Error determination device, error determination method, and program

Abstract

An error determination device provided with: a classification estimation process observation unit which acquires estimation process data from a classification estimation unit that estimates the classification of data to be classified, and generates an estimation process feature vector on the basis of the estimation process data; a probability estimation unit which generates an estimated probability vector consisting of the probability of the data to be classified belonging to each class, on the basis of the estimation process feature vector; and an error determination unit which determines whether the result of the classification by the classification estimation unit is correct or incorrect on the basis of the estimated probability vector, and outputs the classification result, the result of the determination of whether the classification result is correct or incorrect, and the estimated probability vector.

Description

[Technical field]

[0001] The present invention relates to a technology for classifying information. One example of the application field of this technology is a technology for security operators who manage security systems against cyber attacks, such as IPS (Intrusion Prevention System) and antivirus software, to automatically classify threat information using machine learning technology or the like. [Background technology]

[0002] Security operators who handle security systems against cyber attacks compile threat information about cyber attack activities, including attackers, their actions and methods, vulnerabilities, etc. This threat information needs to be generated daily, so security operators need to classify the threat information continuously and sequentially.

[0003] Conventional classification techniques are disclosed in, for example, Patent Documents 1 and 2. These conventional techniques propose a technique for automatically determining whether data classification is correct, which makes it possible to semi-automate the data classification task by leaving the task of classifying data that is thought to be incorrect to a human. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2020-024513 A [Patent Document 2] JP 2020-160642 A Summary of the Invention [Problem to be solved by the invention]

[0005] In the conventional technology, it is possible to classify data and judge whether it is correct or not with high accuracy, but there is a problem in that it is not possible to output the probability of belonging to each classified class.

[0006] The present invention has been made in consideration of the above points, and has an object to provide a technique that makes it possible to output the probability of belonging to each class in addition to the correctness or incorrectness of classification of certain data. [Means for solving the problem]

[0007] According to the disclosed technology, a classification estimation process observation unit that acquires data on the estimation process from a classification estimation unit that estimates the classification of data to be classified and generates an estimation process feature vector based on the data; a probability estimation unit that generates an estimated probability vector including a probability of belonging to each class of the classification target data based on the estimation process feature vector; an error determination unit that determines whether a classification result by the classification estimation unit is correct or incorrect based on the estimated probability vector, and outputs the classification result, a result of the determination of correctness or incorrectness, and the estimated probability vector; The probability estimation unit is the classification estimation unit Obtained during the study of The machine learning model is trained using the proportion of data classified into each class as correct answer data. An error determination device is provided. Effect of the Invention

[0008] According to the disclosed technology, in addition to the correctness or incorrectness of classification of certain data, it is possible to output the probability of belonging to each class. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 is a diagram for explaining an overview of an embodiment of the present invention. [Diagram 2] FIG. 1 is a diagram for explaining an overview of an embodiment of the present invention. [Diagram 3] 1 is a configuration diagram of a classification device 100 according to an embodiment of the present invention. [Figure 4] 13 is a flowchart for explaining a generation method of the classification probability correction vector calculation unit. [Diagram 5]FIG. 2 illustrates an example of a hardware configuration of the apparatus. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Hereinafter, an embodiment of the present invention (the present embodiment) will be described with reference to the drawings. The embodiment described below is merely an example, and the embodiment to which the present invention is applied is not limited to the following embodiment.

[0011] (Outline of the embodiment) The outline of this embodiment will be described with reference to Fig. 1. Fig. 1(a) shows an image of the prior art, in which only one accuracy rate is output from a function (neural network) that calculates the certainty of classification.

[0012] In contrast, in the technique according to this embodiment shown in FIG. 1(b), the function for calculating the certainty of classification outputs all the probabilities of belonging to each class.

[0013] 2 shows an overview of the processing contents of the classification device according to this embodiment. The Classifier (corresponding to the classification estimation unit 110 described later) performs learning using input data and a class that is a correct answer. During this learning, the classification estimation unit 110 predicts the class of the data many times. The proportion of predicted classes is used as training data for a multi-class confidence calculation function in the Rejecter (corresponding to the classification probability correction vector calculation unit 122 described later).

[0014] For example, for a given data set, if the classifier predicted class A 70 times, class B 20 times, and class C 10 times during supervised learning, then the labels would be [0.7, 0.2, 0.1].

