Information processing system, information processing method, and recording medium

Abstract

An information processing system (10) comprises: a score acquisition means (110) that acquires a first score based on the output of a first inference model and a second score based on the out put of a second inference model; a first integrated score calculation means (120) that calculates a first integrated score; a second integrated score calculation means (130) that calculates a second integrated score by performing integration in which the weight of the second score is increased; a score determination means (140) that determines whether or not each of the first integrated score and the second integrated score exceeds a prescribed threshold value, and also determines whether or not the system is in a specific state in which only the second integrated score exceeds the prescribed threshold value; and an output means (150) that outputs an inference result based on whether or not the first integrated score exceeds the prescribed threshold value if the system is not in the specific state, and outputs information different from the inference result if the system is in the specific state.

Description

Technical Field

[0001] This disclosure relates to the technical field of information processing systems, information processing methods, and recording media.

Background Art

[0002] As this type of system, there is known one that calculates various scores using a learned estimation model. For example, in Patent Document 1, in a system for performing face authentication, it is disclosed that a reference face image satisfying a score threshold is used as a face image candidate for authentication, and the face image candidate and its calculated score are selected as authentication information. In Patent Document 2, it is disclosed that an attribute-dependent score and a non-attribute-dependent score are integrated and output as a matching score.

[0003] As other related technologies, for example, in Patent Document 3, it is disclosed that when an error occurs in face authentication, information indicating a method for dealing with the authentication target person is output.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Patent Document 3

Summary of the Invention

Problems to be Solved by the Invention

[0005] This disclosure aims to improve the technologies disclosed in the prior art documents.

Means for Solving the Problems

[0006] One aspect of the information processing system disclosed herein includes: a score acquisition means for acquiring a first score based on the output of a first inference model and a second score based on the output of a second inference model having different characteristics from the first inference model; a first integrated score calculation means for calculating a first integrated score by integrating the first score and the second score; a second integrated score calculation means for calculating a second integrated score by integrating the first score and the second score such that the weight of the second score is greater than when the first integrated score is calculated; a score determination means for determining whether each of the first integrated score and the second integrated score exceeds a predetermined threshold, and whether a specific state exists in which only the second integrated score exceeds the predetermined threshold; and an output means for outputting an inference result based on whether the first integrated score exceeds the predetermined threshold when the specific state does not exist, and outputting information different from the inference result when the specific state exists.

[0007] One aspect of the information processing method of this disclosure involves, using at least one computer, obtaining a first score based on the output of a first inference model and a second score based on the output of a second inference model having different characteristics from the first inference model; calculating a first integrated score by integrating the first score and the second score; calculating a second integrated score by integrating the first score and the second score such that the weight of the second score is greater than when the first integrated score was calculated; determining whether each of the first integrated score and the second integrated score exceeds a predetermined threshold; determining whether a specific state exists where only the second integrated score exceeds the predetermined threshold; outputting an inference result based on whether the first integrated score exceeds the predetermined threshold if the specific state does not exist; and outputting information different from the inference result if the specific state exists.

[0008] One aspect of the recording medium of this disclosure includes a computer program that causes at least one computer to execute an information processing method, which involves acquiring a first score based on the output of a first inference model and a second score based on the output of a second inference model having different characteristics from the first inference model; calculating a first integrated score by integrating the first and second scores; calculating a second integrated score by integrating the first and second scores such that the weight of the second score is greater than when the first integrated score is calculated; determining whether each of the first integrated score and the second integrated score exceeds a predetermined threshold; determining whether a specific state exists where only the second integrated score exceeds the predetermined threshold; outputting an inference result based on whether the first integrated score exceeds the predetermined threshold if the specific state does not exist; and outputting information different from the inference result if the specific state exists. [Brief explanation of the drawing]

[0009] [Figure 1] This is a block diagram showing the hardware configuration of the information processing system according to the first embodiment. [Figure 2] This is a block diagram showing the functional configuration of the information processing system according to the first embodiment. [Figure 3] This is a flowchart showing the operation flow of the information processing system according to the first embodiment. [Figure 4] This is a flowchart showing the operation flow of the information processing system according to the second embodiment. [Figure 5] This is a graph showing the region corresponding to a specific state in the information processing system according to the second embodiment. [Figure 6] This graph shows a method for determining the region corresponding to a specific state in the information processing system according to the third embodiment. [Figure 7] This graph shows a method for calculating the size of a region corresponding to a specific state in an information processing system according to the third embodiment. [Figure 8] This is a block diagram showing the functional configuration of the information processing system according to the fourth embodiment. [Figure 9] It is a flowchart showing the operation flow of the information processing system according to the fourth embodiment. [Figure 10] It is a graph showing a region corresponding to a specific state in the information processing system according to the fourth embodiment. [Figure 11] It is a block diagram showing the configuration of a model used by the information processing system 10 according to the fifth embodiment. [Figure 12] It is a graph showing a region corresponding to a specific state in the information processing system according to the fifth embodiment.

Embodiments for Carrying Out the Invention

[0010] Hereinafter, embodiments of an information processing system, an information processing method, and a recording medium will be described with reference to the drawings.

