Machine learning device, machine learning method, and machine learning program

Abstract

An inference unit (113) performs, for each model parameter set, inference on input data using a machine learning model to which the model parameter set is set, and obtains output data indicating the result of inference by the machine learning model to which the model parameter set is set. A comparison unit (114) compares features of the inference results between sets of output data to obtain a comparison result. An output unit (115) determines, on the basis of the comparison result, information leakage-related information in the information included in the inference result indicated by one of the sets of output data, applies a change for the information leakage-related information to the output data, and outputs the changed output data.

Description

[Technical Field]

[0001] The present disclosure relates to measures to prevent information leakage from machine learning models. [Background technology]

[0002] Generative AI is expected to be used in a variety of industrial fields. AI is an abbreviation for artificial intelligence. In particular, there are high expectations for the use of machine learning models with large model parameters, such as large-scale language models.

[0003] Here, we will discuss large-scale language models as a representative example of machine learning models with large model parameters. However, what we will discuss here is also applicable to machine learning models for other tasks, such as image generation models. Building and training a large-scale language model from scratch requires enormous computational costs and a huge dataset. For this reason, additional training (fine tuning) of existing large-scale pre-trained language models using new training data is being considered. For example, additional training can be used to enrich pre-trained large-scale language models with company-specific knowledge.

[0004] When tuning all model parameters (full fine-tuning) for a large-scale language model, the computational cost remains high due to the parameter size of the large-scale language model. Therefore, Parameter Efficient Fine-Tuning (PEFT), which is a method of fine-tuning with good parameter efficiency, has attracted attention. PEFT tunes only a small number of additional model parameters without updating most of the model parameters of the pre-trained model, thereby significantly reducing computational costs and storage. PEFT replaces a small number of tuned parameters without replacing the entire model, which allows a single pre-trained model to be used to perform multiple tasks.

[0005] A large-scale language model has a huge model structure, or in other words, a huge number of model parameters. For this reason, it has been pointed out that large-scale language models tend to store training data internally, are vulnerable to the leakage of private information related to training data from the model, and are also vulnerable to the leakage of sensitive information.

[0006] Patent Document 1 discloses the configuration of a system for evaluating the risk of information leakage from a learning model. This vulnerability exists even in large, fine-tuned language models. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] Japanese Patent Publication No. 2022-007311 Summary of the Invention [Problem to be solved by the invention]

[0008] Fine-tuned machine learning models (large-scale language models) have the problem of information leakage.

[0009] The present disclosure aims to suppress information leakage from machine learning models. [Means for solving the problem]

[0010] The machine learning device of the present disclosure includes: an inference unit that, for each model parameter set included in a plurality of different model parameter sets, performs inference on input data using a machine learning model to which the model parameter set is set, and obtains output data indicating an inference result by the machine learning model when the model parameter set is set; a comparison unit that compares features of the inference results between output data for a plurality of output data corresponding to the plurality of model parameter sets and obtains a comparison result; an output unit that determines information regarding information leakage from among the information included in the inference results shown in any of the output data based on the comparison result, modifies the output data with respect to the information regarding information leakage, and outputs the modified output data; Equipped with. [Effects of the Invention]

[0011] According to the present disclosure, it is possible to suppress information leakage from machine learning models. [Brief explanation of the drawings]

[0012] [Figure 1] FIG. 1 is a configuration diagram of a machine learning device 100 according to a first embodiment. [Figure 2] 3 is a flowchart of a machine learning method according to the first embodiment. [Figure 3] 3 is a flowchart of an example according to the first embodiment. [Figure 4] FIG. 10 is a configuration diagram of a machine learning device 100 according to a second embodiment. [Figure 5] 10 is a flowchart of a machine learning method according to the second embodiment. [Figure 6] 10 is a flowchart of an example according to the second embodiment. [Figure 7] FIG. 1 is a hardware configuration diagram of a machine learning device 100 according to an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0013] In the embodiments and drawings, the same or corresponding elements are denoted by the same reference numerals. The description of elements denoted by the same reference numerals as those already described will be omitted or simplified as appropriate. Arrows in the drawings primarily indicate the flow of data or the flow of processing.

[0014] Embodiment 1 Measures to prevent information leakage from machine learning models will be explained based on Figures 1 to 3.

[0015] ***Configuration Description*** The configuration of the machine learning device 100 will be described with reference to FIG. The machine learning device 100 is a computer that includes hardware such as a processor 101, a memory 102, an auxiliary storage device 103, and an input / output interface 104. These pieces of hardware are connected to each other via signal lines.