[0015] The predicted class ratios (the above-mentioned labels) are used as correct answer data to learn the multi-class confidence calculation function. This makes it possible to obtain a multi-class confidence calculation function (classification probability correction vector calculation unit 122) that can predict the probability of belonging to each class for a certain piece of data with high accuracy.

[0016] The configuration and operation of the classification device according to this embodiment will be described in detail below.

[0017] (Example of device configuration) Fig. 3 shows a functional configuration diagram of a classification device 100 according to an embodiment of the present invention. As shown in Fig. 1, the classification device 100 has a classification estimation unit 110 and an error determination processing unit 120. The error determination processing unit 120 has a classification estimation process observation unit 121, a classification probability correction vector calculation unit 122, a classification probability estimation unit 123, and an error determination unit 124.

[0018] The classification device 100 may also include a learning unit 130. The learning unit 130 performs learning operations such as parameter adjustment in supervised learning of the classification estimation unit 110, the classification probability correction vector calculation unit 122, and the like. Note that, in a state where learning has been completed, the classification device 100 may not include the learning unit 130. Also, a device including the learning unit 130 as shown in FIG. 3 may be called a learning device.

[0019] The classification estimation unit 110 and the error determination processing unit 120 may be configured as separate devices connected via a network, in which case the error determination processing unit 120 may be referred to as an error determination device. Also, a device including the classification estimation unit 110 and the error determination processing unit 120 may be referred to as an error determination device. An overview of the operation of each unit during inference in the classification device 100 is as follows.

[0020] (Operation overview) First, data to be classified is input to the classification estimation unit 110. Data to be classified is data to be classified in some way using this system, and corresponds to threat information, for example.

[0021] The classification estimation unit 110 estimates the classification of the data to be classified. The method and model for the estimation is assumed to be an artificial intelligence-related technique such as SVM or neural network, but is not limited to this.

[0022] The classification estimation process observation unit 121 observes the calculation process when the classification estimation unit 110 estimates the data to be classified, converts it into a feature vector (a feature vector of the estimation process), and outputs the feature vector.

[0023] The classification probability correction vector calculation unit 122 receives the feature vector of the estimation process from the classification estimation process observation unit 121, and calculates a vector for correcting the classification probability. This classification probability correction vector calculation unit 122 is generated by machine learning. The method of generation will be described later.

[0024] The classification probability correction vector output from the classification probability correction vector calculation unit 122 is a numeric vector used to correct the classification probability, and is a real-valued vector having a dimension equal to the number of classes. Note that the classification probability correction vector output from the classification probability correction vector calculation unit 122 may be used as a vector of the probability of the data to be classified belonging to each class (estimated probability vector for each class).

[0025] The classification probability estimation unit 123 receives the feature vector of the estimation process from the classification estimation process observation unit 121, receives the classification probability correction vector from the classification probability correction vector calculation unit 122, and calculates the probability of belonging to each class of the classification target data. There are multiple implementation methods, and the details will be described later. There are also cases where the feature vector of the estimation process, a part of the feature vector of the estimation process, or the classification probability correction vector is output as it is. In other words, the classification probability estimation unit 123 may not be provided, and the classification probability correction vector calculation unit 122 may be used as the classification probability estimation unit 123.

[0026] The classification probability correction vector calculation unit 122 and the classification probability estimation unit 123 may be collectively referred to as a “probability estimation unit.” A functional unit including the classification probability correction vector calculation unit 122 and the classification probability estimation unit 123 may be referred to as a “probability estimation unit.”

[0027] The error determination unit 124 receives the classification result, the feature vector of the estimation process, and the estimated probability for each classification from the classification estimation unit 110, the classification estimation process observation unit 121, and the classification probability estimation unit 123, respectively, and determines whether the classification estimated by the classification estimation unit 110 is "correct" or "incorrect" based on these. The error determination unit 123 also outputs the error determination result, the classification result, and the estimated probability vector for each class as the result of the entire system.

[0028] The classification result is a classification result of the data to be classified, and indicates one or more "classes" determined from a predetermined class list.

[0029] The estimated probability vector for each class is the probability value of each class output by the classification probability estimation unit 123. For example, assuming that certain data is classified into classes A, B, and C, the probability that the classification is A is 0%, B is □%, and C is △%. The error determination result is the determination result of whether or not the classification is incorrect.

[0030] The processing operation of each unit in the error determination processing unit 120 will be described in detail below.