[0011] <First Embodiment> The information processing system according to the first embodiment will be described with reference to FIGS. 1 to 3.

[0012] (Hardware Configuration) First, the hardware configuration of the information processing system according to the first embodiment will be described with reference to FIG. 1. FIG. 1 is a block diagram showing the hardware configuration of the information processing system according to the first embodiment.

[0013] As shown in FIG. 1, the information processing system 10 according to the first embodiment includes a processor 11, a RAM (Random Access Memory) 12, a ROM (Read Only Memory) 13, and a storage device 14. The information processing system 10 may further include an input device 15 and an output device 16. The above-described processor 11, RAM 12, ROM 13, storage device 14, input device 15, and output device 16 are connected via a data bus 17.

[0014] Processor 11 reads a computer program. For example, processor 11 is configured to read a computer program stored in at least one of RAM 12, ROM 13, and storage device 14. Alternatively, processor 11 may read a computer program stored in a computer-readable recording medium using a recording medium reader (not shown). Processor 11 may obtain (i.e., read) a computer program from a device (not shown) disposed outside information processing system 10 via a network interface. By executing the read computer program, processor 11 controls RAM 12, storage device 14, input device 15, and output device 16. In particular, in this embodiment, when processor 11 executes the read computer program, a functional block for integrating scores of an inference model and outputting an inference result based on the integrated scores is realized within processor 11. That is, processor 11 may function as a controller that executes each control in information processing system 10.

[0015] Processor 11 may be configured as, for example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an FPGA (field-programmable gate array), a DSP (Demand-Side Platform), an ASIC (Application Specific Integrated Circuit), or a quantum processor. Processor 11 may be constituted by one of these, or may be configured to use a plurality of them in parallel.

[0016] RAM12 temporarily stores computer programs executed by processor 11. RAM12 also temporarily stores data that processor 11 uses temporarily while executing computer programs. RAM12 may be, for example, D-RAM (Dynamic Random Access Memory) or SRAM (Static Random Access Memory). Alternatively, other types of volatile memory may be used instead of RAM12.

[0017] ROM 13 stores computer programs executed by processor 11. ROM 13 may also store other static data. ROM 13 may be, for example, P-ROM (Programmable Read Only Memory) or EPROM (Erasable Read Only Memory). Alternatively, other types of non-volatile memory may be used instead of ROM 13.

[0018] The storage device 14 stores data that the information processing system 10 will save for the long term. The storage device 14 may also operate as a temporary storage device for the processor 11. The storage device 14 may include, for example, at least one of a hard disk drive, a magneto-optical disk drive, an SSD (Solid State Drive), and a disk array device.

[0019] The input device 15 is a device that receives input instructions from the user of the information processing system 10. The input device 15 may include, for example, at least one of a keyboard, a mouse, and a touch panel. The input device 15 may be configured as a mobile terminal such as a smartphone or tablet. The input device 15 may also be a device capable of voice input, for example, including a microphone.

[0020] The output device 16 is a device that outputs information related to the information processing system 10 to the outside. For example, the output device 16 may be a display device (e.g., a display) capable of displaying information related to the information processing system 10. Alternatively, the output device 16 may be a speaker or the like capable of outputting information related to the information processing system 10 as sound. The output device 16 may be configured as a mobile terminal such as a smartphone or tablet. Furthermore, the output device 16 may be a device that outputs information in a format other than an image. For example, the output device 16 may be a speaker that outputs information related to the information processing system 10 as sound.

[0021] Although Figure 1 shows an example of an information processing system 10 comprising multiple devices, all or some of these functions may be implemented by a single device (information processing device). In that case, the information processing device may be configured to include only the processor 11, RAM 12, and ROM 13 described above, and the other components (i.e., storage device 14, input device 15, output device 16) may be provided by external devices connected to the information processing device. Furthermore, the information processing device may implement some of its computational functions through external devices (e.g., external servers or cloud services).

[0022] (Functional configuration) Next, the functional configuration of the information processing system 10 according to the first embodiment will be described with reference to Figure 2. Figure 2 is a block diagram showing the functional configuration of the information processing system according to the first embodiment.

[0023] The information processing system 10 according to the first embodiment is configured to output an inference result by integrating the outputs (scores) of multiple inference models having different characteristics. Hereinafter, these inference models will be referred to as the first inference model and the second inference model. The specific forms of the first and second inference models are not particularly limited, but they may be configured, for example, as an authentication model used when performing biometric authentication or as a detection model used when detecting diseases such as cancer. The information processing system 10 may be configured to include inference models within the system itself, or it may be configured to utilize inference models from outside the system.

[0024] As shown in Figure 2, the information processing system 10 according to the first embodiment is configured to include, as components for realizing its functions, a score acquisition unit 110, a first integrated score calculation unit 120, a second integrated score calculation unit 130, a score determination unit 140, and an output unit 150. Each of the score acquisition unit 110, the first integrated score calculation unit 120, the second integrated score calculation unit 130, the score determination unit 140, and the output unit 150 may be a processing block realized by, for example, the processor 11 (see Figure 1) described above.

