it to the label data pool. The processor 110 of the present disclosure may increase the size of the label data pool and decrease the size of the unlabeled data pool while performing the update process.

According to an embodiment of the present disclosure, the processor 110 may effectively control the learning dynamics of the model by querying data capable of controlling the convergence rate. The processor 110 of the present disclosure may more effectively control the convergence rate based on a convergence rate prediction requirement including a learning rate of a model or a learning rate schedule.

According to an embodiment of the present disclosure, the processor 110 may input a medical image set to a classification model to identify and classify medical images. The medical image may be an image obtained by photographing an organ or an organ including a pathology. For example, the medical image may be an image obtained by photographing an organ using MRI, CT, or the like. The medical image may be an image of an organ. Alternatively, the medical image may be an image of an organ including a specific pathology. The medical image set may be a set of medical images. The medical image set may include one or more medical images. Medical image sets may be generated in units of organs or pathologies. For example, a medical image set including a medical image of the lungs and a medical image set including a medical image of the pancreas may be generated. Alternatively, for example, lung cancer may also be classified into squamous cell carcinoma, adenocarcinoma, large cell carcinoma, etc. according to the pathology, and a medical image set may also be generated separately for each of squamous cell carcinoma, adenocarcinoma, and large cell carcinoma. have. The detailed description of the medical image set described above is only an example, and the present disclosure is not limited thereto.

According to an embodiment of the present disclosure, the processor 110, when learning the neural network model for the medical field, in a situation where training data for labeling is insufficient, without additional learning of the neural network model, Since it is possible to collect efficient data for learning from label data, it is possible to efficiently collect learning data necessary for not only common diseases but also rare diseases.

According to an embodiment of the present disclosure, by using the first neural network, a training dataset of the second neural network model for performing the classification task may be efficiently configured. Accordingly, efficiency in the labeling process of the learning data or the construction process of the learning data can be achieved.

According to an embodiment of the present disclosure, the processor 110 may effectively control a convergence rate by selecting sample data for label switching for a query from an unlabeled data pool through Equation 1 below.

는, Empirical NTK(Neutral Tangent Kernel) matrix K이고, X_L은, 레이블 데이터 풀(labeled data pool)의 전체 레이블 데이터이며, X_U는, 언레이블 데이터 풀(unlabeled data pool)의 전체 언레이블 데이터이고, X_u는, 언레이블 데이터 풀의 후보 언레이블 데이터 그룹이며, λ_i는, Empirical NTK matrix의 고유값이고, i는, i ∈ {1, ….., │X_L ∪ X_u│}이다.here,

is the Empirical Neutral Tangent Kernel (NTK) matrix K, X _L is the full label data of the labeled data pool, X _U is the full unlabeled data of the unlabeled data pool, and , X _u is a candidate unlabeled data group of the unlabeled data pool, λ _i is an eigenvalue of the Empirical NTK matrix, and i is i ∈ {1, ... .., │X _L ∪ X _u │}.

Equation 1 describes a formula for selecting unlabeled data that maximizes the minimum value among n eigenvalues for controlling the learning convergence rate of the model, but the processor 110 of the present disclosure provides the maximum value among the n eigenvalues. Unlabeled data that maximizes n may be selected, or unlabeled data that maximizes the average value of n eigenvalues may be selected. The foregoing is merely an example, and the present disclosure is not limited thereto.

As an example, the first neural network model in the present disclosure is operable based on Equation 1 above.

The processor 110 of the present disclosure collects efficient data for learning from unlabeled data through eigenvalue operation without additional training of the classification model by selecting sample data for label conversion for query from the unlabeled data pool. can

That is, the processor 110 of the present disclosure configures a plurality of candidate unlabeled data groups from the unlabeled data pool, and includes one candidate unlabeled data group and one or more label data among the plurality of candidate unlabeled data groups. constructs an input dataset, inputting the configured input dataset to the first neural network model, calculating representative eigenvalues corresponding to the input dataset, and unlabeling a plurality of candidates based on the calculated representative eigenvalues An unlabeled data group for label switching is determined from among the data groups, a plurality of unlabeled data included in the determined unlabeled data group for label switching is determined as data to be labeled, and a plurality of unlabeled data determined as data to be labeled After the labeling is performed, the unlabeled data and one or more label data on which the labeling has been performed may be configured as a training dataset of the second neural network model for performing the classification task.

A neural network model (eg, a first n neural network model and a second neural network model) of the present disclosure may be a deep neural network. Throughout this specification, the terms neural network, network function, and neural network may be used interchangeably. A deep neural network (DNN) may refer to a neural network including a plurality of hidden layers in addition to an input layer and an output layer. Deep neural networks can be used to identify the latent structures of data. In other words, it can identify the potential structure of photos, texts, videos, voices, and music (e.g., what objects are in the photos, what the text and emotions are, what the texts and emotions are, etc.) . Deep neural networks include convolutional neural networks (CNNs), recurrent neural networks (RNNs), restricted Boltzmann machines (RBMs), and deep belief networks (DBNs). , Q network, U network, Siamese network, and the like.

A convolutional neural network is a kind of deep neural network, and includes a neural network including a convolutional layer. A convolutional neural network is a type of multilayer perceptron designed to use minimal preprocessing. A CNN can consist of one or several convolutional layers and artificial neural network layers combined with them. CNNs can additionally utilize weights and pooling layers. Thanks to this structure, CNN can fully utilize the input data of the two-dimensional structure. Convolutional neural networks can be used to recognize objects in images. A convolutional neural network can process image data by representing it as a matrix with dimensions. For example, in the case of RGB (red-green-blue) encoded image data, each R, G, and B color may be represented as a two-dimensional (eg, two-dimensional image) matrix. That is, the color value of each pixel of the image data may be a component of the matrix, and the size of the matrix may be the same as the size of the image. Accordingly, the image data can be represented by three two-dimensional matrices (a three-dimensional data array).