[0031] (Classification estimation process observation unit 121) First, a description will be given of the classification estimation process observation unit 121. The classification estimation process observation unit 121 observes the calculation process (data of the estimation process) when the classification estimation unit 110 estimates data to be classified, and configures and outputs a feature vector (feature vector of the estimation process).

[0032] The constructed feature vector basically differs depending on the model in the classification estimation unit 110. Here, the following (1), (2), and (3) will be explained as examples of representative feature vectors.

[0033] (1) A feature vector that can be commonly constructed by any classification estimation module Examples of feature vectors that can be commonly constructed by any classification estimation module include (1-1) and (1-2) below.

[0034] (1-1) Feature vectors obtained by converting the data to be classified into numerical vectors When the classification estimation unit 110 is constructed using a machine learning model, the classification target data is converted into a feature vector, which is a vector of numerical values, inside the classification estimation unit 110. The vector of numerical values ​​is observed and used as the feature vector for the estimation process.

[0035] (1-2) Estimated probability vector for each class When the classification estimation unit 110 is constructed using a machine learning model that performs multi-class classification, classification scoring is performed for each class. The scoring is observed, converted into a probability value, and arranged to obtain an estimated probability vector for each class, which is used as the feature vector for the estimation process.

[0036] Specifically, the classification estimation process observation unit 121 converts the scores (real values) for each class obtained by observing the classification estimation unit 110 into a vector of probabilities by using a softmax function. That is, in the case of n-class classification, the scores of each class are expressed as a1, . . . , a n Then, the probability of class k is p k can be calculated as follows:

[0037]

number

[0038] Alternatively, the prediction score of any classifier may be used as the feature vector in the estimation process. For example, when the classification estimation unit 110 performs class classification using a Support Vector Machine (SVM), the distance from the boundary surface can be observed as a prediction score and used as the feature vector in the estimation process.

[0039] (3) Feature vector of the ensemble classifier When the classification estimation unit 110 is configured with a plurality of machine learning models, each machine learning model can obtain one or more of the above-mentioned "feature vector obtained by converting data to be classified into a numerical vector," "estimated probability vector for each estimated class," and "logit vector." A vector obtained by concatenating the vectors of the plurality of machine learning models can be output as a feature vector for the estimation process.

[0040] (Error determination unit 124) Next, the error determination unit 124 will be described. As shown in Fig. 3, the error determination unit 124 receives the classification result, the feature vector of the estimation process, and the estimation probability for each class, and determines whether the classification estimated by the classification estimation unit 110 is "correct" or "incorrect" based on these. Note that only one of the feature vector of the estimation process and the estimation probability for each class may be used.

[0041] Moreover, the error determining unit 124 outputs the error determination result, the classification result, and the estimated probability for each class as the result of the entire system.

[0042] The error determination method executed by the error determination unit 124 is not limited to a specific method, but may be, for example, any of the following methods 1 to 3. Any two or all of the methods 1 to 3 may be applied in combination. Also, the following methods 1 to 3 are examples, and methods other than the following methods 1 to 3 may be used.

[0043] [Method 1] In method 1, the error determination unit 124 performs threshold determination of an index called certainty. Specifically, the error determination unit 124 obtains the maximum value of the estimated probabilities for each class, and sets the maximum value as the certainty. If the certainty is equal to or greater than a set threshold, the classification result into that class is determined to be "correct," and if it is less than the set threshold, it is determined to be "incorrect."

[0044] In addition, the user can arbitrarily set the error determination unit 124 to perform any calculation using the classification result, the feature vector of the estimation process, or the estimation probability for each class in calculating the confidence level.

[0045] For example, the error determination unit 124 may determine the confidence level as the difference (m1-m2) between the maximum value (m1) of the estimated probability for each class and the second largest value (m2). The estimated probability of any order, such as the maximum value and the third value, the fourth value, etc., can also be calculated in the same way.

[0046] [Method 2] In method 2, the error determination unit 124 performs threshold determination on an index called uncertainty. Specifically, the error determination unit 124 calculates the average information amount (entropy) of the estimated probability for each class, and sets this value as the uncertainty. If the uncertainty is equal to or greater than a set threshold, the classification result is determined to be "incorrect," and if it is less than the threshold, the classification result is determined to be "correct."

[0047] In n-class classification, the probability of each class is p1, ,p n Then, the average information content can be calculated as follows:

[0048]

number

[0049] [Method 3] The judgment may be made by an error judgment unit created by machine learning, as in the conventional techniques disclosed in Patent Documents 1 and 2. In addition, it is also possible to make the judgment using any conventional technique other than the conventional techniques disclosed in Patent Documents 1 and 2.