[0016] The processor 101 is an IC that performs arithmetic processing and controls other hardware. For example, the processor 101 is a CPU, a DSP, or a GPU. IC is an abbreviation for Integrated Circuit. CPU is an abbreviation for Central Processing Unit. DSP is an abbreviation for Digital Signal Processor. GPU is an abbreviation for Graphics Processing Unit.

[0017] The memory 102 is a volatile or non-volatile storage device. The memory 102 is also called a primary storage device or a main memory. For example, the memory 102 is a RAM. Data stored in the memory 102 is saved in the secondary storage device 103 as needed. RAM is an abbreviation for Random Access Memory.

[0018] The auxiliary storage device 103 is a non-volatile storage device. For example, the auxiliary storage device 103 is a ROM, a HDD, a flash memory, or a combination thereof. Data stored in the auxiliary storage device 103 is loaded into the memory 102 as needed. ROM is an abbreviation for Read Only Memory. HDD is an abbreviation for Hard Disk Drive.

[0019] The input / output interface 104 is a port to which an input device and an output device are connected. For example, the input / output interface 104 is a USB terminal, and the input devices are a keyboard, a mouse, and a communication device (receiver), and the output devices are a display and a communication device (transmitter). Input and output of the machine learning device 100 is performed using the input / output interface 104. USB is an abbreviation for Universal Serial Bus.

[0020] The machine learning device 100 includes elements such as an input unit 111, a switching unit 112, an inference unit 113, a comparison unit 114, and an output unit 115. These elements are realized by software.

[0021] The auxiliary storage device 103 stores a machine learning program for causing the computer to function as an input unit 111, a switching unit 112, an inference unit 113, a comparison unit 114, and an output unit 115. The machine learning program is loaded into the memory 102 and executed by the processor 101. The auxiliary storage device 103 also stores an OS. At least a part of the OS is loaded into the memory 102 and executed by the processor 101. The processor 101 executes a machine learning program while running the OS. OS is an abbreviation for Operating System.

[0022] Input and output data of the machine learning program is stored in the storage unit 120. The memory 102 functions as the storage unit 120. However, a storage device such as the auxiliary storage device 103, a register in the processor 101, or a cache memory in the processor 101 may function as the storage unit 120 instead of or together with the memory 102.

[0023] The machine learning program can be recorded (stored) in a computer-readable manner on a non-volatile recording medium such as an optical disk or flash memory.

[0024] ***Explanation of Operation*** The operation procedure of the machine learning device 100 corresponds to a machine learning method, and also corresponds to a processing procedure by a machine learning program.

[0025] The machine learning method will be explained based on FIG. In step S101, the input unit 111 acquires input data. The input data is data that is input to the machine learning model. For example, the input data is stored in advance in the storage unit 190. The machine learning model is a model (trained model) generated by machine learning. The machine learning model is stored in advance in the storage unit 190. An example of a machine learning model is a large-scale language model, in which text data is used as input data.

[0026] Steps S102 to S104 are repeatedly executed. Steps S102 to S104 are executed by the switching unit 112 and the inference unit 113. The switching unit 112 switches the model parameter set set in the machine learning model using a plurality of model parameter sets. A model parameter set is one or more parameters (model parameters) used in a machine learning model. For example, the switching unit 112 generates a new model parameter set by adding a new model parameter to the current model parameter set of the machine learning model, and then uses the current model parameter set and the new model parameter set as multiple parameter sets. For example, for each new model parameter included in the two or more new model parameters, the switching unit 112 generates a new model parameter set by adding the new model parameter to the current model parameter set of the machine learning model.Then, the switching unit 112 uses the two or more new model parameter sets corresponding to the two or more new model parameters as the multiple parameter sets. For each model parameter set, the inference unit 113 performs inference on the input data using the machine learning model to which the model parameter set is set. As a result, output data is obtained. The output data indicates the inference results by the machine learning model when the model parameters are set.

[0027] The procedure from step S102 to step S104 will be described. In step S102, the switching unit 112 sets the model parameter set to the machine learning model. Each time step S102 is executed, some of the model parameter sets are different.

[0028] In step S103, the inference unit 113 performs inference on the input data using the machine learning model, thereby obtaining output data. When the machine learning model is a large-scale language model, text data is the output data. In this case, the output text (output data) indicates a token sequence corresponding to the input text (input data). The token sequence consists of one or more tokens. A token corresponds to, for example, a word.