In a convolutional neural network, a convolutional process (input/output of a convolutional layer) can be performed by moving the convolutional filter and multiplying the convolutional filter and matrix components at each position of the image. The convolutional filter may be composed of an n × n matrix. A convolutional filter can be composed of a fixed type filter that is generally smaller than the total number of pixels in the image. That is, when an m × m image is input as a convolutional layer (eg, a convolutional layer having a size of n × n of a convolutional filter), a matrix representing n × n pixels including each pixel of the image This convolutional filter can be multiplied by a component (ie, the product of each component of a matrix). A component matching the convolutional filter may be extracted from the image by multiplication with the convolutional filter. For example, a 3 × 3 convolutional filter for extracting upper and lower linear components from an image can be configured as [[0,1,0], [0,1,0], [0,1,0]] can When the 3 × 3 convolutional filter for extracting the upper and lower linear components from the image is applied to the input image, the upper and lower linear components matching the convolutional filter may be extracted and output from the image. The convolutional layer may apply a convolutional filter to each matrix (ie, R, G, B colors for R, G, B coded images) for each channel representing the image. The convolutional layer may extract a feature matching the convolutional filter from the input image by applying the convolutional filter to the input image. The filter value of the convolutional filter (ie, the value of each component of the matrix) may be updated by backpropagation in the learning process of the convolutional neural network.

A subsampling layer is connected to the output of the convolutional layer to simplify the output of the convolutional layer, thereby reducing memory usage and computational amount. For example, if you input the output of the convolutional layer to a pooling layer with a 2 × 2 max pooling filter, you can compress the image by outputting the maximum value included in each patch for every 2 × 2 patch in each pixel of the image. can The above-described pooling may be a method of outputting a minimum value from a patch or an average value of a patch, and any pooling method may be included in the present disclosure.

A convolutional neural network may include one or more convolutional layers and subsampling layers. The convolutional neural network may extract features from an image by repeatedly performing a convolutional process and a subsampling process (eg, the aforementioned max pooling, etc.). Through iterative convolutional and subsampling processes, neural networks can extract global features from images.

An output of the convolutional layer or the subsampling layer may be input to a fully connected layer. A fully connected layer is a layer in which all neurons in one layer and all neurons in neighboring layers are connected. The fully connected layer may refer to a structure in which all nodes of each layer are connected to all nodes of other layers in a neural network.

The processor 110 of the present disclosure may include one or more cores, and a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), and a tensor of the computing device 100 . It may include a processor for analyzing unstructured data such as a tensor processing unit (TPU) and deep learning (DL). The processor 110 may read a computer program stored in the memory 130 to perform model learning according to an embodiment of the present disclosure. According to an embodiment of the present disclosure, the processor 110 may select a set of samples for a query from the unlabeled data pool. The processor 110 may perform calculations for learning of the neural network, such as processing input data for learning in deep learning, feature extraction from input data, error calculation, and weight update of the neural network using backpropagation. have. The processor 110, at least one of a CPU, a GPGPU, and a TPU may process learning of a network function. For example, the CPU and GPGPU can process network function training and model training using network functions together. Also, in an embodiment of the present disclosure, learning of a network function and model learning using the network function may be processed by using the processors of a plurality of computing devices together. In addition, the computer program executed in the computing device according to an embodiment of the present disclosure may be a CPU, GPGPU or TPU executable program.

According to an embodiment of the present disclosure, the memory 130 may store any type of information generated or determined by the processor 110 and any type of information received by the network unit 150 .

According to an embodiment of the present disclosure, the memory 130 may include a flash memory type, a hard disk type, a multimedia card micro type, and a card type memory (eg, a memory card type). SD or XD memory, etc.), Random Access Memory (RAM), Static Random Access Memory (SRAM), Read-Only Memory (ROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Programmable Memory (PROM) read-only memory), a magnetic memory, a magnetic disk, and an optical disk may include at least one type of storage medium. The computing device 100 may operate in relation to a web storage that performs a storage function of the memory 130 on the Internet. The description of the above-described memory is only an example, and the present disclosure is not limited thereto.

According to an embodiment of the present disclosure, the network unit 150 may transmit and receive data for configuring a dataset and performing model learning with other computing devices, servers, and the like. The network unit 150 may transmit and receive data inferred to configure a dataset and perform model learning with other computing devices, servers, and the like. In addition, the network unit 150 may enable communication between a plurality of computing devices so that learning of a network function is performed in a distributed manner in each of the plurality of computing devices. The network unit 150 may enable communication between a plurality of computing devices to distribute analysis data generation using a network function.

According to an embodiment of the present disclosure, the network unit 150 may be configured regardless of its communication mode, such as wired and wireless, and may include a personal area network (PAN), a wide area network (WAN), and the like. ), etc., may be composed of various communication networks. In addition, the network unit 150 may be a well-known World Wide Web (WWW), and may use a wireless transmission technology used for short-range communication such as infrared (IrDA) or Bluetooth (Bluetooth). may be The techniques described herein may be used in the networks mentioned above, as well as in other networks.

As such, in the present disclosure, by selecting sample data for label conversion for query from the unlabeled data pool, it is possible to collect efficient data for learning from unlabeled data through eigenvalue operation without additional training of the classification model. .

