Abnormality degree calculation system and method
The abnormality degree calculation system improves detection accuracy by using sensor data from devices of the same higher-level concept but different lower-level concepts, optimizing normal distribution estimation through concept type assignment and likelihood calculations, addressing the challenge of insufficient data and distribution differences.
Patent Information
- Application Number
- JP2024147507
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-08-29
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2041-02-22
AI Technical Summary
Existing anomaly detection methods struggle with insufficient data for estimating normal distributions, leading to reduced detection accuracy when normal distributions of sensor data differ between devices of the same type, especially when only normal data from the target device is used.
An abnormality degree calculation system that utilizes sensor data from devices of the same higher-level concept but different lower-level concepts to improve normal distribution estimation, incorporating a concept type assignment, feature vector extraction, likelihood calculation, loss function optimization, and relearning necessity determination to enhance detection accuracy.
The system accurately estimates the normal distribution of the target device's sensor data even with limited data, enhancing detection accuracy by treating data from similar devices as normal or abnormal based on likelihood calculations and relearning necessity.
Smart Images

Figure 2024159935000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to an anomaly calculation system and method. [Background technology]
[0002] In general, in anomaly detection, it is difficult to obtain a sufficient amount of anomaly data for all types of anomalies, so a method is often adopted in which a normal distribution is estimated using only normal data and then used to determine anomalies.
[0003] On the other hand, even in methods that estimate the normal distribution using only normal data, there are cases where sufficient normal data cannot be collected or where the normal distribution changes while a sufficient amount of normal data is collected, so a method that can obtain sufficient detection accuracy with a small amount of normal data is needed.
[0004] Therefore, a technology has been proposed that uses normal data from machines similar to the target machine in addition to the normal data from the target machine to improve the accuracy of estimating the normal distribution, thereby achieving sufficient detection accuracy even when the amount of normal data for the target machine is small (Patent Document 1).
[0005] Patent document 1 states that "The processing unit 11 calculates a first value which is a value indicating to what extent the derived status value of the gazed-after device deviates from a first distribution area which represents the distribution of status values of each device of the same type as the gazed-after device. The processing unit 11 calculates a second value which is a value indicating to what extent the derived status value of the gazed-after device deviates from a second distribution area which represents the distribution of each past status value of the gazed-after device. The processing unit 11 determines whether the gazed-after device is abnormal based on the first value and the second value." [Prior art documents] [Patent documents]
[0006] [Patent Document 1] JP 2019-008354 A Summary of the Invention [Problem to be solved by the invention]
[0007] The invention described in Patent Document 1 assumes that the sensor data of a device that is the target of anomaly detection and the sensor data of a device of the same type as the target device follow the same normal distribution. Therefore, the technology in Patent Document 1 can be used only when the target device and the device of the same type as the target device are the same in a certain concept, such as the model, and are also the same in a concept lower than the above concept, such as the type, i.e., when both the model and type can be considered to be the same.
[0008] Therefore, Patent Document 1 does not anticipate a case in which the detection target device and a device of the same type as the detection target device are identical in terms of a higher-level concept, such as model, but differ in terms of a lower-level concept, such as model, and the normal distributions of the two devices are different.
[0009] When the normal distribution of the sensor data of the target device differs from the normal distribution of the sensor data of the same type of device, the technology of Patent Document 1 learns the non-normal distribution area of the target device as a normal distribution, resulting in lower detection accuracy compared to when only the sensor data of the target device is used.
[0010] Therefore, an object of the present disclosure is to provide an anomaly degree calculation system and method that are capable of calculating the anomaly degree of a device with higher accuracy. [Means for solving the problem]
[0011] In order to solve the above problem, an anomaly degree calculation system according to one aspect of the present invention is an anomaly degree calculation system that calculates the degree of anomaly of a target device, and includes a concept type assignment unit that assigns a predetermined concept type based on the identification number of the target device, a feature vector extraction unit that extracts a feature vector based on sensor data obtained from a sensor corresponding to the target device, a likelihood calculation unit that calculates the likelihood of the feature vector using a machine learning model obtained from a learning database, a loss calculation unit that calculates a loss using a loss function defined as a function of the likelihood calculated by the likelihood calculation unit, a model update unit that updates the machine learning model using the loss calculated by the loss calculation unit and the trained machine learning model, a relearning necessity determination unit that determines whether re-learning is necessary from the likelihood calculated by the likelihood calculation unit, and an anomaly degree calculation unit that calculates the degree of anomaly when the relearning necessity determination unit determines that re-learning is unnecessary. Effect of the Invention
[0012] According to the present invention, when an abnormality is detected in a target device, the need for re-learning is determined from the calculated likelihood, and if it is determined that re-learning is not necessary, the degree of anomaly is calculated, thereby improving the accuracy of calculating the degree of anomaly. [Brief description of the drawings]
[0013] [Figure 1] FIG. 1 is an explanatory diagram showing an overall outline of the present embodiment; [Diagram 2] FIG. 13 is an explanatory diagram of a process for extracting a concept type from an input. [Diagram 3] FIG. 1 is an explanatory diagram of a process for calculating loss using data from a training database. [Figure 4] FIG. 13 is a diagram illustrating the relationship between the negative log likelihood and the number of model updates. [Diagram 5] FIG. 4 is an explanatory diagram of a method for determining whether or not re-learning is necessary. [Figure 6] FIG. 2 is a diagram showing the hardware and software configuration of the anomaly degree calculation device. [Figure 7] FIG. 4 is a block diagram of a feature vector extraction unit. [Figure 8] Block diagram during learning. [Figure 9]13 is a processing flow of the anomaly calculation system during learning. [Figure 10] FIG. 13 is a block diagram showing the configuration when calculating the degree of abnormality. [Figure 11] 13 is a process flow of the anomaly degree calculation system when calculating an anomaly degree. [Figure 12] FIG. 1 is an explanatory diagram showing the difference between the present invention and a conventional method. [Figure 13] FIG. 11 is a block configuration diagram for calculating an abnormality degree according to the second embodiment. [Figure 14] 13 is a process flow of the anomaly degree calculation system when calculating an anomaly degree. [Figure 15] FIG. 13 is a block configuration diagram during learning according to the third embodiment. [Figure 16] 13 is a processing flow of the anomaly calculation system during learning. [Figure 17] FIG. 13 is a block diagram showing the configuration when calculating the degree of abnormality. [Figure 18] 13 is a process flow of the anomaly degree calculation system when calculating an anomaly degree. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0014] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In an environment where a detection target device and a device with the same higher concept as the detection target device but a different lower concept (also called an identical conceptual device or other device) coexist, even if the sensor data of the detection target device and the sensor data of the identical conceptual device follow different normal distributions, the anomaly degree calculation system according to this embodiment improves the accuracy of detecting anomalies by using both sensor data. As a result, the anomaly degree calculation system according to this embodiment can obtain sufficient detection accuracy using a relatively small amount of normal data of the detection target device.