[0029] In step S104, the switching unit 112 determines whether to switch the model parameter set set in the machine learning model. For example, if the number of times the model parameter set has been switched has not reached a predetermined number of times, the switching unit 112 determines to switch the model parameter set. If the model parameter set to be set in the machine learning model is to be switched, the process proceeds to step S102. If the model parameter set set in the machine learning model is not to be switched, the process proceeds to step S105.

[0030] Steps S102 to S104 result in obtaining a plurality of output data corresponding to a plurality of model parameter sets.

[0031] In step S105, the comparison unit 114 compares the characteristics of the inference results between the plurality of pieces of output data, thereby obtaining a comparison result. Examples of features to be compared are probability transitions, probability distributions, or cosine similarities of inference results, which are features related to information leakage. When the machine learning model is a large-scale language model, for example, probability transitions between tokens in a token sequence shown in an output text are compared between the output texts. Then, abnormal probability transitions are detected. For example, an abnormal probability transition is a state that is unlikely to occur under normal operation. An abnormal probability transition can be detected by evaluating whether the magnitude of the possibility (probability) of the transition is lower than a predetermined threshold.

[0032] In step S106, the output unit 115 selects one of the plurality of output data. For example, when the plurality of output data are two output data corresponding to the current model parameter set and a new model parameter set, the output unit 115 selects the output data corresponding to the new model parameter set.

[0033] Next, the output unit 115 determines, based on the comparison result, information relating to information leakage from among the information included in the inference result shown in the selected output data. For example, the output unit 115 determines that a portion of the token sequence shown in the output text in which an abnormal probability transition is detected is information related to information leakage.

[0034] Next, the output unit 115 applies the changes to the information related to information leakage to the selected output data.

[0035] For example, the output unit 115 modifies the output data in one of the following ways. The output unit 115 eliminates (deletes) information relating to information leakage. The output unit 115 replaces the information about information leakage with information whose features are suppressed. The output unit 115 integrates (adds) information about information leakage of other output data to information about information leakage of the selected output data.

[0036] When the machine learning model is a large-scale language model, for example, tokens in parts where abnormal probability transitions are detected are replaced with similar tokens that suppress those characteristics. Tokens in parts where abnormal probability transitions are detected are sensitive information for additional learning (additional parameters). For example, if the sensitive information is an individual's name, it is replaced with a similar token such as a fictitious name that does not actually exist. In this way, the characteristics of sensitive information can be eliminated from the output data, thereby preventing information leakage from the machine learning model.

[0037] Then, the output unit 115 outputs the output data after the change has been made. For example, the output unit 115 stores the output data after the change in the storage unit 190 .

[0038] ***Description of Example*** An example of comparing output features with and without PEFT parameter settings for a large-scale language model will be described with reference to FIG. Large-scale language models are pre-trained machine learning models. PEFT does not update the parameters of pre-trained machine learning models, so the model parameters of large-scale language models can be switched by adding or not adding parameters. In step S101, the input unit 111 acquires the input text 121. In step S102, the switching unit 112 adds the PEFT parameters and sets the model parameter set to the large-scale language model. The large-scale language model with the PEFT parameters added is referred to as large-scale language model 122. In step S103, the inference unit 113 performs inference on the input text 121 using the large-scale language model 122. As a result, an output text is obtained. In step S102, the switching unit 112 sets the model parameter set to the large-scale language model without adding the PEFT parameters. The large-scale language model without adding the PEFT parameters is referred to as the large-scale language model 123. In step S103, the inference unit 113 performs inference on the input text 121 using the large-scale language model 123. As a result, an output text is obtained. In step S105, the comparison unit 114 compares the predicted probabilities of word strings (token strings) between the two output texts. In step S106, if the transitions in the predicted probabilities of word strings differ between output texts, the output unit 115 replaces the target token in the output text of the large-scale language model 122 with a similar token to reduce the difference in the transitions in the predicted probabilities. The output text of the large-scale language model 122 after the replacement is referred to as output text 124. The output unit 115 outputs the output text 124.

[0039] ***Effects of the First Embodiment*** The purpose of the first embodiment is to make it easier to apply information leakage countermeasures for fine-tuned machine learning models (large-scale language models) to existing fine-tuning methods without degrading model performance. The first embodiment can realize a machine learning device 100, a machine learning method, and a machine learning program that prevent information about learning data from leaking when inferring certain input data. In the first embodiment, output data is compared when switching additional parameters for a fine-tuned (PEFT) machine learning model, and features related to information leakage are eliminated or suppressed. The main effect of the first embodiment is to prevent information leakage from the machine learning model. The first embodiment consists of switching additional parameters and comparing inference results. Therefore, the first embodiment can be implemented without manipulating model learning, and the first embodiment results in less degradation of model performance compared to existing methods. Adding the configuration of parameter switching and output comparison can be easily applied to existing fine-tuning methods.