That is, the present disclosure provides a process of configuring a plurality of candidate unlabeled data groups from an unlabeled data pool, and an input dataset including one candidate unlabeled data group and one or more label data among a plurality of candidate unlabeled data groups. a process of constructing , the process of inputting the configured input dataset into the first neural network model to calculate a representative eigenvalue corresponding to the input dataset, a plurality of candidate unlabeled data groups based on the plurality of calculated representative eigenvalues The process of determining an unlabeled data group for label switching among By performing a process of composing the unlabeled data and one or more label data on which the labeling has been performed into the training dataset of the second neural network model for performing the classification task, data efficient for learning from the unlabeled data is obtained. can be obtained

2 is a schematic diagram illustrating a network function according to an embodiment of the present disclosure;

Throughout this specification, computational model, neural network, network function, and neural network may be used interchangeably. A neural network may be composed of a set of interconnected computational units, which may generally be referred to as nodes. These nodes may also be referred to as neurons. A neural network is configured to include at least one or more nodes. Nodes (or neurons) constituting the neural networks may be interconnected by one or more links.

In the neural network, one or more nodes connected through a link may relatively form a relationship between an input node and an output node. The concepts of an input node and an output node are relative, and any node in an output node relationship with respect to one node may be in an input node relationship in a relationship with another node, and vice versa. As described above, an input node-to-output node relationship may be created around a link. One or more output nodes may be connected to one input node through a link, and vice versa.

In the relationship between the input node and the output node connected through one link, the value of the data of the output node may be determined based on data input to the input node. Here, a link interconnecting the input node and the output node may have a weight. The weight may be variable, and may be changed by the user or algorithm in order for the neural network to perform a desired function. For example, when one or more input nodes are interconnected to one output node by respective links, the output node sets values input to input nodes connected to the output node and links corresponding to the respective input nodes. An output node value may be determined based on the weight.

As described above, in a neural network, one or more nodes are interconnected through one or more links to form an input node and an output node relationship in the neural network. The characteristics of the neural network may be determined according to the number of nodes and links in the neural network, the correlation between the nodes and the links, and the value of a weight assigned to each of the links. For example, when the same number of nodes and links exist and there are two neural networks having different weight values of the links, the two neural networks may be recognized as different from each other.

A neural network may consist of a set of one or more nodes. A subset of nodes constituting a neural network may constitute a layer. Some of the nodes constituting the neural network may configure one layer based on distances from the initial input node. For example, a set of nodes having a distance n from the initial input node may constitute n layers. The distance from the initial input node may be defined by the minimum number of links that must be traversed to reach the corresponding node from the initial input node. However, the definition of such a layer is arbitrary for description, and the order of the layer in the neural network may be defined in a different way from the above. For example, a layer of nodes may be defined by a distance from the final output node.

The initial input node may mean one or more nodes to which data is directly input without going through a link in a relationship with other nodes among nodes in the neural network. Alternatively, in a relationship between nodes based on a link in a neural network, it may mean nodes that do not have other input nodes connected by a link. Similarly, the final output node may refer to one or more nodes that do not have an output node in relation to other nodes among nodes in the neural network. In addition, the hidden node may mean nodes constituting the neural network other than the first input node and the last output node.

The neural network according to an embodiment of the present disclosure may be a neural network in which the number of nodes in the input layer may be the same as the number of nodes in the output layer, and the number of nodes decreases and then increases again as the input layer progresses to the hidden layer. can In addition, the neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes in the input layer may be less than the number of nodes in the output layer, and the number of nodes decreases as the number of nodes progresses from the input layer to the hidden layer. have. In addition, the neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes in the input layer may be greater than the number of nodes in the output layer, and the number of nodes increases as the number of nodes progresses from the input layer to the hidden layer. can The neural network according to another embodiment of the present disclosure may be a neural network in a combined form of the aforementioned neural networks.

A deep neural network (DNN) may refer to a neural network including a plurality of hidden layers in addition to an input layer and an output layer. Deep neural networks can be used to identify the latent structures of data. In other words, it can identify the potential structure of photos, texts, videos, voices, and music (e.g., what objects are in the photos, what the text and emotions are, what the texts and emotions are, etc.) . Deep neural networks include convolutional neural networks (CNNs), recurrent neural networks (RNNs), auto encoders, generative adversarial networks (GANs), and restricted Boltzmann machines (RBMs). machine), a deep belief network (DBN), a Q network, a U network, a Siamese network, and a Generative Adversarial Network (GAN). The description of the deep neural network described above is only an example, and the present disclosure is not limited thereto.

In an embodiment of the present disclosure, the network function may include an autoencoder. The auto-encoder may be a kind of artificial neural network for outputting output data similar to input data. The auto encoder may include at least one hidden layer, and an odd number of hidden layers may be disposed between the input/output layers. The number of nodes in each layer may be reduced from the number of nodes in the input layer to an intermediate layer called the bottleneck layer (encoding), and then expanded symmetrically with reduction from the bottleneck layer to the output layer (symmetrical to the input layer). Autoencoders can perform non-linear dimensionality reduction. The number of input layers and output layers may correspond to a dimension after preprocessing the input data. In the auto-encoder structure, the number of nodes of the hidden layer included in the encoder may have a structure that decreases as the distance from the input layer increases. If the number of nodes in the bottleneck layer (the layer with the fewest nodes between the encoder and decoder) is too small, a sufficient amount of information may not be conveyed, so a certain number or more (e.g., more than half of the input layer, etc.) ) may be maintained.

The neural network may be trained by at least one of supervised learning, unsupervised learning, semi-supervised learning, and reinforcement learning. Learning of the neural network may be a process of applying knowledge for the neural network to perform a specific operation to the neural network.