[0015] The anomaly degree calculation system of this embodiment acquires input D0 from n devices M including a target device, which are the same in terms of a higher-level concept, for example, model. The input D0 includes a model number D1, which is an example of a lower-level concept of the model, and sensor data D2.
[0016] The abnormality degree calculation system of this embodiment includes a concept classifier 11 that outputs a concept classifier D3 indicating whether each device is a device to be detected (target device) or a device of the same conceptual level as the device to be detected (other device) based on the model number D1.
[0017] Furthermore, the anomaly degree calculation system of this embodiment includes a feature vector extraction unit 12 that outputs a feature D4 from the sensor data D2, a likelihood calculation unit 13 that outputs a likelihood D5 from the feature D4 using a machine learning model, a loss calculation unit 14 that calculates a loss value D6 using a loss function defined by a function of the likelihood D5 so as to increase the likelihood of data from the detection target device and decrease the likelihood of data from devices of the same concept, a learning unit 15 that updates the parameters of the machine learning model of the likelihood calculation unit 13 so as to minimize the loss value D6, a re-learning necessity determination unit 16 that determines whether re-learning is necessary from the likelihood D5 and the concept type D3, and an anomaly degree calculation unit 17 that determines the degree of anomaly from the likelihood D7 of the detection target when it is determined that re-learning is not necessary.
[0018] In this embodiment, learning is performed to increase the likelihood of normal data of the detection target device and regard it as normal data, and to decrease the likelihood of normal data of devices of the same concept that have a similar normal distribution to the detection target device and regard it as abnormal data, thereby obtaining a model that gives a high likelihood only to the data of the detection target device and regards it as normal. Therefore, compared to the conventional technology that performs learning using only the data of the detection target device, it is possible to accurately estimate the normal distribution of the data of the detection target device, and sufficient detection accuracy can be obtained even when the data of the detection target device is small. EXAMPLES
[0019] The first embodiment will be described with reference to Figures 1 to 12. Figure 1 is an explanatory diagram showing an overall outline of the present embodiment.
[0020] The abnormality degree calculation system 1 (see FIG. 6) calculates the abnormality degree of multiple devices M that exhibit measurable physical changes in response to operation. Examples of physical changes that can be measured in response to operation include sound, vibration (sound and vibration can be collectively referred to as a signal associated with vibration), temperature, and color that accompany the operation of the device, such as an electric motor, hydraulic motor, pneumatic cylinder, hydraulic cylinder, solenoid, linear actuator, etc. Here, detection of an abnormality by sound will be described as an example, but abnormalities can also be detected from physical changes other than sound.
[0021] The anomaly degree calculation system shown in Fig. 1 calculates an anomaly degree for a device selected from multiple devices M1 to Mn that belong to a common superordinate concept C2. Each device M1 to Mn has a different subordinate concept C1 that belongs to the superordinate concept C2. For example, the superordinate concept C2 is a model, and the subordinate concepts C1 are different models that belong to the same model. To show that the subordinate concept C1 differs for each device, Fig. 1 shows the subordinate concepts C11 to C13. If the model is common, differences in serial numbers are not an issue.
[0022] The superordinate concept C2 and subordinate concept C1 can also be expressed as follows. For example, the smallest unit for classifying device M is subordinate concept C1, and the concept immediately above to which subordinate concept C1 belongs is superordinate concept C2. Here, the device to be detected is device M1, and other devices M2 to Mn are called devices of the same level concept. When devices M1 to Mn are not particularly distinguished, they are called device M or device.
[0023] The input D0 includes sensor data D2 acquired from n devices M that are the same in terms of model, which is an example of a higher-level concept, and a model number D1, which is an example of a lower-level concept. Hereinafter, the model number D1 may be abbreviated to model D1.
[0024] The concept type assigning unit 11 converts the model number D1 into a concept type D3. The feature vector extracting unit 12 outputs a feature vector D4 from the sensor data D2. Hereinafter, the feature vector D4 may be abbreviated as feature D4. The training database DB1 stores the sensor data D2 and the concept type D3. The training database DB1 may also be called a training digital input signal database DB1.
[0025] The likelihood calculation unit 13 outputs a likelihood D5 of the sensor data of each device M from the feature amount D4 using a machine learning model provided from the learning database DB2.
[0026] The loss calculation unit 14 uses the concept type D3 and the likelihood D5 to calculate a loss function such that the likelihood of the device M1 whose concept type is the detection target is high and the likelihood of the devices M2 to Mn whose concept types are the same level is low, and outputs the loss D6.
[0027] The model update unit 15 uses the loss D6 to update parameters of the machine learning model stored in the learning database DB2. Hereinafter, the machine learning model may be abbreviated to "model."
[0028] When an abnormality is detected, the relearning necessity determination unit 16 determines whether relearning is necessary using the likelihood D5 and the concept type D3.
[0029] When it is determined that re-learning is not necessary, the abnormality degree calculation unit 17 calculates the abnormality degree of the detection object by using the likelihood D7 of the detection object out of the likelihood D5.
[0030] The relationship between the input D0, model number D1, sensor data D2, and concept type D3 will be described with reference to Fig. 2. The input D0 obtained from each device M includes a model number D1 and sensor data D2. The model number D1 is converted to a concept type D3 by the concept type assigning unit 11. For example, the concept type D3 is set to "1" for the detection target device M1, and "0" is set for other devices M2 to Mn of the same concept as the detection target.