[0040] Embodiment 2 The following describes a mode in which a machine learning model is retrained using output data obtained in the first embodiment, focusing mainly on the differences from the first embodiment, with reference to FIGS. 4 to 6.

[0041] ***Configuration Description*** The configuration of the machine learning device 100 will be described with reference to FIG. The machine learning device 100 further includes a re-learning unit 116. The machine learning program also causes the computer to function as a re-learning unit 116 .

[0042] ***Explanation of Operation*** The machine learning method will be explained based on FIG. Steps S101 to S106 are the same as those described in the first embodiment.

[0043] In step S107, the re-learning unit 116 uses the input data obtained in step S101 as learning data and the changed output data obtained in step S106 as correct answer data to re-learn the machine learning model.

[0044] For example, the re-learning unit 116 performs fine tuning by using the training data set (training data and correct answer data) to update additional parameters (part of the parameters of the machine learning model).

[0045] ***Description of Example*** An example of comparing output features for different PEFT parameter settings for a large-scale language model will be described with reference to FIG. In step S101, the input unit 111 acquires the input text 125. The input text 125 is text data collected for training purposes. In step S102, the switching unit 112 adds the PEFT parameter (1) and sets the model parameter set to the large-scale language model. The large-scale language model to which the PEFT parameter (1) has been added is referred to as the large-scale language model 126. In step S103, the inference unit 113 performs inference on the input text 125 using the large-scale language model 126. As a result, an output text is obtained. In step S102, the switching unit 112 adds the PEFT parameter (2) and sets the model parameter set to the large-scale language model. The large-scale language model to which the PEFT parameter (2) is added is referred to as the large-scale language model 127. In step S103, the inference unit 113 performs inference on the input text 125 using the large-scale language model 127. As a result, an output text is obtained. In step S105, the comparison unit 114 compares the predicted probabilities of word strings (token strings) between the two output texts. In step S106, the output unit 115 modifies the target token of the output text of the large-scale language model 126 or the large-scale language model 127 when the transition of the predicted probability of the word string differs between the output texts. For example, the output unit 115 integrates the target token of the output text of the large-scale language model 127 with the target token of the output text of the large-scale language model 126. Alternatively, the output unit 115 replaces the target token of the output text of the large-scale language model 126 with the target token of the output text of the large-scale language model 127. The modified output text is referred to as output text 128. In step S107, the re-training unit 116 creates a training data set (training data and supervised data) using the input text 125 as training data and the output text 128 as supervised data. Then, the re-training unit 116 re-trains the large-scale language model by updating the PEFT parameters (1) and (2) using the training data set.

[0046] ***Effects of the Second Embodiment*** The second embodiment aims to generate a machine learning model that has high accuracy in output data and does not leak information. The second embodiment has a configuration for switching between multiple additional parameters and a configuration for retraining a machine learning model using output data whose features have been compared. The second embodiment can realize a machine learning device 100, a machine learning method, and a machine learning program that train a machine learning model that does not leak information about training data when inferring certain input data. In the second embodiment, output data is compared when switching additional parameters for a fine-tuned (PEFT) machine learning model, and data in which features related to information leakage have been eliminated, suppressed, or integrated is used as new ground truth data. The main effect of the second embodiment is to generate a machine learning model that does not leak information by retraining a fine-tuned (PEFT) machine learning model. The second embodiment is configured by adding a configuration for parameter switching and output comparison. Therefore, the second embodiment can be easily applied to existing fine-tuning methods. Furthermore, the second embodiment generates a machine learning model that does not leak information. Therefore, when applying the machine learning model to a product, it is not necessary to implement additional functions for preventing information leakage.

[0047] ***Supplementary explanation of implementation form*** The type of machine learning operation is not limited to deep learning, but may be an operation such as regression, decision tree learning, Bayesian method, or clustering.

[0048] The hardware configuration of the machine learning device 100 will be described with reference to FIG. The machine learning device 100 includes a processing circuit 109 . The processing circuit 109 is hardware that realizes an input unit 111, a switching unit 112, an inference unit 113, a comparison unit 114, an output unit 115, and a re-learning unit . The processing circuitry 109 may be dedicated hardware, or may be a processor 101 that executes a program stored in memory 102 .