A neural network can be trained in a way that minimizes output errors. In the training of a neural network, iteratively input the training data into the neural network, calculate the output of the neural network and the target error for the training data, and calculate the error of the neural network from the output layer of the neural network to the input layer in the direction to reduce the error. It is the process of updating the weight of each node of the neural network by backpropagation in the direction. In the case of supervised learning, training data in which correct answers are labeled in each training data is used (ie, labeled learning data), and in the case of unsupervised learning, the correct answers may not be labeled in each training data. That is, for example, the training data in the case of supervised learning regarding data classification may be data in which categories are labeled in each of the training data. Labeled training data is input to the neural network, and an error can be calculated by comparing the output (category) of the neural network with the label of the training data. As another example, in the case of unsupervised learning about data classification, an error may be calculated by comparing the input of training data with the output of the neural network. The calculated error is back propagated in the reverse direction (ie, from the output layer to the input layer) in the neural network, and the connection weight of each node of each layer of the neural network may be updated according to the back propagation. A change amount of the connection weight of each node to be updated may be determined according to a learning rate. The computation of the neural network on the input data and the backpropagation of errors can constitute a learning cycle (epoch). The learning rate may be applied differently according to the number of repetitions of the learning cycle of the neural network. For example, in the early stage of learning of a neural network, a high learning rate can be used to increase the efficiency by allowing the neural network to quickly obtain a certain level of performance, and in the late learning period, a low learning rate can be used to increase accuracy.

In training of a neural network, in general, the training data may be a subset of real data (that is, data to be processed using the trained neural network), and thus the error on the training data is reduced, but the error on the real data is reduced. There may be increasing learning cycles. Overfitting is a phenomenon in which errors on actual data increase by over-learning on training data as described above. For example, a phenomenon in which a neural network that has learned a cat by seeing a yellow cat does not recognize that it is a cat when it sees a cat other than yellow may be a type of overfitting. Overfitting can act as a cause of increasing errors in machine learning algorithms. In order to prevent such overfitting, various optimization methods can be used. In order to prevent overfitting, methods such as increasing the training data, regularization, and dropout that deactivate some of the nodes of the network in the process of learning, and the use of a batch normalization layer are applied. can

3 is a flowchart illustrating a data set configuration process according to an embodiment of the present disclosure.

As shown in FIG. 3 , the computing device of the present disclosure may configure a plurality of candidate unlabeled data groups from the unlabeled data pool ( S10 ). The computing device of the present disclosure, when configuring the plurality of candidate unlabeled data groups from the unlabeled data pool, randomly extracts a plurality of unlabeled data to be included in each of the plurality of candidate unlabeled data groups from the unlabeled data pool can be configured. When configuring the plurality of candidate unlabeled data groups from the unlabeled data pool, the computing device of the present disclosure includes a preset number of a plurality of unlabeled data to be included in each of the plurality of candidate unlabeled data groups in the unlabeled data pool. It can be extracted and constructed based on Here, the preset number may be set based on the total number of unlabeled data included in the unlabeled data pool. The computing device of the present disclosure may optimally determine the size of a candidate unlabeled data group in order to improve network performance by efficiently handling a memory problem while minimizing the complexity of eigenvalue calculation.

In addition, the computing device of the present disclosure configures an input dataset including one candidate unlabeled data group and one or more label data among a plurality of candidate unlabeled data groups, and uses the configured input dataset as a first neural network model. It is possible to calculate a representative eigenvalue corresponding to the input data set by inputting to ( S20 ). The computing device of the present disclosure calculates first eigenvalues corresponding to each of one or more label data and second eigenvalues corresponding to each of a plurality of unlabeled data included in a candidate unlabeled data group, and calculates the calculated first A representative eigenvalue corresponding to the input data set may be calculated based on the eigenvalues and the second eigenvalues. Here, the reason for calculating the eigenvalue is to select data that can affect the learning curve convergence of the model in order to minimize the classification error of the classification model. That is, a set of eigenvalues including the label dataset and one candidate unlabeled dataset may be a factor enabling it to be known that the convergence of the learning curve varies depending on which data is queried. Accordingly, the processor 110 of the present disclosure may effectively control the learning dynamics by querying data capable of effectively controlling the convergence rate. The computing device of the present disclosure, when calculating a representative eigenvalue corresponding to the input dataset, sets an eigenvalue having a minimum value or a maximum value among the calculated first eigenvalues and the second eigenvalues corresponding to the input dataset. It may be determined as a representative eigenvalue, or an average value of the calculated first and second eigenvalues may be determined as a representative eigenvalue corresponding to the input data set.

Next, the computing device of the present disclosure may determine an unlabeled data group for label switching from among a plurality of candidate unlabeled data groups based on the plurality of calculated representative eigenvalues ( S30 ). The computing device of the present disclosure selects an input dataset having the highest eigenvalue among a plurality of eigenvalues calculated in response to a plurality of input datasets, and labels a candidate unlabeled data group included in the selected input dataset It can be determined as an unlabeled data group for conversion. In some cases, the computing device of the present disclosure selects a plurality of input datasets having a higher-level eigenvalue from among a plurality of eigenvalues calculated corresponding to the plurality of input datasets, and adds the selected input datasets to the plurality of input datasets. The included candidate unlabeled data group may be determined as the unlabeled data group for label switching. Here, when the processor 110 selects a plurality of input datasets having a high-level eigenvalue from among a plurality of eigenvalues calculated in response to the input dataset, the processor 110 interlocks with the target number of unlabeled data to be labeled. It is possible to select a plurality of input datasets with eigenvalues of the level.