[0031] A method for calculating the likelihood and loss using input from the training database DB1 is shown in Fig. 3. First, a batch size of N combinations of sensor data D2 and concept types D3 are obtained from the training database DB1.
[0032] Here, for example, N / 2 combinations of sensor data D2 from the detection target device and concept type D3 are acquired, and for example, N / 2 combinations from devices with the same concept as the detection target are acquired. When there are multiple other devices with the same concept as the detection target device, if data of a batch size of N pieces is randomly acquired from a collection of data of the detection target device and data of other devices with the same concept, the number of data from the detection target device will be smaller than the number of data from other devices with the same concept, and learning of the machine learning model will be slow.
[0033] Next, the feature vector extraction unit obtains N feature quantities D4 by extracting a feature vector from the sensor data D2. The likelihood calculation unit 13 obtains N likelihood quantities D5 from the feature quantities D4 by calculating the likelihood using a machine learning model.
[0034] The machine learning model can be any model that allows the model parameters to be updated using a loss defined as a function of likelihood. For example, Normalizing Flow (NF), Gaussian Mixture Model (GMM), or Mahalanobis can be used.
[0035] The N likelihoods D5 and the concept types D3 corresponding to each likelihood D5 are input to the loss calculation unit 14. The loss calculation unit 14 calculates and outputs the loss by using a loss function. The loss function is given by, for example, the following equation 1.
[0036]
number
[0037] Ntarget is the number N / 2 of combinations of likelihood D5 and concept type D3 from the device to be detected included in the batch. Noutlier is the number of samples that satisfy the conditions within the indicator function I among the number N / 2 of combinations of likelihood D5 and concept type D3 from other devices of the same concept as the detection target included in the batch. NLL(x) is the negative log-likelihood of feature D4. k is a hyperparameter where 0 ≦ k ≦ 1, and by adjusting k, the ratio between the effect of increasing the likelihood of the data of the device to be detected in the first term of Equation 1 and the effect of decreasing the likelihood of the data of the device to be detected in the second term can be adjusted. By decreasing k, the effect of increasing the likelihood of the data of the device to be detected in the first term can be enhanced.
[0038] When using, for example, GMM as the machine learning model, it is necessary to make k smaller than 1 (k < 1). The indicator function I[NLL(x) < c] takes 1 when NLL(x) < c and 0 when NLL(x) ≥ c. c is a threshold value and is determined, for example, by the method shown below. Both the feature from the device to be detected and the feature from the device of the same concept as the detection target are given to the likelihood calculation unit, and the loss function is given as in Equation 2 below.
[0039]
Equation
[0040] The value of NLL(x) when the model is updated several times to minimize this loss function is used as the threshold value c.
[0041] Figure 4 is a graph showing the transition of the negative log-likelihood of each data when the model is updated using Equation 1. The vertical axis of Figure 4 represents NLL(x), and the horizontal axis represents the number of model updates. Above the graph of Figure 4, the change in the negative log-likelihood of the data of other devices of the same concept as the detection target is shown. Below the graph of Figure 4, the change in the negative log-likelihood of the data of the device to be detected is shown.
[0042] By using the formula 1, it is possible to add a constraint that NLL(x) of the data of the detection target device is smaller than NLL(x) of the data of other devices of the same conceptual order as the detection target. Since the anomaly degree calculation unit 17 calculates the anomaly degree using the negative log likelihood as the anomaly degree, this constraint corresponds to the model learning to lower the anomaly degree of the data of the detection target device and to raise the anomaly degree of the data of other devices of the same conceptual order as the detection target.
[0043] Furthermore, by using the indicator function I, it is possible to prevent the NLL(x) of the data of the detection target device and other devices of the same conceptual order from diverging.
[0044] A method in which the relearning necessity determination unit 16 determines whether relearning is necessary will be described with reference to Fig. 5. The vertical axis of Fig. 5 indicates NLL(x), and the horizontal axis of Fig. 5 indicates time. The upper side of the graph in Fig. 5 shows the negative log-likelihood of data of the detection target device and other devices of the same concept. The lower side of the graph in Fig. 5 shows the negative log-likelihood of data of the detection target device.
[0045] When an anomaly detection system is operated using the model created during model training, the distribution of data from the detection target device and data from other devices of the same conceptual level as the detection target device transitions over time from the time of training.
[0046] For this reason, as shown on the left side of the graph in Figure 5 at the time of "re-learning execution," the negative log-likelihood of the data from the detection target device will increase, while the negative log-likelihood of the data from other devices of the same conceptual level as the detection target will decrease.
[0047] If the negative log-likelihood of the data of the detection target device becomes large, there is a possibility that an abnormality has occurred in the detection target device, so re-learning is performed only when the negative log-likelihood of the data of other devices of the same conceptual order as the detection target device becomes small. For example, the user sets a re-learning determination threshold Thc, and re-learning is performed when the negative log-likelihood of the data of the detection target device and other devices of the same conceptual order falls below the re-learning determination threshold c.
[0048] Fig. 6 shows the hardware configuration and software configuration of the anomaly degree calculation system. By making a computer execute a specific computer program, the computer can be used as the anomaly degree calculation device 100. Fig. 6 shows an example in which the anomaly degree calculation device 100 is configured from one computer, but instead, one or more anomaly degree calculation devices 100 may be formed from multiple computers. The computer may be a virtual computer.
[0049] The abnormality degree calculation device 100 is connected to n sensor terminals T attached to n devices, respectively, via a communication network CN.
[0050] The sensor terminal T is configured, for example, as a portable recording terminal.
[0051] The abnormality degree calculation device 100 is a computer including, for example, a calculation unit 1001, a main memory device 1002, an auxiliary memory device 1003, an input unit 1004, an output unit 1005, and a communication unit 1006.
[0052] The calculation unit 1001 includes one or more microprocessors, and realizes functions such as the concept classification unit 11, feature vector extraction unit 12, likelihood calculation unit 13, loss calculation unit 14, model update unit 15, re-learning necessity determination unit 16, and anomaly calculation unit 17 described in FIG. 1 by reading a predetermined computer program stored in the auxiliary storage unit 1003 into the main storage unit 1002 and executing it.