[0025] The score acquisition unit 110 is configured to acquire a first score based on the output of the first inference model and a second score based on the output of the second inference model (i.e., a model with different characteristics from the first inference model). The first score and the second score acquired by the score acquisition unit 110 are output to the first integrated score calculation unit 120 and the second integrated score calculation unit, respectively.

[0026] The first integrated score calculation unit 120 is configured to calculate a first integrated score by integrating the first score and the second score. The method for calculating the first integrated score is not particularly limited, but the first integrated score calculation unit 120 performs the integration process by weighting the first score and the second score. The first integrated score calculated by the first integrated score calculation unit 120 is output to the score determination unit 140.

[0027] The second integrated score calculation unit 130 is configured to calculate a second integrated score by integrating the first score and the second score. The method for calculating the second integrated score is not particularly limited, but the second integrated score calculation unit 130 is configured to perform the integration process in such a way that the weight of the second score is greater than when calculating the first integrated score (i.e., the processing of the first integrated score calculation unit 120). As a result, the first integrated score and the second integrated score will be different scores. The first integrated score calculated by the second integrated score calculation unit 130 is output to the score determination unit 140.

[0028] The score determination unit 140 is configured to perform various determination processes using the first integrated score calculated by the first integrated score calculation unit 120 and the second integrated score calculated by the second integrated score calculation unit 130. Specifically, the score determination unit 140 is configured to determine whether each of the first integrated score and the second integrated score exceeds a predetermined threshold. The "predetermined threshold" here is a value set in advance to produce an inference result. The score determination unit 140 is also configured to determine whether a specific state is in place. The "specific state" here refers to a state in which only the second integrated score exceeds the predetermined threshold. The determination result from the score determination unit 140 is output to the output unit 150.

[0029] The output unit 150 is configured to output various types of information according to the determination result of the score determination unit 140. Specifically, when the score determination unit 140 determines that the system is not in a specific state, the output unit 140 outputs an inference result based on whether or not the first integrated score exceeds a predetermined threshold. On the other hand, when the score determination unit 140 determines that the system is in a specific state, the output unit 140 outputs information different from the inference result. Specific examples of information different from the inference result will be explained in detail in other embodiments described later.

[0030] (Flow of operations) Next, with reference to Figure 3, the operation flow of the information processing system 10 according to the first embodiment will be described. Figure 3 is a flowchart showing the operation flow of the information processing system according to the first embodiment.

[0031] As shown in Figure 3, when the operation of the information processing system 10 according to the first embodiment begins, the score acquisition unit 120 first acquires a first score based on the output of the first inference model and a second score based on the output of the second inference model (step S101). The score acquisition unit 120 may acquire the first score and the second score simultaneously, or it may acquire them sequentially.

[0032] Next, the first integrated score calculation unit 120 calculates the first integrated score based on the first score and second score obtained by the score acquisition unit 120 (step S102). Then, the second integrated score calculation unit 130 calculates the second integrated score based on the first score and second score obtained by the score acquisition unit 120 (step S103). Note that the processes in steps S102 and S103 may be executed sequentially or simultaneously in parallel.

[0033] Next, the score determination unit 140 determines whether or not a specific state exists based on the first integrated score calculated by the first integrated score calculation unit 120 and the second integrated score calculated by the second integrated score calculation unit 130 (step S104). If it is determined that the specific state does not exist (step 104: NO), the score determination unit 140 performs inference based on the first integrated score, and the output unit 150 outputs the inference result (step S105). On the other hand, if it is determined that the specific state exists (step 104: YES), the output unit 150 outputs information other than the inference result.

[0034] (Technical effects) Next, the technical effects obtained by the information processing system 10 according to the first embodiment will be described.

[0035] As explained in Figures 1 to 3, in the information processing system 10 according to the first embodiment, a first integrated score and a second integrated score are calculated from the first score of the first inference model and the second score of the second inference model. Then, depending on the determination result using the first integrated score and the second integrated score, the inference result or information other than the inference result is output. In this way, it is possible to output more appropriate information by taking into account the characteristics of each of the multiple models. The output that takes into account the characteristics of the inference model will be explained in detail in other embodiments described later.

[0036] <Second Embodiment> The information processing system 10 according to the second embodiment will be described with reference to Figures 4 and 5. Note that the second embodiment differs from the first embodiment described above only in some configurations and operations; other parts may be identical to the first embodiment. Therefore, the following will describe in detail the parts that differ from the first embodiment already described, and will omit explanations of other overlapping parts as appropriate.

[0037] (Authentication Model) First, we will explain the inference model used in the information processing system 10 according to the second embodiment.

[0038] In the information processing system 10 according to the second embodiment, the first inference model and the second inference model are configured as authentication models. The authentication model is a model that takes information about the subject as input and authenticates whether or not the subject is a registered user. The authentication model may be configured as a biometric authentication model that performs authentication processing using, for example, a biometric image (or features extracted from a biometric image). For example, the authentication model may be configured as a facial recognition model that performs authentication using a facial image, or as an iris recognition model that performs authentication using an iris image. In this case, the first model and the second model may be configured as models that take information about different modals as input. For example, the first inference model may be configured as a facial recognition model, and the second authentication model may be configured as an iris recognition model. In this case, the information processing system 10 may be configured as an authentication system capable of performing multimodal authentication.