Next, the computing device of the present disclosure determines a plurality of unlabeled data included in the determined unlabeled data group for label conversion as data to be labeled, and performs labeling by labeling the plurality of unlabeled data determined as data to be labeled. The unlabeled data and one or more labeled data can be configured as a training dataset of a second neural network model for performing a classification task.

The computing device of the present disclosure may effectively control the learning dynamics of a model by querying data capable of controlling a convergence rate. The computing device of the present disclosure may more effectively control the convergence rate based on a convergence rate prediction requirement including a learning rate of a model or a learning rate schedule.

The computing device of the present disclosure may perform an eigenvalue calculation for convergence rate control using a neural network tangent kernel (NTK). In the neural network tangent kernel, when the number of channels of all layers goes to infinity, training dynamics can be expressed as in Equation 2 below.

의 고유값/고유벡터 쌍

과 상수

를 이용하여 위의 미분 방정식의 해를 표현하면 다음 수학식 3과 같다.Here, the NTK matrix

eigenvalue/eigenvector pairs of

and constant

The solution of the above differential equation is expressed using Equation 3 below.

는, 고유값이고, v_i는, 고유벡터이며, c_i는 상수일 수 있다.here,

is an eigenvalue, v _i is an eigenvector, and c _i may be a constant.

는, 해당 고유값에 상응하는 데이터 쿼리에 따라 학습곡선의 수렴성이 달라질 수 있음을 보여준다.eigenvalues

shows that the convergence of the learning curve can vary depending on the data query corresponding to the corresponding eigenvalue.

Accordingly, the computing device of the present disclosure may effectively control the learning dynamics by querying data capable of effectively controlling the convergence rate.

The computing device of the present disclosure may effectively control a convergence rate by selecting sample data for label switching for a query from an unlabeled data pool through Equation 4 below.

는, Empirical NTK matrix K이고, X_L은, 레이블 데이터 풀(labeled data pool)의 전체 레이블 데이터이며, X_U는, 언레이블 데이터 풀(unlabeled data pool)의 전체 언레이블 데이터이고, X_u는, 언레이블 데이터 풀의 후보 언레이블 데이터 그룹이며, λ_i는, Empirical NTK matrix의 고유값이고, i는, i ∈ {1, ……, │X_L ∪ X_u│}이다.here,

is the Empirical NTK matrix K, X _L is the full label data of the labeled data pool, X _U is the full unlabeled data of the unlabeled data pool, X _u is, A candidate unlabeled data group of the unlabeled data pool, λ _i is an eigenvalue of the Empirical NTK matrix, and i is, i ∈ {1, ... … , │X _L ∪ X _u │}.

의 고유값을 산출하기 위하여, 각각의 블록(block) 행렬을 다음 수학식 5와 같이 대체할 수도 있다.

In order to calculate the eigenvalue of , each block matrix may be substituted as in Equation 5 below.

는, Empirical NTK(Neutral Tangent Kernel) matrix K로서,

는, 데이터쌍 x, x'에 의해서 계산되는 블록 행렬이며,

는,

의 차원(dimension)으로서, C는 출력 및 클래스 개수일 수 있다.here,

is an Empirical NTK (Neutral Tangent Kernel) matrix K,

is a block matrix calculated by the data pairs x, x',

Is,

As the dimension of , C may be the number of outputs and classes.

According to Equation 5, if Equation 4 is approximated to reduce the dimension, negative eigenvalues may come out. It is also possible to calculate the eigenvalues of (2) and calculate the eigenvalues of the input data set based on the eigenvalues of positive numbers. In the case of (2), a positive number can have a greater effect than a negative number on the convergence rate control for the model learning curve, so the eigenvalues of the matrix are calculated based on the positive eigenvalues without considering the negative eigenvalues. By doing so, it is possible to reduce the amount of computation required compared to Equation (4) and improve network performance.

As an example, the first neural network model in the present disclosure is operable based on Equations 2 to 5 above.

The computing device of the present disclosure may calculate an eigenvalue by inputting an input dataset including label data and one candidate unlabeled data group into a neural network model. Here, the reason for calculating the eigenvalue is to select data that can affect the learning curve convergence of the model in order to minimize the classification error of the classification model. That is, the representative eigenvalue of the input dataset including the label data and one candidate unlabeled data group may be a factor enabling it to be known that the convergence of the learning curve varies depending on which data is queried.

In addition, the computing device of the present disclosure may determine an unlabeled data group for label switching from among a plurality of candidate unlabeled data groups based on a plurality of representative eigenvalues calculated in response to an input data set. For example, the computing device of the present disclosure selects one input dataset having the highest eigenvalue among a plurality of eigenvalues calculated in response to the input dataset, and selects a candidate word included in the selected one input dataset. A label data group can be determined as an unlabeled data group for label switching. As another example, the computing device of the present disclosure selects a plurality of input datasets having a higher-level eigenvalue from among a plurality of eigenvalues calculated in response to the input dataset, and includes a plurality of input datasets included in the selected input dataset. A plurality of candidate unlabeled data groups may be determined as unlabeled data groups for label switching.

The computing device of the present disclosure may select unlabeled data maximizing a minimum value among n eigenvalues or select unlabeled data maximizing a maximum value among n eigenvalues for controlling the learning convergence rate of a model Also, it is possible to select unlabeled data that maximizes the average value of n eigenvalues.

Since the computing device of the present disclosure lacks training data when learning a neural network model for the medical field, it is possible to collect efficient data for learning from unlabeled data through eigenvalue operation without additional training of the model, so that general diseases In addition, it is possible to efficiently collect the learning data required for rare diseases.