[0053] The input unit 1004 may include, for example, a keyboard, a touch panel, a pointing device, etc., and accepts input from a user who uses the anomaly degree calculation device 100. The output unit 1005 may include, for example, a monitor display, a speaker, a printer, etc., and provides information to the user.
[0054] The communication unit 1006 communicates with the sensor terminal T via the communication network CN. The communication unit 1006 can also communicate with other computers (not shown).
[0055] The storage medium MM is, for example, a storage medium such as a flash memory or a hard disk, and transfers and stores computer programs or data to the abnormality degree calculation device 100, and reads and stores computer programs or data from the abnormality degree calculation device 100. The storage medium MM may be directly connected to the abnormality degree calculation device 100, or may be connected to the abnormality degree calculation device 100 via the communication network CN.
[0056] The following describes the configuration of the sensor terminal T. The sensor terminal T includes, for example, a sensor unit 21, a control unit 22, a storage unit 23, and a communication unit 24.
[0057] The sensor unit 21 is a device that acquires sensor data from devices such as a microphone, an acceleration sensor, a magnetic sensor, a camera, etc. The data acquired by the sensor unit 21 is stored in the storage unit 23. The control unit 22 that controls the sensor terminal T transmits the sound data stored in the storage unit 23 to the abnormality degree calculation device 100. Then, the abnormality degree calculation device 100 can calculate the degree of abnormality based on the received sensor data.
[0058] Fig. 7 is a block diagram of the feature quantity vector extraction unit 12. Fig. 8 is a block diagram of the anomaly degree calculation device 100 during learning, and Fig. 9 shows a process flow within the anomaly degree calculation device 100 during learning.
[0059] The configuration of the feature vector extraction unit 12 will be described. Although the sensor unit 21 of the sensor terminal T is a microphone, it is not limited to a microphone and may be any device capable of acquiring sensor data from the device M, such as an acceleration sensor, a magnetic sensor, or a camera. When a sensor other than a microphone is used, a known feature vector extraction method is used. When the sensor unit 21 is another sensor, a known feature vector extraction method according to the type of sensor is used.
[0060] The input sound acquisition unit 101 converts an analog input signal input from a microphone into a digital input signal by an A / D (analog / digital) converter (S101), and stores the digital input signal in a training digital input signal database DB1 (S102).
[0061] The frame division unit 102 divides the digital input signal extracted from the training digital input signal database DB1 into a specified number of time points (hereinafter, frame size), and outputs the frame signals (S104). The frames may overlap.
[0062] The window function multiplier 103 multiplies the input frame signal by a window function and outputs a window function multiplied signal (S105). For example, a Hanning window is used as the window function.
[0063] The frequency domain signal calculation unit 104 outputs a frequency domain signal by performing a short-time Fourier transform on the input window function multiplied signal (S106). If the frame size is N, the frequency domain signal is a set of M complex numbers, with one complex number corresponding to each of (N / 2+1)=M frequency bins. The frequency domain signal calculation unit 104 can also use a frequency transform method such as a constant Q transform (CQT) instead of the short-time Fourier transform.
[0064] The power spectrogram calculation unit 105 outputs a power spectrogram based on the input frequency domain signal (S107). The filter bank multiplication unit 106 multiplies the input power spectrogram by a Mel filter bank to output a Mel power spectrogram (S108). The filter bank multiplication unit 106 may use a filter bank such as a 1 / 3 octave band filter instead of the Mel filter bank.
[0065] The instantaneous feature amount calculation unit 107 applies a logarithm to the input Mel power spectrogram to output a logarithmic Mel power spectrogram (S109). Note that instead of the logarithmic Mel power spectrogram, Mel frequency cepstral coefficients (MFCCs) may be calculated. In that case, instead of the filter bank multiplication unit 106 and the logarithm calculation unit 107, the logarithmic value of the power spectrogram is calculated, multiplied by a filter bank, discrete cosine transform is performed, and the MFCCs are output.
[0066] The feature time series calculation unit 108 outputs the feature D4 by concatenating L frames for the input logarithmic mel-power spectrogram or MFCC (S110). Instead of the logarithmic mel-power spectrogram or MFCC, a time series (delta) of their time difference or time derivative may be input, and the feature D4 may be output by concatenating L frames for the input.
[0067] The feature vector extraction unit 12 may input a time difference or a time series of time differentials (delta-delta) to concatenate L adjacent frames and output the feature vector D4. Furthermore, any combination of these may be selected and concatenated in the feature axis direction, and the feature vector D4 may be concatenated with L adjacent frames to output the feature vector D4. The feature vector extraction unit 12 generates the feature vector D4 and inputs it to the likelihood calculation unit 13.
[0068] The likelihood calculation unit 13 calculates a likelihood D5 based on the feature amount D4. The calculated likelihood D5 is input to the loss calculation unit 14 (S111).
[0069] The loss calculation unit 14 calculates a value D6 of a predetermined loss function defined as the likelihood of the data of the detection target device and the likelihood of the data of other devices of the same concept as the detection target device, based on the likelihood D5 and the concept type D3. The loss D6 calculated by the loss calculation unit 14 is input to the model update unit 15 (S112).
[0070] The model update unit 15 repeatedly learns the parameters of the machine learning model so as to minimize the value of the loss D6 (S113 to S116). The parameters of the machine learning model are stored in the learning database DB2 (S117).
[0071] That is, the model update unit 15 determines whether the convergence condition is satisfied or whether the number of iterations C1 of this process exceeds the upper limit value ThC (S113).
[0072] If the convergence condition is not satisfied or if the number of iterations C1 is equal to or less than the upper limit ThC, the model update unit 15 updates the parameters of the machine learning model (S114), calculates the convergence condition (S115), increments the number of iterations C1 by one, and returns to step S112.
[0073] The model update unit 15 stores the parameters of the machine learning model in the learning database DB2 (S117).