[0049] When processing circuitry 109 is dedicated hardware, processing circuitry 109 may be, for example, a single circuit, a multiple circuit, a programmed processor, a parallel programmed processor, an ASIC, an FPGA, or a combination thereof. ASIC is an abbreviation for Application Specific Integrated Circuit. FPGA is an abbreviation for Field Programmable Gate Array.

[0050] The machine learning device 100 may include multiple processing circuits that replace the processing circuit 109.

[0051] In the processing circuit 109, some functions may be realized by dedicated hardware, and the remaining functions may be realized by software or firmware.

[0052] In this way, the functions of the machine learning device 100 can be realized by hardware, software, firmware, or a combination of these.

[0053] Each embodiment is an example of a preferred embodiment and is not intended to limit the technical scope of the present disclosure. Each embodiment may be implemented in part or in combination with other embodiments. Procedures described using flowcharts, etc. may be modified as appropriate.

[0054] The "part" of each element of the machine learning device 100 may be read as a "process," a "step," a "circuit," or a "circuitry." [Explanation of symbols]

[0055] 100 machine learning device, 101 processor, 102 memory, 103 auxiliary storage device, 104 input / output interface, 109 processing circuit, 111 input unit, 112 switching unit, 113 inference unit, 114 comparison unit, 115 output unit, 116 retraining unit, 121 input text, 122 large-scale language model, 123 large-scale language model, 124 output text, 125 input text, 126 large-scale language model, 127 large-scale language model, 128 output text, 190 memory unit.

Claims

1. an inference unit that, for each model parameter set included in a plurality of different model parameter sets, performs inference on input data using a machine learning model to which the model parameter set is set, and obtains output data indicating an inference result by the machine learning model when the model parameter set is set; a comparison unit that compares features of the inference results between output data for a plurality of output data corresponding to the plurality of model parameter sets and obtains a comparison result; an output unit that determines information regarding information leakage from among the information included in the inference results shown in any of the output data based on the comparison result, modifies the output data with respect to the information regarding information leakage, and outputs the modified output data; a switching unit that generates a new model parameter set by adding new model parameters to a current model parameter set of the machine learning model, and switches between model parameter sets that are set in the machine learning model by using the current model parameter set and the new model parameter set as the plurality of model parameter sets; A machine learning device comprising:

2. an inference unit that, for each model parameter set included in a plurality of different model parameter sets, performs inference on input data using a machine learning model to which the model parameter set is set, and obtains output data indicating an inference result by the machine learning model when the model parameter set is set; a comparison unit that compares features of the inference results between output data for a plurality of output data corresponding to the plurality of model parameter sets and obtains a comparison result; an output unit that determines information regarding information leakage from among the information included in the inference results shown in any of the output data based on the comparison result, modifies the output data with respect to the information regarding information leakage, and outputs the modified output data; a switching unit that generates a new model parameter set by adding, for each new model parameter included in the two or more new model parameters, the new model parameters to a current model parameter set of the machine learning model, and switches the model parameter set to be set in the machine learning model by using the two or more new model parameter sets corresponding to the two or more new model parameters as the multiple model parameter sets; A machine learning device comprising:

3. an inference unit that, for each model parameter set included in a plurality of different model parameter sets, performs inference on input data using a machine learning model to which the model parameter set is set, and obtains output data indicating an inference result by the machine learning model when the model parameter set is set; a comparison unit that compares features of the inference results between output data for a plurality of output data corresponding to the plurality of model parameter sets and obtains a comparison result; an output unit that determines information regarding information leakage from among the information included in the inference results shown in any of the output data based on the comparison result, modifies the output data with respect to the information regarding information leakage, and outputs the modified output data; a re-learning unit that uses the input data as training data and the output data after the change as correct answer data to re-learn the machine learning model; A machine learning device comprising:

4. The comparison unit compares the probability transition of the inference result between output data as the feature. The machine learning device according to claim 1 .

5. The comparison unit compares the probability distribution of the inference results between the output data as the features. The machine learning device according to claim 1 .

6. The comparison unit compares the cosine similarity of the inference results between the output data as the feature. The machine learning device according to claim 1 .

7. The output unit performs one of exclusion, suppression, and integration on the output data as a change to the information related to the information leakage. The machine learning device according to claim 1 .