4A and 4B are tables showing performance results of a data set configuration method and another method according to an embodiment of the present disclosure.

As shown in FIG. 4A, when the query size, which means the quantity of sample data for label conversion selected from the unlabeled data pool, is 50 and 100, classification accuracy for labeled images is controlled by the convergence rate control ( It can be seen that the Convergence Rate Control (CRC) method is more improved than other existing random methods, entropy methods, confidence methods, and EGL (Expected Gradient Length) methods. Other existing methods are methods for learning from one candidate unlabeled data randomly extracted from the unlabeled data pool, and it can be seen that the classification accuracy for images is lower than that of the convergence rate control method of the present disclosure.

As shown in Figure 4b, when the query size for SVHN (Street View House Numbers) data is 20, when the number of randomly sampled images per class in the first query is 12, the classification accuracy for labeled images is It can be seen that the convergence rate control (CRC) scheme of the present disclosure is more improved than other existing random schemes. It can be seen that other existing random methods have lower classification accuracy for images compared to the convergence rate control method of the present disclosure.

5A and 5B are graphs showing performance results between a data set configuration method and another method according to an embodiment of the present disclosure.

As shown in FIG. 5A , when the query size is 50, it is a graph comparing the performance of a number of learning models according to an increase in the quantity of labeled images, and a conventional random method with a different convergence rate control method of the present disclosure , it can be seen that the performance is improved compared to the entropy method, the confidence method, and the EGL (Expected Gradient Length) method.

And, as shown in FIG. 5B , as a graph showing the performance comparison of learning models when the query step for selecting data samples for query is continuously performed, the conventional random method having a different convergence rate control method of the present disclosure , it can be seen that it shows stable performance compared to the entropy method, the confidence method, and the EGL (Expected Gradient Length) method.

6 is a graph showing a convergence rate prediction performance result of a data set configuration method according to an embodiment of the present disclosure.

6 is a graph showing the distribution of minimum eigenvalues according to the number of epochs. As shown in FIG. 6 , the convergence rate control method of the present disclosure determines the number of epochs required for convergence through the distribution of the minimum eigenvalues. predictable. Epoch means the number of times learning is completed for the entire data set, and the convergence rate control method of the present disclosure can predict the number of epochs required for convergence through the distribution of minimum eigenvalues, so that efficient learning can be performed. .

7 is a simplified, general schematic diagram of an exemplary computing environment in which embodiments of the present disclosure may be implemented.

Although the present disclosure has been described above generally as being capable of being implemented by a computing device, those skilled in the art will appreciate that the present disclosure is a combination of hardware and software and/or in combination with computer-executable instructions and/or other program modules that may be executed on one or more computers. It will be appreciated that it can be implemented as a combination.

Generally, program modules include routines, programs, components, data structures, etc. that perform particular tasks or implement particular abstract data types. In addition, those skilled in the art will appreciate that the methods of the present disclosure are suitable for single-processor or multiprocessor computer systems, minicomputers, mainframe computers as well as personal computers, handheld computing devices, microprocessor-based or programmable consumer electronics, etc. (each of which is It will be appreciated that other computer system configurations may be implemented, including those that may operate in connection with one or more associated devices.

The described embodiments of the present disclosure may also be practiced in distributed computing environments where certain tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.

Computers typically include a variety of computer-readable media. Any medium accessible by a computer can be a computer readable medium, and such computer readable media includes volatile and nonvolatile media, transitory and non-transitory media, removable and non-transitory media. including removable media. By way of example, and not limitation, computer-readable media may include computer-readable storage media and computer-readable transmission media. Computer-readable storage media includes volatile and non-volatile media, temporary and non-transitory media, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. includes media. A computer-readable storage medium may be RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital video disk (DVD) or other optical disk storage device, magnetic cassette, magnetic tape, magnetic disk storage device, or other magnetic storage device. device, or any other medium that can be accessed by a computer and used to store the desired information.

Computer readable transmission media typically embodies computer readable instructions, data structures, program modules or other data, etc. in a modulated data signal such as a carrier wave or other transport mechanism, and Includes any information delivery medium. The term modulated data signal means a signal in which one or more of the characteristics of the signal is set or changed so as to encode information in the signal. By way of example, and not limitation, computer-readable transmission media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared, and other wireless media. Combinations of any of the above are also intended to be included within the scope of computer-readable transmission media.

An exemplary environment 1100 implementing various aspects of the disclosure is shown including a computer 1102 , the computer 1102 including a processing unit 1104 , a system memory 1106 , and a system bus 1108 . do. A system bus 1108 couples system components, including but not limited to system memory 1106 , to the processing device 1104 . The processing device 1104 may be any of a variety of commercially available processors. Dual processor and other multiprocessor architectures may also be used as processing unit 1104 .

The system bus 1108 may be any of several types of bus structures that may be further interconnected to a memory bus, a peripheral bus, and a local bus using any of a variety of commercial bus architectures. System memory 1106 includes read only memory (ROM) 1110 and random access memory (RAM) 1112 . A basic input/output system (BIOS) is stored in non-volatile memory 1110, such as ROM, EPROM, EEPROM, etc., which BIOS is the basic input/output system (BIOS) that helps transfer information between components within computer 1102, such as during startup. contains routines. RAM 1112 may also include high-speed RAM, such as static RAM, for caching data.