[0074] Fig. 10 is a block configuration diagram of the abnormality degree calculation device 100 when calculating the abnormality degree. Fig. 11 shows a processing flow within the abnormality degree calculation device 100 when calculating the abnormality degree. Description will be made with reference to Figs. 10 and 11.
[0075] The anomaly degree calculation device 100 has a relearning necessity determination unit 16 and an anomaly degree calculation unit 17 instead of the loss calculation unit 14 and the model update unit 15 .
[0076] When calculating the degree of abnormality of the feature amount D4 extracted from the sensor data D2, the likelihood calculation unit 13 reads parameters from the learning database DB2 (S201).
[0077] The contents of steps S101 to S110 have already been described, so a duplicated description will be omitted. In these steps S101 to S110, a feature amount D4 is generated for the sensor data D2 acquired from the sensor terminal T, and input to the machine learning model of the likelihood calculation unit 13. In the following description, duplicated descriptions will also be omitted. In S102, the input sound is stored in the training database DB1 even during anomaly calculation, but this is for the purpose of storing training data for re-learning.
[0078] The likelihood calculation unit 13 calculates a likelihood D5 from the feature amount D4, and inputs the calculated likelihood D5 to the re-learning necessity determination unit 16 (S111).
[0079] Using the likelihood D5 and the concept type D3, the re-learning necessity determination unit 16 performs re-learning if the negative logarithmic likelihood of the detection target device and other devices of the same concept exceeds a certain re-learning necessity determination threshold Thc, and inputs the likelihood D7 of the data of the detection target device to the anomaly degree calculation unit 17 if it falls below the threshold (S202).
[0080] The abnormality degree calculation unit 17 calculates the abnormality degree of the sensor data of the detection target device based on the likelihood D7 of the data of the detection target device (S203).
[0081] Here, a configuration will be disclosed in which, for example, Normalizing Flow (NF) is used as the machine learning model.
[0082] The likelihood calculation unit 13 is composed of a multilayer neural network. The input of the likelihood calculation unit 13 consists of elements whose number is equal to the number of dimensions of the feature vector, and each of these elements accepts each element of the feature vector. The input layer is connected to a predetermined transformation function, a multilayer neural network having the same number of elements in the input and output, and a nonlinear activation function (e.g., ReLU function), and the final layer of the multilayer neural network has the same number of elements as the input layer. In this case, the predetermined transformation function is determined so that the transformation from the input layer to the output of the multilayer neural network is composed of a transformation that can be reversed. The likelihood calculation unit 13 calculates the likelihood from the output of the final layer of the multilayer neural network, assuming that the distribution function of the output of this final layer of the multilayer neural network follows a known distribution (e.g., normal distribution). Furthermore, the likelihood D4 of the input feature vector is calculated using a known formula for random variable transformation.
[0083] The loss calculation unit 14 calculates the loss D6 using a loss function defined as a function of the likelihood D5, for example, using Equation 1. At this time, the loss function is defined, for example, as in Equation 1, so as to make the likelihood D5 of the data of the detection target device high and the likelihood D5 of the data of other devices of the same conceptual level as the detection target device low.
[0084] The model update unit 15 updates the parameters of the machine learning model so as to minimize the loss D6 calculated by the loss calculation unit 14. The minimization can be performed by using a known optimization algorithm such as SGD, Momentum SGD, AdaGrad, RMSprop, AdaDelta, or Adam.
[0085] According to this embodiment, which is configured in this manner, learning can be performed so that the likelihood of normal data from the device to be detected is increased and regarded as normal data, and the likelihood of normal data from other devices of the same conceptual level as the device to be detected is decreased and regarded as abnormal data.
[0086] Fig. 12 shows the effect of this embodiment. The upper side of Fig. 12 shows a case where the machine learning model learns using only normal data of the detection target device. The lower side of Fig. 12 shows a case where the machine learning model learns by regarding normal data of a device of the same conceptual level as the detection target device as abnormal. The thin shaded area indicates the area that the machine learning model of this embodiment considers to be normal data.
[0087] As shown in Figure 12 (1), when learning is performed using only the normal data of the detection target device, the normal data of the detection target device and the normal data of other devices of the same conceptual level as the detection target device have similar normal distributions, so the trained model regards part of the normal data of other devices of the same conceptual level as the detection target device as normal data and gives a high likelihood.
[0088] However, in this embodiment, the trained model regards normal data of other devices of the same conceptual level as the detection target device as abnormal data and assigns a low likelihood. Therefore, according to this embodiment, as shown in Fig. 12(2), it is possible to reduce the area that the machine learning model regards as normal, and to more accurately estimate the normal distribution of the data of the detection target device.
[0089] Therefore, according to this embodiment, by using normal data of other devices of the same conceptual level as the device to be detected, it becomes possible to more accurately estimate the normal distribution of the data of the device to be detected, and sufficient detection accuracy can be obtained even when there is little data on the device to be detected. EXAMPLES
[0090] The second embodiment will be described with reference to Figures 13 and 14. In the following embodiments including this embodiment, differences from the first embodiment will be mainly described.
[0091] Fig. 13 is a block diagram showing the configuration when performing the sequential learning process, and Fig. 14 is a process flow of the sequential learning process.
[0092] In this embodiment, first, a machine learning model is trained by batch learning similar to the learning method described in the first embodiment with reference to Figures 7 and 8. This embodiment differs from the first embodiment in that, while the trained model is used to calculate an anomaly degree, the trained model is updated using data obtained during operation, and anomaly degree calculation is also performed from the updated model.
[0093] An input D0 from a sensor terminal T is divided into a model number D1 and sensor data D2, the model number D1 is converted into a concept type D3 by a concept type assigning unit 11, and the sensor data D2 is converted into a feature D4 by a feature vector extracting unit 12. The sensor data D2 and the concept type D3 are stored in a data storage database DB3 for identifying the cause of an abnormality when an abnormality occurs.
[0094] In this embodiment, there are two types of likelihood calculation units. One is a learning model likelihood calculation unit 21 that performs likelihood calculation of the feature amount D4 using a model obtained from a learning database that is successively updated. The other is a trained model likelihood calculation unit 22 that performs likelihood calculation of the feature amount D4 using a model obtained from a trained database DB4 that stores trained models.