8. A machine learning device, For each model parameter set included in a plurality of mutually different model parameter sets, inference is performed on input data using a machine learning model to which the model parameter set is set, and output data indicating an inference result by the machine learning model when the model parameter set is set is obtained; comparing features of the inference results between output data for a plurality of output data corresponding to the plurality of model parameter sets, and obtaining comparison results; Determine information regarding information leakage from among the information included in the inference results shown in any of the output data based on the comparison result, modify the output data with the information regarding information leakage, and output the modified output data.

1. A machine learning method comprising: The machine learning device includes: generating a new model parameter set by adding a new model parameter to a current model parameter set of the machine learning model, and switching the model parameter set set for the machine learning model by using the current model parameter set and the new model parameter set as the plurality of model parameter sets; Machine learning methods.

9. A machine learning device, For each model parameter set included in a plurality of mutually different model parameter sets, inference is performed on input data using a machine learning model to which the model parameter set is set, and output data indicating an inference result by the machine learning model when the model parameter set is set is obtained; comparing features of the inference results between output data for a plurality of output data corresponding to the plurality of model parameter sets, and obtaining comparison results; Determine information regarding information leakage from among the information included in the inference results shown in any of the output data based on the comparison result, modify the output data with the information regarding information leakage, and output the modified output data.

1. A machine learning method comprising: The machine learning device includes: generating a new model parameter set by adding, for each new model parameter included in the two or more new model parameters, the new model parameter to a current model parameter set of the machine learning model, and switching the model parameter set to be set in the machine learning model by using, as the plurality of model parameter sets, the two or more new model parameter sets corresponding to the two or more new model parameters; Machine learning methods.

10. A machine learning device, For each model parameter set included in a plurality of mutually different model parameter sets, inference is performed on input data using a machine learning model to which the model parameter set is set, and output data indicating an inference result by the machine learning model when the model parameter set is set is obtained; comparing features of the inference results between output data for a plurality of output data corresponding to the plurality of model parameter sets, and obtaining comparison results; Determine information regarding information leakage from among the information included in the inference results shown in any of the output data based on the comparison result, modify the output data with the information regarding information leakage, and output the modified output data.

1. A machine learning method comprising: The machine learning device includes: Re-training the machine learning model using the input data as training data and the output data after the change as correct answer data. Machine learning methods.

11. an inference process for performing inference on input data using a machine learning model to which each model parameter set included in a plurality of mutually different model parameter sets is set, and obtaining output data indicating an inference result by the machine learning model when each model parameter set is set; a comparison process for comparing features of the inference results between output data for a plurality of output data corresponding to the plurality of model parameter sets, and obtaining a comparison result; an output process of determining information regarding information leakage from among the information included in the inference results shown in any of the output data based on the comparison result, modifying the output data with respect to the information regarding information leakage, and outputting the modified output data; a switching process of generating a new model parameter set by adding a new model parameter to a current model parameter set of the machine learning model, and switching the model parameter set set to the machine learning model using the current model parameter set and the new model parameter set as the plurality of model parameter sets; A machine learning program that allows a computer to execute the following.

12. an inference process for performing inference on input data using a machine learning model to which each model parameter set included in a plurality of mutually different model parameter sets is set, and obtaining output data indicating an inference result by the machine learning model when each model parameter set is set; a comparison process for comparing features of the inference results between output data for a plurality of output data corresponding to the plurality of model parameter sets, and obtaining a comparison result; an output process of determining information regarding information leakage from among the information included in the inference results shown in any of the output data based on the comparison result, modifying the output data with respect to the information regarding information leakage, and outputting the modified output data; a switching process of generating a new model parameter set by adding, for each new model parameter included in two or more new model parameters, the new model parameters to a current model parameter set of the machine learning model, and switching the model parameter set set to the machine learning model by using the two or more new model parameter sets corresponding to the two or more new model parameters as the multiple model parameter sets; A machine learning program that allows a computer to execute the following.

13. an inference process for performing inference on input data using a machine learning model to which each model parameter set included in a plurality of mutually different model parameter sets is set, and obtaining output data indicating an inference result by the machine learning model when each model parameter set is set; Regarding a plurality of output data corresponding to the plurality of model parameter sets, a comparison process for comparing the characteristics of the inference results to obtain a comparison result; an output process of determining information regarding information leakage from among the information included in the inference results shown in any of the output data based on the comparison result, modifying the output data with respect to the information regarding information leakage, and outputting the modified output data; a re-learning process that uses the input data as learning data and the output data after the change as correct answer data to re-learn the machine learning model; A machine learning program that allows a computer to execute the following.

Images