The computer 1102 may also be configured with an internal hard disk drive (HDD) 1114 (eg, EIDE, SATA) - this internal hard disk drive 1114 may also be configured for external use within a suitable chassis (not shown). Yes-, magnetic floppy disk drive (FDD) 1116 (eg, for reading from or writing to removable diskette 1118), and optical disk drive 1120 (eg, CD-ROM) for reading from, or writing to, disk 1122, or other high capacity optical media such as DVD. The hard disk drive 1114 , the magnetic disk drive 1116 , and the optical disk drive 1120 are connected to the system bus 1108 by the hard disk drive interface 1124 , the magnetic disk drive interface 1126 , and the optical drive interface 1128 , respectively. ) can be connected to The interface 1124 for implementing an external drive includes at least one or both of Universal Serial Bus (USB) and IEEE 1394 interface technologies.

These drives and their associated computer readable media provide non-volatile storage of data, data structures, computer executable instructions, and the like. In the case of computer 1102, drives and media correspond to storing any data in a suitable digital format. Although the description of computer readable media above refers to HDDs, removable magnetic disks, and removable optical media such as CDs or DVDs, those skilled in the art will use zip drives, magnetic cassettes, flash memory cards, cartridges, etc. It will be appreciated that other tangible computer-readable media such as etc. may also be used in the exemplary operating environment and any such media may include computer-executable instructions for performing the methods of the present disclosure.

A number of program modules may be stored in the drive and RAM 1112 , including an operating system 1130 , one or more application programs 1132 , other program modules 1134 , and program data 1136 . All or portions of the operating system, applications, modules, and/or data may also be cached in RAM 1112 . It will be appreciated that the present disclosure may be implemented in various commercially available operating systems or combinations of operating systems.

A user may enter commands and information into the computer 1102 via one or more wired/wireless input devices, for example, a pointing device such as a keyboard 1138 and a mouse 1140 . Other input devices (not shown) may include a microphone, IR remote control, joystick, game pad, stylus pen, touch screen, and the like. Although these and other input devices are often connected to the processing unit 1104 through an input device interface 1142 that is connected to the system bus 1108, parallel ports, IEEE 1394 serial ports, game ports, USB ports, IR interfaces, It may be connected by other interfaces, etc.

A monitor 1144 or other type of display device is also coupled to the system bus 1108 via an interface, such as a video adapter 1146 . In addition to the monitor 1144, the computer typically includes other peripheral output devices (not shown), such as speakers, printers, and the like.

Computer 1102 may operate in a networked environment using logical connections to one or more remote computers, such as remote computer(s) 1148 via wired and/or wireless communications. Remote computer(s) 1148 may be workstations, computing device computers, routers, personal computers, portable computers, microprocessor-based entertainment devices, peer devices, or other common network nodes, and are typically connected to computer 1102 . Although it includes many or all of the components described for it, only memory storage device 1150 is shown for simplicity. The logical connections shown include wired/wireless connections to a local area network (LAN) 1152 and/or a larger network, eg, a wide area network (WAN) 1154 . Such LAN and WAN networking environments are common in offices and companies, and facilitate enterprise-wide computer networks, such as intranets, all of which can be connected to a worldwide computer network, for example, the Internet.

When used in a LAN networking environment, the computer 1102 is connected to the local network 1152 through a wired and/or wireless communication network interface or adapter 1156 . Adapter 1156 may facilitate wired or wireless communication to LAN 1152 , which also includes a wireless access point installed therein for communicating with wireless adapter 1156 . When used in a WAN networking environment, the computer 1102 may include a modem 1158, be connected to a communications computing device on the WAN 1154, or establish communications over the WAN 1154, such as over the Internet. have other means. A modem 1158 , which may be internal or external and a wired or wireless device, is coupled to the system bus 1108 via a serial port interface 1142 . In a networked environment, program modules described for computer 1102 or portions thereof may be stored in remote memory/storage device 1150 . It will be appreciated that the network connections shown are exemplary and other means of establishing a communication link between the computers may be used.

Computer 1102 may be associated with any wireless device or object that is deployed and operates in wireless communication, eg, printers, scanners, desktop and/or portable computers, portable data assistants (PDAs), communication satellites, wireless detectable tags. It operates to communicate with any device or place, and phone. This includes at least Wi-Fi and Bluetooth wireless technologies. Accordingly, the communication may be a predefined structure as in a conventional network or may simply be an ad hoc communication between at least two devices.

Wi-Fi (Wireless Fidelity) makes it possible to connect to the Internet, etc. without a wire. Wi-Fi is a wireless technology such as cell phones that allows these devices, eg, computers, to transmit and receive data indoors and outdoors, ie anywhere within range of a base station. Wi-Fi networks use a radio technology called IEEE 802.11 (a, b, g, etc) to provide secure, reliable, and high-speed wireless connections. Wi-Fi can be used to connect computers to each other, to the Internet, and to wired networks (using IEEE 802.3 or Ethernet). Wi-Fi networks may operate in unlicensed 2.4 and 5 GHz radio bands, for example, at 11 Mbps (802.11a) or 54 Mbps (802.11b) data rates, or in products that include both bands (dual band). .

One of ordinary skill in the art of this disclosure will understand that information and signals may be represented using any of a variety of different technologies and techniques. For example, data, instructions, instructions, information, signals, bits, symbols, and chips that may be referenced in the above description are voltages, currents, electromagnetic waves, magnetic fields or particles, optical field particles or particles, or any combination thereof.

Those of ordinary skill in the art of the present disclosure will recognize that the various illustrative logical blocks, modules, processors, means, circuits and algorithm steps described in connection with the embodiments disclosed herein include electronic hardware, (convenience For this purpose, it will be understood that it may be implemented by various forms of program or design code (referred to herein as software) or a combination of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. A person skilled in the art of the present disclosure may implement the described functionality in various ways for each specific application, but such implementation decisions should not be interpreted as a departure from the scope of the present disclosure.