[0039] In the following explanation, we will use examples where the first and second inference models are configured as facial recognition models. Furthermore, the first inference model is configured as a model that is broadly applicable to general faces in general (hereinafter referred to as the "general-purpose model"), and the second inference model is configured as a model specialized for certain types of faces (hereinafter referred to as the "expert model"). Examples of expert models include models specialized for faces wearing masks, or models specialized for profiles.

[0040] (Flow of operations) Next, with reference to Figure 4, the operation flow of the information processing system 10 according to the second embodiment will be described. Figure 4 is a flowchart showing the operation flow of the information processing system according to the second embodiment.

[0041] As shown in Figure 4, when the operation of the information processing system 10 according to the second embodiment is started, the face image to be authenticated is first input to each authentication model (i.e., the general-purpose model and the expert model) (step S201). Note that the verification model may be input with features extracted from the face image.

[0042] Next, the score acquisition unit 120 acquires the first score calculated by the general-purpose model (step S202). The score acquisition unit 120 also acquires the second score calculated by the expert model (step S203).

[0043] Next, the first integrated score calculation unit 120 calculates the first integrated score based on the first and second scores obtained by the score acquisition unit 120 (step S204). Then, the second integrated score calculation unit 130 calculates the second integrated score based on the first and second scores obtained by the score acquisition unit 120 (step S205). As already explained, the second integrated score is calculated so that the weight of the first score (i.e., the score calculated by the expert model) is greater than that of the first integrated score.

[0044] Next, the score determination unit 140 determines whether a specific state exists based on the first integrated score calculated by the first integrated score calculation unit 120 and the second integrated score calculated by the second integrated score calculation unit 130 (step S206). If it is determined that the specific state does not exist (step 206: NO), the score determination unit 140 performs authentication processing based on the first integrated score, and the output unit 150 outputs the authentication result (step S207). For example, if the first integrated score exceeds a predetermined threshold, the output unit 150 outputs information indicating authentication OK (i.e., the subject is a registered user). If the first integrated score does not exceed the predetermined threshold, the output unit 150 outputs information indicating authentication NG (i.e., the subject is not a registered user).

[0045] On the other hand, if a specific state is determined (Step 206: YES), guidance information for the subject is output as information other than the authentication result. Guidance information is, for example, information that requests a specific action from the subject. Guidance information may be information that prompts the subject to move so that their face is in a position suitable for imaging. Specifically, a message such as "Please bring your face closer" may be output. Alternatively, guidance information may be information that prompts the subject to remove their mask. Specifically, a message such as "Please remove your mask" may be output. This guidance information may be displayed on a screen, for example, or output as audio.

[0046] (Specific condition) Next, with reference to Figure 5, a specific state in the information processing system 10 according to the second embodiment will be described in detail. Figure 5 is a graph showing the region corresponding to the specific state in the information processing system according to the second embodiment.

[0047] In Figure 5, the specific state determined by the information processing system 10 according to the second embodiment is defined as the shaded area. Specifically, the specific state corresponds to a state in which the second score (i.e., the expert model score) is high but is rejected (judged as incorrect) because the first score (i.e., the general-purpose model score) is low.

[0048] The specific conditions described above can occur, for example, when the subject is wearing a mask. If the subject is wearing a mask, even if the subject is the actual person (i.e., a registered user), the first score will be calculated low because the face is hidden by the mask. On the other hand, if the expert model is configured as a mask-specific model, the second score will be calculated high because the model can accurately recognize the subject even with a mask on. In such a situation, the person should ideally be identified as the same person, but if the determination is made based only on the first score, the first combined score will also be low due to the low first score, which may lead to the person being mistakenly identified as someone else.

[0049] However, in this embodiment, a second integrated score, calculated by increasing the weight of the expert score, is used to determine whether or not there is a possibility of an inappropriate judgment being made as described above (i.e., a specific state). If a specific state is present, no authentication result is output, and information other than the authentication result is output. In this embodiment, guidance information is given as an example of information other than the authentication result, but other information may also be output. For example, if a specific state is present, alert information (i.e., information warning that accurate authentication may not be possible) may be output.

[0050] (Technical effects) Next, the technical effects obtained by the information processing system 10 according to the second embodiment will be described.

[0051] As explained in Figures 4 and 5, in the information processing system 10 according to the second embodiment, an authentication result is output when the system is not in a specific state, and information other than the authentication result is output when the system is in a specific state. In this way, it is possible to suppress the output of an incorrect authentication result when the system is in a specific state. Furthermore, by outputting guidance information when the system is in a specific state, it is possible to request a predetermined action from the user, thereby resolving the specific state.