[0050] (Classification probability estimation unit 123) Next, the classification probability estimation unit 123 will be described in detail. As shown in Fig. 3, the classification probability estimation unit 123 receives the feature vector of the estimation process and the classification probability correction vector, and calculates the estimated probability vector for each class. The implementation method is not limited to a specific method, but for example, methods 1 to 3 described below can be used. Note that the method that can be implemented depends on what is included in the feature vector of the estimation process.

[0051] [Method 1] When the feature vector in the estimation process includes an "estimated probability for each class," the classification probability estimation unit 123 extracts the "estimated probability for each class" and outputs it as an estimated probability vector for each class. In this case, the extracted "estimated probability for each class" may be output as is, or may be corrected with a classification probability correction vector and output. The correction may be, for example, taking the average of the extracted "estimated probability for each class" and the estimated probability for each class in the classification probability correction vector, or may be a result of other processing.

[0052] [Method 2] In method 2, the classification probability estimation unit 123 outputs the classification probability correction vector as it is as an estimated probability vector for each class. In this case, the classification probability estimation unit 123 may not be provided and the classification probability correction vector calculation unit 122 may be used as the classification probability estimation unit 123.

[0053] [Method 3] In method 3, when the feature vector of the estimation process includes the “logit vector” shown in (2) of the classification estimation process observation unit 121 described above, an estimated probability vector for each class is calculated by either method 3-1 or method 3-2 below.

[0054] [Method 3-1] For n-class classification, the logit vector is [a1, ,a n ] T , the classification probability correction vector is [b1, ,b n ] T Then, the probability of class k is p k can be calculated as follows:

[0055]

number

[0056] [Method 3-2] For n-class classification, the logit vector is [a1, ,a n ] T , the classification probability correction vector is [b1, ,b n ] T The maximum value of the elements in the classification probability correction vector b max and obtain the probability p k is calculated as follows:

[0057]

number

[0058] (Classification probability correction vector calculation unit 122) Next, the classification probability correction vector calculation unit 122 will be described in detail. As shown in Fig. 3, the classification probability correction vector calculation unit 122 receives the feature vector of the estimation process, calculates and outputs the classification probability correction vector. The classification probability correction vector is an n-dimensional real-valued vector in the case of n-class classification.

[0059] The classification probability correction vector calculation unit 122 is constructed using a machine learning model capable of estimating a plurality of real values. A method of generation (parameter tuning) of the classification probability correction vector calculation unit 122 will be described later.

[0060] As a machine learning model capable of estimating a plurality of real values ​​used as the classification probability correction vector calculation unit 122, for example, a neural network, a logistic regression, a support vector regression (SVR), or the like can be used. And so on.

[0061] When a neural network is used as the classification probability correction vector calculation unit 122, multiple real values ​​can be estimated with a single model. However, logistic regression and SVR cannot estimate multiple real values ​​by themselves. In such a case, n machine learning models are prepared and real values ​​corresponding to each class are inferred.

[0062] Note that the listed models, such as neural networks, logistic regression, and support vector regression, are merely examples, and any machine learning model can be used as long as it has a structure that allows multiple real values ​​to be estimated using a machine learning model.

[0063] (Method of generation by classification probability correction vector calculation unit 122) Next, a generation method (parameter adjustment method) of the classification probability correction vector calculation unit 122 will be described following the steps of the flowchart in Fig. 4. Here, it is assumed that the number of classifications is n. In the following explanation, for ease of understanding, the learning classification target data list is labeled (A), the classification ratio list for each learning classification target data is labeled (B), and the estimation process feature vector list is labeled (C). In the following explanation, it is assumed that each unit is implemented with a neural network, but this is merely one example.

[0064] Moreover, the following processes relating to learning are executed by the learning unit 130. The learning unit 130 includes a function of holding learning data, a parameter adjustment function (such as a function of executing an error backpropagation method), and the like.

[0065] <s1> In S1 (step 1), (A) a classification target data list for learning and the classification estimation unit 110 before parameter adjustment are prepared and held in the learning unit .

[0066] <s2> The classification estimation unit 110 adjusts parameters using a general supervised learning method. During this process, the learning unit 130 acquires (B) a classification ratio list for each learning classification target data. (B) A classification ratio list for each learning classification target data will be described.