The various embodiments presented herein may be implemented as methods, apparatus, or articles of manufacture using standard programming and/or engineering techniques. The term article of manufacture includes a computer program, carrier, or media accessible from any computer-readable storage device. For example, computer-readable storage media include magnetic storage devices (eg, hard disks, floppy disks, magnetic strips, etc.), optical disks (eg, CDs, DVDs, etc.), smart cards, and flash drives. memory devices (eg, EEPROMs, cards, sticks, key drives, etc.). Also, various storage media presented herein include one or more devices and/or other machine-readable media for storing information.

It is to be understood that the specific order or hierarchy of steps in the presented processes is an example of exemplary approaches. Based on design priorities, it is to be understood that the specific order or hierarchy of steps in the processes may be rearranged within the scope of the present disclosure. The appended method claims present elements of the various steps in a sample order, but are not meant to be limited to the specific order or hierarchy presented.

The description of the presented embodiments is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Thus, the present disclosure is not intended to be limited to the embodiments presented herein, but is to be construed in the widest scope consistent with the principles and novel features presented herein.

Claims (10)

constructing a plurality of candidate unlabeled data groups from an unlabeled data pool;
constructing an input dataset including one candidate unlabeled data group from among the plurality of candidate unlabeled data groups and one or more labeled data;
inputting the configured input dataset into a first neural network model to calculate a representative eigenvalue corresponding to the input dataset; and
determining an unlabeled data group for label switching from among a plurality of candidate unlabeled data groups based on the calculated plurality of representative eigenvalues;
includes,
Each of the plurality of representative eigenvalues corresponds to one input data set,
How to construct a dataset. The method of claim 1,
determining a plurality of unlabeled data included in the determined unlabeled data group for label conversion as data to be labeled; and
After labeling is performed on a plurality of unlabeled data determined as the data to be labeled, a training dataset of a network model for performing a classification task on the unlabeled data on which the labeling has been performed and the one or more label data comprising the steps of;
further comprising,
How to construct a dataset. 3. The method of claim 2,
The first neural network model is a model for maximizing a minimum value, a maximum value, or an average value of eigenvalues corresponding to input data of the first neural network model using an Empirical Neural Tangent Kernel (Empirical NTK),
How to construct a dataset. The method of claim 1,
Constructing a plurality of candidate unlabeled data groups from the unlabeled data pool includes:
A plurality of unlabeled data to be included in each of the plurality of candidate unlabeled data groups is randomly extracted from the unlabeled data pool, or to be included in each of the plurality of candidate unlabeled data groups from the unlabeled data pool. A method of extracting a plurality of unlabeled data based on a preset number, configuring the plurality of candidate unlabeled data groups,
How to construct a dataset. The method of claim 1,
Calculating a representative eigenvalue corresponding to the input data set includes:
calculating first eigenvalues corresponding to each of the one or more label data and second eigenvalues corresponding to each of a plurality of unlabeled data included in the candidate unlabeled data group; and
calculating a representative eigenvalue corresponding to the input data set based on the calculated first and second eigenvalues;
containing,
How to construct a dataset. 6. The method of claim 5,
Calculating a representative eigenvalue corresponding to the input data set includes:
determining an eigenvalue having a minimum value or a maximum value among the calculated first eigenvalues and second eigenvalues as a representative eigenvalue corresponding to the input data set; or
determining an average value of the calculated first eigenvalues and second eigenvalues as a representative eigenvalue corresponding to the input data set;
How to construct a dataset. The method of claim 1,
Determining an unlabeled data group for label switching from among the plurality of candidate unlabeled data groups includes:
Selecting an input dataset having the highest eigenvalue from among a plurality of eigenvalues calculated in response to the plurality of input datasets, or selecting a higher-level selecting a plurality of input data sets having eigenvalues; and
determining a candidate unlabeled data group included in the one or more selected input datasets as the unlabeled data group for label switching;
containing,
How to construct a dataset. 8. The method of claim 7,
The number of the plurality of input datasets having the higher-level eigenvalue is linked with the target number of unlabeled data to be labeled,
How to construct a dataset. A computer program stored in a computer readable storage medium, wherein the computer program, when executed on one or more processors, performs the following steps for constructing a data set, the steps comprising:
constructing a plurality of candidate unlabeled data groups from the unlabeled data pool;
constructing an input dataset including one candidate unlabeled data group and one or more label data from among the plurality of candidate unlabeled data groups;
inputting the configured input dataset into a first neural network model to calculate representative eigenvalues corresponding to the input dataset; and
determining an unlabeled data group for label switching from among a plurality of candidate unlabeled data groups based on the calculated plurality of representative eigenvalues;
includes,
Each of the plurality of representative eigenvalues corresponds to one input data set,
A computer program stored on a computer readable storage medium. A computing device for providing a method of constructing a dataset, comprising:
a processor including one or more cores; and
Memory;
including,
The processor is:
configure a plurality of candidate unlabeled data groups from the unlabeled data pool;
configure an input dataset including one candidate unlabeled data group and one or more label data from among the plurality of candidate unlabeled data groups;
inputting the configured input dataset into a first neural network model to calculate a representative eigenvalue corresponding to the input dataset; and
determining an unlabeled data group for label switching from among a plurality of candidate unlabeled data groups based on the calculated plurality of representative eigenvalues; and
Each of the plurality of representative eigenvalues corresponds to one input data set,
computing device.

Images