[0052] <Third Embodiment> The information processing system 10 according to the third embodiment will be described with reference to Figures 6 and 7. The third embodiment is an example of a method for determining the region corresponding to the specific state described in the second embodiment (see Figure 5), and other parts may be the same as those of the first and second embodiments. For this reason, the following will describe in detail the parts that differ from the embodiments already described, and will omit explanations of other overlapping parts as appropriate.

[0053] (How to set the area) First, with reference to Figure 6, the method for setting the area corresponding to a specific state in the information processing system 10 according to the third embodiment will be explained. Figure 6 is a graph showing the method for determining the area corresponding to a specific state in the information processing system according to the third embodiment.

[0054] As shown in Figure 6, the size of the area corresponding to a specific state can be determined by determining α (i.e., the value corresponding to the width of the area). In this case, the value of α may be set based on the false positive rate in the authentication process (i.e., the probability of misidentifying someone as the real person). For example, α may be determined such that the probability of the area corresponding to the specific state in the figure occurring is the same as the probability of the area corresponding to the false positive rate occurring.

[0055] Incidentally, the rate of misinterpretation by others is P FA (α) Therefore, the first integrated score τ1 can be calculated as shown in equation (1) below, and the second integrated score τ2 can be calculated as shown in equation (2) below.

[0056]

number

[0057]

number

[0058] As described above, the second integrated score τ2 may be calculated based on the rate of misinterpretation by others. Note 1-p (α) This corresponds to weights based on inference accuracy, and -logP FA (α) These correspond to the first and second scores.

[0059] (Calculation of the size of the region) Next, with reference to Figure 7, a method for calculating the size of the area corresponding to a specific state in the information processing system 10 according to the third embodiment will be described. Figure 7 is a graph showing the method for calculating the size of the area corresponding to a specific state in the information processing system according to the third embodiment.

[0060] In Figure 7, the probability of occurrence of a region corresponding to a specific state can be determined in the same way as calculating the probability of occurrence of the trapezoidal region in the figure. Specifically, the probability P of occurrence of a region corresponding to a specific state can be calculated using the following formula (3).

[0061]

number

[0062] Note that c = ln10. Furthermore, each axis is assumed to be the negative logarithm of the misinterpretation rate. Also, each score is considered independent (uncorrelated) (i.e., it is assumed that each model has different characteristics).

[0063] (Technical effects) Next, the technical effects obtained by the information processing system 10 according to the third embodiment will be described.

[0064] As explained in Figures 6 and 7, in the information processing system 10 according to the third embodiment, the size of the area corresponding to a specific state is set based on the rate of false positives. In this way, the probability of a specific state occurring can be set to an appropriate value. Therefore, for example, it is possible to suppress the situation in which an excessive number of specific states occur, resulting in an overabundance of information other than the authentication result being output.

[0065] <Fourth Embodiment> The information processing system 10 according to the fourth embodiment will be described with reference to Figures 8 to 10. Note that the fourth embodiment differs from the first to third embodiments described above only in some configurations and operations; other parts may be identical to those of the first to third embodiments. Therefore, the following will describe in detail the parts that differ from each embodiment already described, while omitting explanations of other overlapping parts as appropriate.

[0066] (Functional configuration) First, the functional configuration of the information processing system 10 according to the fourth embodiment will be described with reference to Figure 8. Figure 8 is a block diagram showing the functional configuration of the information processing system according to the fourth embodiment. Note that in Figure 8, the same reference numerals are used for elements similar to those described in Figure 2.

[0067] As shown in Figure 8, the information processing system 10 according to the fourth embodiment is configured to include, as components for realizing its functions, a score acquisition unit 110, a first integrated score calculation unit 120, a second integrated score calculation unit 130, a score determination unit 140, an output unit 150, and a third integrated score calculation unit 160. That is, the information processing system 10 according to the fourth embodiment includes the third integrated score calculation unit 160 in addition to the configuration of the first embodiment (see Figure 2). The third integrated score calculation unit 160 may be, for example, a processing block realized by the processor 11 (see Figure 1) described above.

[0068] The third integrated score calculation unit 160 is configured to calculate a third integrated score by integrating the first score and the second score. The method for calculating the third integrated score is not particularly limited, but the third integrated score calculation unit 160 is configured to perform the integration process in such a way that the weight of the first score is greater than when calculating the first integrated score (i.e., the processing of the first integrated score calculation unit 120). Therefore, the third integrated score will be a score different from both the first integrated score and the second integrated score. The third integrated score calculated by the third integrated score calculation unit 160 is output to the score determination unit 140.

[0069] Furthermore, the score determination unit 140 according to the fourth embodiment is configured to perform a determination using a third integrated score in addition to the first integrated score and the second integrated score. The determination operation of the score determination unit 140 will be explained in detail below.

[0070] (Flow of operations) Next, the operation flow of the information processing system 10 according to the fourth embodiment will be described with reference to Figure 9. Figure 9 is a flowchart showing the operation flow of the information processing system according to the fourth embodiment. Note that in Figure 9, the same reference numerals are used for the same processes as shown in Figure 3.