[0067] Neural networks are a typical example, but in general supervised learning, data is classified many times during the process. Through this repetition, the classification ratios for each training target data are compiled into a list, which is (B) a classification ratio list for each training target data.

[0068] For example, in the case of classification into three classes, suppose that the neural network classifies data 1 and data 2 100 times during the learning process. During that process, suppose that data 1 is classified into class 1 50 times, into class 2 30 times, and into class 3 20 times. Also, suppose that data 2 is classified into class 1 10 times, into class 2 70 times, and into class 3 20 times. In this case, (B) the classification ratio list for each learning classification target data is [[0.5,0.3,0.2] T ,[0.1,0.7,0.2] T ]

[0069] <s3> In S3, (A) each element of the learning classification target data list is input to the classification estimation unit 110 whose parameters were adjusted in S2, and the classification estimation process observation unit 121 acquires a feature vector of the estimation process, which is used as (C) an estimation process feature vector list.

[0070] <s4> In S4, multiple pseudo feature vectors generated using random numbers or the like are added to the estimation process feature vector list (C). In addition, n-dimensional vectors with all elements set to 1 / n are added to the classification ratio list for each learning classification target data (B) in the same number as the pseudo feature vectors added to (C).

[0071] For example, when performing three-class classification, the vector added to (B) is [1 / 3, 1 / 3, 1 / 3]. T The number of additions is set by the user of the classification device.

[0072] By making the above-mentioned additions, it becomes more robust against random feature vectors, and the accuracy of classification of threat information having unknown features is improved. Note that S4 is not essential, and S4 may not be performed.

[0073] <s5> In S5, the estimation process feature vector list (C) processed in S4 is input, and the classification ratio list for each learning classification target data (B) processed in S4 is output (correct answer), and the classification probability correction vector calculation unit 122 is generated by supervised learning. In other words, the parameters of the classification probability correction vector calculation unit 122 are adjusted by supervised learning.

[0074] (Hardware configuration example) The above-mentioned classification device 100 (as well as the error determination device) can be realized by, for example, having a computer execute a program describing the processing contents described in this embodiment. This computer may be a physical computer or a virtual machine on the cloud.

[0075] That is, the classification device 100 can be realized by executing a program corresponding to the processing performed by the classification device 100, using hardware resources such as a CPU and memory built into a computer. The program can be recorded on a computer-readable recording medium (such as a portable memory) and stored or distributed. The program can also be provided via a network such as the Internet or email.

[0076] Fig. 5 is a diagram showing an example of the hardware configuration of the computer. The computer in Fig. 5 has a drive device 1000, an auxiliary storage device 1002, a memory device 1003, a CPU 1004, an interface device 1005, a display device 1006, an input device 1007, an output device 1008, etc., which are all connected to each other via a bus BS.

[0077] The program for realizing the processing on the computer is provided by a recording medium 1001 such as a CD-ROM or a memory card. When the recording medium 1001 storing the program is set in the drive device 1000, the program is installed from the recording medium 1001 to the auxiliary storage device 1002 via the drive device 1000. However, the program does not necessarily have to be installed from the recording medium 1001, but may be downloaded from another computer via a network. The auxiliary storage device 1002 stores the installed program as well as necessary files, data, etc.

[0078] The memory device 1003 reads out and stores the program from the auxiliary storage device 1002 when an instruction to start the program is received. The CPU 1004 realizes functions related to the light touch maintenance device 100 in accordance with the program stored in the memory device 1003. The interface device 1005 is used as an interface for connecting to a network, various measuring devices, exercise intervention devices, etc. The display device 1006 displays a GUI (Graphical User Interface) etc. according to a program. The input device 1007 is composed of a keyboard and mouse, buttons, a touch panel, etc., and is used to input various operation instructions. The output device 1008 outputs the calculation results.

[0079] (Effects of the embodiment) The technology according to the present embodiment makes it possible to output the probability of each class for certain data in addition to judging whether the data is correct or not. For example, assume that certain data is classified into classes A, B, and C. The classification device 100 can estimate the probability that the classification is A, such as 〇%, B, □%, and C, and present the estimate to a human.

[0080] Furthermore, in the technology according to this embodiment, the proportion of classification estimated for each piece of training data during training by the classification estimation unit 110 is acquired, and is used for training by the classification probability correction vector calculation unit 122. With this kind of ingenuity, the accuracy of determining whether something is true or false is improved compared to the conventional technology, and the accuracy of estimating the probability of each class estimated within the system is also improved.