[0095] Using a predetermined loss function defined as a function of likelihood, the loss calculation unit 14 calculates a loss value from the likelihood calculated by the learning model likelihood calculation unit 21. The predetermined loss function is a loss function that gives a high likelihood to data of the detection target device and a low likelihood to data of other devices that are conceptually of the same order as the detection target device.
[0096] Note that in the sequential learning process used in this embodiment, since the sequential model is updated using sequentially input sensor data, unlike during batch learning, learning cannot be performed using batches. Therefore, out of the n sensor data input from n devices through the sensor terminal T, one piece of data of the device to be detected and one piece of data of another device having the same concept as the device to be detected are extracted, and the loss is calculated from their likelihoods. For example, when using Equation 1, Ntarget is 1, Noutlier becomes 1 when NLL(x) < c, and Equation 1 becomes only the first term when NLL(x) ≥ c.
[0097] The model update unit 15 updates the parameters based on the value of the loss and stores the obtained parameters in the learning database DB2. For minimizing the loss function, a known optimization algorithm for performing sequential parameter update, such as Online Gradient Descent (OGD), can be used.
[0098] The abnormality degree calculation unit 17 calculates the abnormality degree from the likelihoods calculated by the learning model likelihood calculation unit 21 and the learned model likelihood calculation unit 22, respectively. When using only the likelihood from the sequentially updated model, it is difficult to deal with abnormalities that occur over a long period of time such as aging deterioration. Therefore, the likelihood from the learned model at the start of operation is added to deal with the above abnormalities. As a method of calculating the abnormality degree using both the likelihood obtained from the learned model and the likelihood obtained from the sequentially updated model, for example, a function such as Equation 3 can be used.
[0099]
Equation
[0100] A is the final abnormality degree. A1 is the abnormality degree calculated using the model obtained from the learning database. A2 is the abnormality degree calculated using the model obtained from the learned database. α is a parameter determined during operation.
[0101] The present embodiment thus configured also has the same operational effects as those of Example 1. Furthermore, in this embodiment, when the degree of anomaly is calculated, a model is sequentially learned while the degree of anomaly is being calculated, so that a model that is robust against changes in the normal distribution of each device can be learned.
[0102] In this embodiment, the learning database DB2 at the start of operation is used as the learned database DB4, and the likelihood given by the model at the start of operation is also calculated at the same time and reflected in the final degree of anomaly, making it possible to detect anomalies that occur over a long period of time, such as aging deterioration.
[0103] Furthermore, this embodiment can also handle cases where the input data of the same concept changes, such as when another device of the same concept is newly added. EXAMPLES
[0104] A third embodiment will be described with reference to Figures 15 to 18. In the first and second embodiments, all devices that are the same as the detection target device in terms of a higher concept, for example, model, and different from the detection target in terms of a lower concept, for example, model, are treated as other devices of the same level as the detection target device.
[0105] On the other hand, among other devices of the same concept, there are cases where multiple types of subordinate concepts, for example, models, are included, and the similarity with the data of the detection target device varies for each type. For example, there may be a case where a device has the same model as the detection target device, but a different model, but the same or a different driving method. In this embodiment, the purpose is to further improve the detection accuracy by utilizing the types of such subordinate concepts.
[0106] The learning process of this embodiment will be described with reference to Figures 15 and 16. Of the input D0 from the sensor terminal T, the model number D1 is converted into a concept type D3 by the concept type assigning unit 11. The sensor data D2 is converted into a feature D4 by the feature vector extracting unit 12. Here, unlike the first and second embodiments, the concept type assigning unit 11 outputs both the concept type D3 and the model number D1.
[0107] The sensor data D2 is converted into a feature D4 by the feature vector extraction unit 12. When a set of the detection target device and other devices of the same concept includes a plurality of models, the same-level concept classification assignment unit 31 examines the relationship between each model of the other devices of the same concept and the model of the detection target device, and assigns a weight w to each model of the other devices of the same concept. This weight w may be set, for example, based on the commonality of the elements related to the source of the sound emitted from the detection target device, that is, the sound source. For example, the sensor data differs between the case where the drive method of the device is an electric motor and the case where the drive method of the device is a hydraulic motor. Therefore, in this embodiment, a high value is assigned to the model of the other devices of the same drive type as the detection target device, as the similarity to the detection target device is high, and a low value is assigned to the model of the other devices of a different drive type from the detection target device, as the similarity to the detection target device is low.
[0108] In addition to the sensor data D2 and the concept type D3, the training database DB1 stores weights w corresponding to the respective models of other devices of the same concept.
[0109] The likelihood calculation unit 13 calculates a likelihood D5 from the feature quantity D4 calculated by the feature quantity vector extraction unit 12 using a model obtained from the learning database DB2.
[0110] The loss calculation unit 14 calculates a loss D6 from the likelihood calculated by the likelihood calculation unit 13 using a loss function that is defined as a function of likelihood and that assigns a high likelihood to data of the detection target device and a low likelihood to data of other devices that are conceptually equivalent to the detection target device. The loss function used here can be, for example, Equation 4, which is a modification of Equation 1.
[0111]
number
[0112] Here, wi is the weight of the i-th type determined in the same-level concept classification unit 31.
[0113] The model update unit 15 updates the parameters of the machine learning model so as to minimize the loss D6 calculated by the loss calculation unit 14. The minimization can be performed by using a known optimization algorithm such as SGD, Momentum SGD, AdaGrad, RMSprop, AdaDelta, or Adam.
[0114] The process of calculating the degree of anomaly in this embodiment will be described with reference to Fig. 17 and Fig. 18. A concept type D3 is assigned to a model number D1 among inputs D0 from a sensor terminal T by a concept type assigning unit 11. The sensor data D2 is converted into a feature D4 by a feature vector extracting unit 12.
[0115] When a set of the detection target device and other devices of the same concept include multiple models, the same-level concept type assignment unit 31 examines the relationship between each model of the other devices of the same concept and the model of the detection target, and assigns a weight w to each model of the other devices of the same concept.
[0116] In the training database DB1, in addition to the sensor data D2 and the concept type D3, weights w corresponding to each model of other devices of the same concept are stored in order to prepare for re-learning.