[0071] As shown in Figure 9, when the operation of the information processing system 10 according to the fourth embodiment is started, the score acquisition unit 120 first acquires a first score based on the output of the first inference model and a second score based on the output of the second inference model (step S101).

[0072] Next, the first integrated score calculation unit 120 calculates the first integrated score based on the first and second scores obtained by the score acquisition unit 120 (step S102). Then, the second integrated score calculation unit 130 calculates the second integrated score based on the first and second scores obtained by the score acquisition unit 120 (step S103). Furthermore, the third integrated score calculation unit 160 calculates the third integrated score based on the first and second scores obtained by the score acquisition unit 120 (step S401).

[0073] The third integrated score can be calculated using the same method as when calculating the second integrated score. Specifically, the third integrated score can be calculated by swapping the first and second scores in the formula (2) described above.

[0074] Next, the score determination unit 140 determines whether or not a specific state is in place based on the first integrated score calculated by the first integrated score calculation unit 120, the second integrated score calculated by the second integrated score calculation unit 130, and the third integrated score calculated by the third integrated score calculation unit 160 (step S402). In particular, the score determination unit 140 according to this embodiment determines that a specific state is in place if only the second integrated score exceeds a predetermined threshold, or if only the third integrated score exceeds a predetermined threshold.

[0075] If it is determined that the system is not in a specific state (Step 402: NO), the score determination unit 140 performs inference based on the first integrated score, and the output unit 150 outputs the inference result (Step S105). On the other hand, if it is determined that the system is in a specific state (Step 402: YES), the output unit 150 outputs information other than the inference result.

[0076] (Multiple specific states) Next, with reference to Figure 10, a specific state in the information processing system 10 according to the fourth embodiment will be described in detail. Figure 10 is a graph showing the region corresponding to a specific state in the information processing system according to the fourth embodiment.

[0077] As shown in Figure 10, in the information processing system 10 according to the fourth embodiment, multiple specific states are determined. Specifically, in the embodiments described above, examples were given of determining a specific state in which the first score is low and the second score is high, but in this embodiment, a specific state in which the first score is high and the second score is low is also determined.

[0078] (Technical effects) Next, the technical effects obtained by the information processing system 10 according to the fourth embodiment will be described.

[0079] As explained in Figures 8 to 10, in the information processing system 10 according to the fourth embodiment, a specific state is determined using the third integrated score. In this way, multiple specific states can be determined, making it possible to output information while taking into account the characteristics of each of the multiple models more than when only one specific state is determined.

[0080] <Fifth Embodiment> The information processing system 10 according to the fifth embodiment will be described with reference to Figures 11 and 12. Note that the fifth embodiment differs from the first to fourth embodiments described above only in some operations; other parts may be identical to the first to fourth embodiments. Therefore, the following will describe in detail the parts that differ from the embodiments already described, while other overlapping parts will be omitted as appropriate.

[0081] (Model configuration) First, the configuration of the model used by the information processing system 10 according to the fifth embodiment will be described with reference to Figure 11. Figure 11 is a block diagram showing the configuration of the model used by the information processing system 10 according to the fifth embodiment.

[0082] As shown in Figure 11, the information processing system 10 according to the fifth embodiment uses one first inference model and two second inference models A and B. The first inference model outputs a first score. Second inference model A outputs a second score A. Second inference model B outputs a second score B. Therefore, the score acquisition unit 110 in the information processing system 10 according to the fifth embodiment acquires three scores: the first score, the second score A, and the second score B.

[0083] Furthermore, the second integrated score calculation unit 130 according to the fifth embodiment calculates a second integrated score for each second score for each second inference model. In other words, the second integrated score calculation unit 130 calculates a second integrated score for each second score. Specifically, the second integrated score calculation unit 130 calculates a second integrated score A from the first score and the second score A. Also, the second integrated score calculation unit 130 calculates a second integrated score B from the first score and the second score B.

[0084] (Specific condition) Next, with reference to Figure 12, a specific state in the information processing system 10 according to the fifth embodiment will be described in detail. Figure 12 is a graph showing the region corresponding to a specific state in the information processing system according to the fifth embodiment.

[0085] As shown in Figure 12, in the information processing system 10 according to the fifth embodiment, a specific state is determined to exist when either the first integrated score A or the second integrated score B exceeds a predetermined threshold. The size of the region corresponding to this specific state is defined by α and β2 in the figure. The values ​​of α and β2 may be calculated using the formula (3) described in the third embodiment (see Figure 7). In this case, the values ​​of α and β2 may be calculated such that the probability of occurrence of the specific region is the same as the probability of occurrence of the region corresponding to the rate of misinterpretation by others.

[0086] Although this explanation describes the case where there are two second inference models, there may be three or more second inference models. In that case as well, by calculating a second integrated score corresponding to each of the multiple second inference models (i.e., by calculating a second integrated score for each second score), a specific state can be determined using a similar method. The i-th second integrated score can be calculated, for example, using equation (4) below.

[0087]

number

[0088] (Technical effects) Next, the technical effects obtained by the information processing system 10 according to the fifth embodiment will be described.