[0081] (Summary of the embodiment) This specification discloses at least the following error determination device, error determination method, and program. (Section 1) a classification estimation process observation unit that acquires data on the estimation process from a classification estimation unit that estimates the classification of the data to be classified, and generates an estimation process feature vector based on the data; a probability estimation unit that generates an estimated probability vector including a probability of belonging to each class of the classification target data based on the estimation process feature vector; an error determination unit that determines whether a classification result by the classification estimation unit is correct or incorrect based on the estimated probability vector, and outputs the classification result, a result of determining whether the classification result is correct or incorrect, and the estimated probability vector; An error determination device comprising: (Section 2) The probability estimation unit has a machine learning model trained using the proportion of classification into each class for each training data acquired during training of the classification estimation unit as correct answer data. 2. An error determination device according to claim 1. (Section 3) The error determination unit determines whether the classification result is correct by comparing a maximum value of the estimated probabilities for each class in the estimated probability vector with a threshold value. 3. An error determination device according to claim 1 or 2. (Section 4) The error determination unit determines whether the classification result is correct by comparing an average information amount of the estimated probability for each class in the estimated probability vector with a threshold value. 3. An error determination device according to claim 1 or 2. (Section 5) 1. A computer-implemented method for determining an error, comprising: A step of acquiring data on an estimation process from a classification estimation unit that estimates a classification of data to be classified, and generating an estimation process feature vector based on the data; generating an estimated probability vector including a probability of belonging to each class of the classification target data based on the estimation process feature vector; determining whether a classification result by the classification estimation unit is correct or incorrect based on the estimated probability vector, and outputting the classification result, a result of the determination of correctness or incorrectness, and the estimated probability vector; An error determination method comprising: (Section 6) A program for causing a computer to function as a probability estimation unit, a classification estimation process observation unit, and an error determination unit in the error determination device according to any one of claims 1 to 4.

[0082] Although the present embodiment has been described above, the present invention is not limited to such a specific embodiment, and various modifications and variations are possible within the scope of the gist of the present invention described in the claims. [Explanation of symbols]

[0083] 100 Classifier 110 Classification estimation part 120 Error judgment processing unit 121 Classification and Estimation Process Observation Unit 122 Classification Probability Correction Vector Calculation Unit 123 Classification Probability Estimation Unit 124 Error judgment unit 130 Learning Department 1000 Drive Device 1001 Recording media 1002 Auxiliary storage 1003 Memory device 1004 CPU 1005 Interface device 1006 Display device 1007 Input Device 1008 Output device

Claims

1. a classification estimation process observation unit that acquires data on the estimation process from a classification estimation unit that estimates the classification of the data to be classified, and generates an estimation process feature vector based on the data; a probability estimation unit that generates an estimated probability vector including a probability of belonging to each class of the classification target data based on the estimation process feature vector; an error determination unit that determines whether a classification result by the classification estimation unit is correct or incorrect based on the estimated probability vector, and outputs the classification result, a result of the determination of correctness or incorrectness, and the estimated probability vector; The probability estimation unit has a machine learning model trained using the proportion of classification into each class for each training data acquired during training of the classification estimation unit as correct answer data. Error determination device.

2. The error determination unit determines whether the classification result is correct by comparing a maximum value of the estimated probabilities for each class in the estimated probability vector with a threshold value. The error determination device according to claim 1 .

3. The error determination unit determines whether the classification result is correct by comparing an average information amount of the estimated probability for each class in the estimated probability vector with a threshold value. The error determination device according to claim 1 .

4. An error determination method executed by a computer functioning as a classification estimation process observation unit, a probability estimation unit, and an error determination unit, comprising: a classification estimation process observation step in which the classification estimation process observation unit acquires data on the estimation process from a classification estimation unit that estimates the classification of data to be classified, and generates an estimation process feature vector based on the data; a probability estimation step in which the probability estimation unit generates an estimated probability vector including a probability of belonging to each class of the classification target data based on the estimation process feature vector; an error determination step in which the error determination unit determines whether a classification result by the classification estimation unit is correct or incorrect based on the estimated probability vector, and outputs the classification result, a result of the determination of correctness or incorrectness, and the estimated probability vector; The probability estimation unit has a machine learning model trained using the proportion of classification into each class for each training data acquired during training of the classification estimation unit as correct answer data. Error detection method.

5. 4. A program for causing a computer to function as a probability estimation unit, a classification estimation process observation unit, and an error determination unit in the error determination device according to claim 1.

Images