Policy consistency verification apparatus, policy consistency verification method, and policy consistency verification program

1 . A policy consistency verification apparatus, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor,

wherein the at least one processor is configured to execute program instructions stored in the memory to:

generate an If-Then rule approximating a policy function used as an input, wherein the policy function is generated from a policy definition defining an attribute value of access and a rule regarding accessibility to obtain an output of accessibility from inputs of the attribute value; and

verify consistency between the If-Then rule and the policy definition to generate an inconsistent attribute value as an error sample;

wherein generating the If-Then rule comprises:

randomly generating a combination of attribute values;

configuring a decision tree from outputs obtained by inputting the combination of attribute values into the policy function; and

generating the If-Then rule from the decision tree.

2 . The policy consistency verification apparatus according to claim 1 , wherein the at least one processor is further configured to execute the program instructions to:

obtain the error sample by performing satisfiability check on a logical formula generated based on the If-Then rule and the policy definition.

3 . The policy consistency verification apparatus according to claim 1 , wherein the at least one processor is further configured to execute the program instructions to:

generate the policy function as a machine learning model based on the policy definition;

generate learning data based on the error sample; and

update the machine learning model.

4 . The policy consistency verification apparatus according to claim 3 , wherein the at least one processor is further configured to execute the program instructions to:

update the If-Then rule based on the error sample.

5 . The policy consistency verification apparatus according to claim 4 , wherein the at least one processor is further configured to execute the program instructions to:

determine whether the error sample is an error caused by the policy function or an error caused by the If-Then rule; and

judge, based on a result of the determination, whether the policy function is updated or the If-Then rule is updated.

6 . A policy consistency verification method, comprising:

converting, by at least one processor, a policy for generating an If-Then rule;

generating, by the at least one processor, the If-Then rule approximating a policy function used as an input, wherein the policy function is generated from policy definition defining an attribute value of access and a rule regarding accessibility to obtain an output of accessibility from inputs of the attribute value, and wherein the If-Then rule is generated from a decision tree configured from outputs obtained by inputting a randomly generated combination of attribute values into the policy function; and

generating, by the at least one processor, an error sample for verifying consistency between the If-Then rule and the policy definition, and generating based on the verification, an inconsistent attribute value as the error sample.

7 . The policy consistency verification method according to claim 6 , comprising:

determining whether the error sample is an error caused by the policy function or an error caused by the If-Then rule;

in a case where the error is caused by the policy function, updating a policy generation part that generates training data based on the error sample to generate the policy function as a machine learning model based on the policy definition; and

in a case where the error is caused by the If-Then rule, updating the If-Then rule based on the error sample.

8 . The policy consistency verification method according to claim 6 , wherein

obtaining the error sample by performing a satisfiability check on a logical formula generated based on the If-Then rule and the policy definition.

9 . A non-transitory computer readable medium storing a policy consistency verification program that causes a computer to verify a policy consistency, the computer comprising at least one processor that executes a program, and a memory that stores the program, the program comprising:

a policy conversion process configured to generate an If-Then rule using a policy function as an input to approximate the policy function, wherein the policy function is generated from a policy definition defining a rule regarding an attribute value of an access and an accessibility, and wherein the If-Then rule is generated from a decision tree configured from outputs obtained by inputting a randomly generated combination of attribute values into the policy function; and

an error sample generation process configured to verify consistency between the If-Then rule and the policy definition to generate an inconsistent attribute value as an error sample.

10 . The non-transitory computer readable medium storing the policy consistency verification program according to claim 9 , comprising:

determining whether the error sample is an error caused by the policy function or an error caused by the If-Then rule;

in a case where the error is caused by the policy function, generating training data based on the error sample;

updating, based on the error sample, a machine learning model configured to generate the policy function based on the policy definition; and

in a case where the error is caused by the If-Then rule, updating the If-Then rule based on the error sample.

11 . The non-transitory computer readable medium storing the policy consistency verification program according to claim 9 , wherein

obtaining the error sample by performing a satisfiability check on a logical formula generated based on the If-Then rule and the policy definition.