[0089] As explained in Figures 11 and 12, in the information processing system 10 according to the fifth embodiment, multiple second integrated scores are calculated using multiple second scores. In this way, even when there are multiple second inference models, more appropriate information can be output by taking into account the respective characteristics of each model.

[0090] The processing method of recording a program that operates the configuration of each embodiment in order to realize the functions of each embodiment described above on a recording medium, reading the program recorded on the recording medium as code, and executing it on a computer is also included in the scope of each embodiment. In other words, a computer-readable recording medium is also included in the scope of each embodiment. Furthermore, not only the recording medium on which the above-mentioned program is recorded, but also the program itself is included in each embodiment.

[0091] As recording media, for example, floppy disks, hard disks, optical disks, magneto-optical disks, CD-ROMs, magnetic tapes, non-volatile memory cards, and ROMs can be used. Furthermore, the scope of each embodiment is not limited to programs that perform processing on the recording media alone, but also includes programs that operate on the OS and perform processing in cooperation with other software and the functions of expansion boards. In addition, the program itself may be stored on a server, and part or all of the program may be made available for download from the server to the user terminal. The program may be provided to the user in, for example, SaaS (Software as a Service) format.

[0092] <Note> The embodiments described above may also be described in the following appendix, but are not limited to these.

[0093] (Note 1) The information processing system described in Appendix 1 is an information processing system comprising: a score acquisition means for acquiring a first score based on the output of a first inference model and a second score based on the output of a second inference model having different characteristics from the first inference model; a first integrated score calculation means for calculating a first integrated score by integrating the first score and the second score; a second integrated score calculation means for calculating a second integrated score by integrating the first score and the second score such that the weight of the second score is greater than when the first integrated score is calculated; a score determination means for determining whether each of the first integrated score and the second integrated score exceeds a predetermined threshold, and whether a specific state exists in which only the second integrated score exceeds the predetermined threshold; and an output means for outputting an inference result based on whether the first integrated score exceeds the predetermined threshold when the specific state does not exist, and outputting information different from the inference result when the specific state exists.

[0094] (Note 2) The information processing system described in Appendix 2 is the information processing system described in Appendix 1, wherein the first inference model and the second inference model are authentication models that take information about the subject as input and authenticate whether the subject is a registered user or not, and the output means outputs the authentication result of the subject when the subject is not in the specified state, and outputs information different from the authentication result when the subject is in the specified state.

[0095] (Note 3) The information processing system described in Appendix 3 is the information processing system described in Appendix 2, wherein the output means outputs information requesting the target person to perform a predetermined action when the specified state is not met.

[0096] (Note 4) The information processing system described in Appendix 4 is the information processing system described in any one of the items in Appendix 1 to 3, wherein the second integrated score is a value calculated based on the rate of misinterpretation by others.

[0097] (Note 5) The information processing system described in Appendix 5 further comprises a third integrated score calculation means that calculates a third integrated score by integrating the first score and the second score such that the weight of the first score is greater than when calculating the first integrated score, and the score determination means determines that the specific state is met when either the second integrated score or the third integrated score exceeds the predetermined threshold, the information processing system described in any one of Appendix 1 to 4.

[0098] (Note 6) The information processing system described in Appendix 6 is the information processing system described in any one of Appendix 1 to 5, wherein the second inference model includes a plurality of models having different characteristics from each other, the second integrated scoring means calculates the second integrated score for each of the plurality of models, and the score determination means determines that the specific state is met if any of the plurality of second integrated scores exceeds the predetermined threshold.

[0099] (Note 7) The information processing method described in Appendix 7 is an information processing method which obtains a first score based on the output of a first inference model and a second score based on the output of a second inference model having different characteristics from the first inference model using at least one computer, calculates a first integrated score by integrating the first score and the second score, calculates a second integrated score by integrating the first score and the second score such that the weight of the second score is greater than when the first integrated score is calculated, determines whether each of the first integrated score and the second integrated score exceeds a predetermined threshold, determines whether a specific state exists where only the second integrated score exceeds the predetermined threshold, outputs an inference result based on whether the first integrated score exceeds the predetermined threshold if the specific state does not exist, and outputs information different from the inference result if the specific state exists.

[0100] (Note 8) The recording medium described in Appendix 8 is a recording medium on which a computer program is recorded that causes at least one computer to execute an information processing method that acquires a first score based on the output of a first inference model and a second score based on the output of a second inference model having different characteristics from the first inference model, calculates a first integrated score by integrating the first score and the second score, calculates a second integrated score by integrating the first score and the second score such that the weight of the second score is greater than when the first integrated score is calculated, determines whether each of the first integrated score and the second integrated score exceeds a predetermined threshold, determines whether a specific state exists where only the second integrated score exceeds the predetermined threshold, outputs an inference result based on whether the first integrated score exceeds the predetermined threshold if the specific state does not exist, and outputs information different from the inference result if the specific state exists.