[0117] The likelihood calculation unit 13 calculates a likelihood D5 from the feature quantity D4 calculated by the feature quantity vector extraction unit 12 using a model obtained from the learning database DB2.
[0118] The re-learning necessity determination unit 16 uses the likelihood D5 and the concept type D3 to re-learn the model if the negative log-likelihood of the detection target device and other devices of the same concept exceeds a certain re-learning necessity determination threshold Thc, and inputs the likelihood D5 of the data of the detection target device to the anomaly calculation unit 17 if the negative log-likelihood falls below the threshold Thc (S202).
[0119] The abnormality degree calculation unit 17 calculates the abnormality degree of the sensor data of the detection target device based on the likelihood D5 of the data of the detection target device (S203).
[0120] The present embodiment thus configured also has the same effect as the first embodiment. Furthermore, in this embodiment, a model information database DB5 that holds model information and a same-level concept classification unit 31 that assigns weights according to the models of other devices of the same concept are added. As a result, according to this embodiment, the influence of the data of other devices of each same concept on the value of the loss D6 can be adjusted for each model of other devices of each same concept. Therefore, in this embodiment, for example, a large weight is assigned to devices of the same model as the detection target device and a small weight is assigned to models of different models, so that a model can be learned to lower the likelihood of data that is more similar to the detection target device.
[0121] The present invention is not limited to the above-described embodiments, and includes various modified examples. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to those including all of the configurations described. It is also possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. It is also possible to add, delete, or replace a part of the configuration of each embodiment with another configuration.
[0122] The components of the present invention may be selected arbitrarily, and the invention having the selected components is also included in the present invention. Furthermore, the components described in the claims may be combined in combinations other than those explicitly stated in the claims. (Appendix 1) An abnormality degree calculation system for calculating an abnormality degree of a target device, a concept classification assigning unit that assigns a predetermined concept classification based on the identification number of the target device; a feature vector extraction unit that extracts a feature vector based on sensor data obtained from a sensor corresponding to the target device; a likelihood calculation unit that calculates the likelihood of the feature vector by using a machine learning model obtained from a learning database; a loss calculation unit that calculates a loss using a loss function defined as a function of the likelihood calculated by the likelihood calculation unit; a model update unit that updates the machine learning model using the loss calculated by the loss calculation unit and a trained machine learning model; a re-learning necessity determination unit that determines whether re-learning is necessary based on the likelihood calculated by the likelihood calculation unit; an abnormality degree calculation unit that calculates an abnormality degree when the relearning necessity determination unit determines that relearning is unnecessary; Equipped Anomaly calculation system. (Appendix 2) The predetermined concept type is a concept type obtained from the identification number, and is a concept of the same level that belongs to a common higher level concept with the target device, The loss calculation unit calculates a loss to be fed back to the machine learning model so as to consider the sensor data of the target device as normal and give a high likelihood, and to consider sensor data from other devices of the same concept as the target device as abnormal and give a low likelihood. 2. An anomaly calculation system according to claim 1. (Appendix 3) The target device and the other devices of the same concept have a common model, which is the smallest unit for classifying devices, and the models included in the model are different from each other. 3. An anomaly calculation system according to claim 2. (Appendix 4) The identification number is information assigned according to the model. 2. An anomaly calculation system according to claim 1. (Appendix 5) The concept type is set to "1" for the target device and "0" for the other devices. 3. An anomaly calculation system according to claim 2. (Appendix 6) A plurality of devices that exhibit a measurable physical change in response to their operation, the plurality of devices belonging to a common model but different models being pre-grouped; One of the target devices is selected from the grouped devices; Among the plurality of grouped devices, at least one device other than the target device is selected as another device. 3. An anomaly calculation system according to claim 2. (Appendix 7) The re-learning necessity determination unit determines whether or not re-learning is necessary during operation by using a likelihood of data of other devices of the same concept. 3. An anomaly calculation system according to claim 2. (Appendix 8) Furthermore, in order to perform model learning not only during learning but also during operation, a learning model likelihood calculation unit that calculates the likelihood of a learning model and a trained model likelihood calculation unit that calculates the likelihood of a trained model are provided. 2. An anomaly calculation system according to claim 1. (Appendix 9) the learning model likelihood calculation unit calculates a likelihood by reading a model from a learning database that stores models that are successively updated during operation; The trained model likelihood calculation unit calculates the likelihood by reading a model from a trained database that stores a model at the start of operation. 9. An anomaly calculation system according to claim 8. (Appendix 10) a same-level concept classification assigning unit that assigns a weight to the other devices in accordance with a relationship between a lower level concept of the target device and a lower level concept of each of the other devices; 3. An anomaly calculation system according to claim 2. (Appendix 11) the peer concept classification unit assigns a weight to each of the other devices and adjusts the influence of the likelihood value of each of the other devices on the loss value by using the weight in a loss function; 11. An anomaly calculation system according to claim 10. (Appendix 12) The target device is a device that generates sound or vibration in response to its operation. 2. An anomaly calculation system according to claim 1. (Appendix 13) An anomaly degree calculation system that calculates an anomaly degree of a target device by a computer, By the computer, assigning a predetermined concept type to the target device based on the identification number of the target device; extracting a feature vector based on sensor data obtained from a sensor corresponding to the target device; Calculating the likelihood of the feature vector using a machine learning model obtained from a learning database; Calculating a loss using a loss function defined as a function of the calculated likelihood; updating the machine learning model using the calculated loss and the trained machine learning model; When an abnormality is detected in the target device, whether or not re-learning is required is determined based on the calculated likelihood; If it is determined that re-learning is not necessary, the degree of anomaly is calculated. Anomaly degree calculation method. [Explanation of symbols]
[0123] 1: Anomaly degree calculation system, 11: Concept classification assignment unit, 12: Feature vector extraction unit, 13: Likelihood calculation unit, 14: Loss calculation unit, 15: Model update unit, 16: Relearning necessity determination unit, 17: Anomaly calculation unit, 21: Learning model likelihood calculation unit, 22: Learned model likelihood calculation unit, 31: Same-level concept classification assignment unit, M: Equipment
Claims