[0101] (Note 9) The computer program described in Appendix 9 is a computer program that causes at least one computer to execute an information processing method which involves obtaining a first score based on the output of a first inference model and a second score based on the output of a second inference model having different characteristics from the first inference model, calculating a first integrated score by integrating the first score and the second score, calculating a second integrated score by integrating the first score and the second score such that the weight of the second score is greater than when the first integrated score is calculated, determining whether each of the first integrated score and the second integrated score exceeds a predetermined threshold, determining whether a specific state exists where only the second integrated score exceeds the predetermined threshold, outputting an inference result based on whether the first integrated score exceeds the predetermined threshold if the specific state does not exist, and outputting information different from the inference result if the specific state exists.

[0102] (Note 10) The information processing device described in Appendix 10 is an information processing device comprising: a score acquisition means for acquiring a first score based on the output of a first inference model and a second score based on the output of a second inference model having different characteristics from the first inference model; a first integrated score calculation means for calculating a first integrated score by integrating the first score and the second score; a second integrated score calculation means for calculating a second integrated score by integrating the first score and the second score such that the weight of the second score is greater than when the first integrated score is calculated; a score determination means for determining whether each of the first integrated score and the second integrated score exceeds a predetermined threshold, and whether a specific state exists in which only the second integrated score exceeds the predetermined threshold; and an output means for outputting an inference result based on whether the first integrated score exceeds the predetermined threshold when the specific state does not exist, and outputting information different from the inference result when the specific state exists.

[0103] This disclosure may be modified as appropriate, insofar as it does not contradict the gist or idea of ​​the invention as can be inferred from the claims and the specification as a whole, and information processing systems, information processing methods, and recording media that include such modifications are also included in the technical idea of ​​this disclosure. [Explanation of Symbols]

[0104] 10 Information Processing Systems 11 processors 110 Score Acquisition Section 120 First Integrated Score Calculation Unit 130 Second Integrated Score Calculation Unit 140 Score Judging Section 150 Output section 160 Third Integrated Score Calculation Unit

Claims

1. A score acquisition means that acquires a first score based on the output of a first inference model and a second score based on the output of a second inference model having different characteristics from the first inference model, A first integrated score calculation means that calculates a first integrated score by integrating the first score and the second score, A second integrated score calculation means calculates a second integrated score by integrating the first score and the second score such that the weight of the second score is greater than when calculating the first integrated score, A score determination means that determines whether each of the first integrated score and the second integrated score exceeds a predetermined threshold, and determines whether only the second integrated score exceeds the predetermined threshold, An output means that outputs an inference result based on whether the first integrated score exceeds the predetermined threshold when the specified state is not met, and outputs information different from the inference result when the specified state is met, An information processing system equipped with the following features.

2. The first and second inference models are authentication models that take information about the subject as input to authenticate whether or not the subject is a registered user. The output means outputs the authentication result of the subject when the subject is not in the specified state, and outputs information different from the authentication result when the subject is in the specified state. The information processing system according to claim 1.

3. The output means outputs information requesting the subject to perform a predetermined action when the specified state is not met. The information processing system according to claim 2.

4. The second integrated score is a value calculated based on the rate of misinterpretation by others. The information processing system according to claim 2 or 3.

5. The system further includes a third integrated score calculation means that calculates a third integrated score by integrating the first score and the second score such that the weight of the first score is greater than when calculating the first integrated score, The score determination means determines that the specific state exists when only the second integrated score exceeds a predetermined threshold, or when only the third integrated score exceeds a predetermined threshold. The information processing system according to any one of claims 1 to 3.

6. The second inference model includes multiple models with different characteristics from each other. The second integrated score calculation means calculates the second integrated score for each of the plurality of models, The score determination means determines that the specific state exists when the first integrated score does not exceed the predetermined threshold, and any of the multiple second integrated scores exceeds the predetermined threshold. The information processing system according to any one of claims 1 to 3.

7. By at least one computer, A first score based on the output of the first inference model and a second score based on the output of a second inference model having different characteristics from the first inference model are obtained. The first combined score is calculated by combining the first score and the second score. Compared to the method used to calculate the first integrated score, the first score and the second score are combined to calculate the second integrated score, such that the second score has a greater weight. The first integrated score and the second integrated score are determined to be whether or not each exceeds a predetermined threshold, and the second integrated score is determined to be in a specific state where only the second integrated score exceeds the predetermined threshold. If the specified state is not met, an inference result based on whether the first integrated score exceeds the predetermined threshold is output; if the specified state is met, information different from the inference result is output. Information processing methods.

8. On at least one computer, A first score based on the output of the first inference model and a second score based on the output of a second inference model having different characteristics from the first inference model are obtained. The first combined score is calculated by combining the first score and the second score. Compared to the method used to calculate the first integrated score, the first score and the second score are combined to calculate the second integrated score, such that the second score has a greater weight. The first integrated score and the second integrated score are determined to be whether or not each exceeds a predetermined threshold, and the second integrated score is determined to be in a specific state where only the second integrated score exceeds the predetermined threshold. If the specified state is not met, an inference result based on whether the first integrated score exceeds the predetermined threshold is output; if the specified state is met, information different from the inference result is output. A computer program that executes information processing methods.

Images