1. An abnormality degree calculation system for calculating an abnormality degree of a target device, a concept type assigning unit that assigns a predetermined concept type based on an identification number of the device; a feature vector extraction unit that extracts a feature vector based on sensor data obtained from a sensor corresponding to the device; a likelihood calculation unit that calculates the likelihood of the feature vector by using a machine learning model obtained from a learning database; a loss calculation unit that calculates a loss using a loss function defined as a function of the likelihood calculated by the likelihood calculation unit; a model update unit that updates the machine learning model using the loss calculated by the loss calculation unit and a trained machine learning model; a re-learning necessity determination unit that determines whether re-learning is necessary based on the likelihood calculated by the likelihood calculation unit; an abnormality degree calculation unit that calculates an abnormality degree when the relearning necessity determination unit determines that relearning is unnecessary; Equipped with the concept type assigning unit, when assigning a concept type based on an identification number of the device, assigns different concept types depending on whether a normal distribution of the sensor data of the device is the same as or different from a normal distribution of the sensor data of the target device that is a target for calculating an anomaly degree, the loss calculation unit calculates the value of the loss using the loss function that increases the value of the loss as the value of the negative logarithmic likelihood based on the likelihood derived from sensor data of the device assigned the same concept type as the concept type of the target device increases, and increases the value of the loss as the value of the negative logarithmic likelihood based on the likelihood derived from sensor data of the device assigned a concept type different from the concept type of the target device decreases, The model update unit updates the machine learning model so as to reduce the value of the loss calculated by the loss calculation unit, The re-learning necessity determination unit determines that re-learning of the machine learning model is necessary when a value of the negative logarithmic likelihood based on the likelihood derived from sensor data of another device of the same concept belonging to a common higher concept with the target device, the other device being the device to which a concept type different from that of the target device is assigned, falls below a re-learning determination threshold. Anomaly calculation system.
2. One model may include multiple models, The model of the target device is common to the model of other devices of the same concept, The type of the target device and the type of other devices of the same concept are different from each other, The abnormality degree calculation system according to claim 1 .
3. The identification number is information assigned according to the model. The abnormality degree calculation system according to claim 1 .
4. The concept type is set to "1" for the target device and to "0" for the other devices. The abnormality degree calculation system according to claim 1 .
5. A plurality of devices that exhibit a measurable physical change in response to their operation, the plurality of devices belonging to a common model but different models being pre-grouped, One of the target devices is selected from the grouped devices; Among the plurality of grouped devices, at least one device other than the target device is selected as another device. The abnormality degree calculation system according to claim 1 .
6. The anomaly degree calculation system includes a trained model likelihood calculation unit and a training model likelihood calculation unit which is the likelihood calculation unit, the learning model likelihood calculation unit calculates the likelihood of the feature vector by using the machine learning model obtained from the learning database not only during learning but also during operation, The trained model likelihood calculation unit calculates the likelihood of the feature vector by using a trained machine learning model obtained from a trained database during operation. The abnormality degree calculation system according to claim 1 .
7. the learning model likelihood calculation unit calculates a likelihood by reading the machine learning model from the learning database that stores the machine learning model that is successively updated during operation; The trained model likelihood calculation unit calculates the likelihood by reading the trained machine learning model from the trained database that stores the trained machine learning model at the start of operation. The abnormality degree calculation system according to claim 6 .
8. The anomaly degree calculation system further includes a same-level concept classification assigning unit that assigns a weight to the other devices in accordance with a similarity between a subordinate concept of the target device and a subordinate concept of each of the other devices, the loss calculation unit calculates the value of the loss by using the loss function including a product obtained by multiplying the value of the negative log-likelihood based on the likelihood derived from the sensor data of the other device by the weight set for the other device. The abnormality degree calculation system according to claim 1 .
9. The target device is a device that generates sound or vibration in response to its operation. The abnormality degree calculation system according to claim 1 .
10. An abnormality degree calculation method in which a computer calculates an abnormality degree of a target device, comprising: The computer, A predetermined concept type is assigned based on the device identification number, extracting a feature vector based on sensor data obtained from a sensor corresponding to the device; Calculating the likelihood of the feature vector using a machine learning model obtained from a training database; Calculating a loss using a loss function defined as a function of the calculated likelihood; updating the machine learning model using the calculated loss and the trained machine learning model; When detecting an abnormality in the target device, determining whether or not re-learning is required based on the calculated likelihood; A method for calculating an anomaly degree when it is determined that re-learning is not necessary, when the computer assigns a concept type based on the identification number of the device, different concept types are assigned depending on whether a normal distribution of the sensor data of the device is the same as or different from a normal distribution of the sensor data of the target device that is a target for calculating an anomaly degree; the computer calculates the loss using the loss function such that, when calculating the loss using the loss function, the greater the value of the negative log-likelihood based on the likelihood derived from sensor data of the device to which the same concept type as the concept type of the target device is assigned, the greater the value of the loss, and the smaller the value of the negative log-likelihood based on the likelihood derived from sensor data of the device to which a concept type different from the concept type of the target device is assigned, the greater the value of the loss; When updating the machine learning model, the computer updates the machine learning model so as to reduce the calculated value of the loss; When the computer determines whether or not re-learning is necessary, it determines that re-learning of the machine learning model is necessary if the value of the negative log-likelihood based on the likelihood derived from sensor data of another device of the same concept belonging to a common higher concept to the target device, which is the device to which a concept type different from that of the target device is assigned, falls below a re-learning determination threshold. Abnormality degree calculation method.
Citation Information
Patent Citations
Estimation of current and future machine states
US20190311273A1
Abnormal sound detection device, abnormality degree calculation device, abnormal sound generation device, abnormal sound detection learning device, abnormal signal detection device, abnormal signal detection learning device, and methods and programs therefor
WO2018150616A1
Abnormality detection device, probability distribution learning device, self-encoder learning device, data conversion device, and program
WO2020031570A1
Monitoring apparatus, monitoring system, and abnormality detection method
JP2019008354A
Cited By
Abnormality management device and abnormality management method
JP7755771B1
