Mandatory Access Control Policy Generation via Machine Learning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing operating systems lack the ability to automatically generate mandatory access control policies for new applications, leading to potential security threats and resource wastage due to the execution of unnecessary files.

Innovation Solution

A method utilizing heuristic and machine learning models to identify required and unnecessary files within an application, generating rules to block unnecessary files and create a mandatory access control policy, which is then implemented by the operating system.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional operating systems are used without automated access control policy generation, then device complexity remains low and ease of operation is maintained, but security reliability is compromised and computing resources are wasted due to execution of unnecessary files

Engineering Contradiction:
ImprovesecurityVSAvoidaccess control policy generation
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system enables self-service by automatically generating mandatory access control policies through machine learning models that analyze application behavior and file requirements, eliminating the need for manual security policy configuration while enhancing security reliability

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system changes the parameter of policy generation from static manual configuration to dynamic automated generation using heuristic and machine learning models, transforming how access control policies are created and applied

Inventive Principle:
Principle #35Parameter changes

2Extent of automation

If manual access control policy generation is used, then automation extent is low, but manufacturing precision of security policies can be maintained through expert analysis, whereas automated generation without ML models would reduce precision and increase errors

Engineering Contradiction:
Improvepolicy generation automationVSAvoidpolicy accuracy
Core Design Contradiction:
Extent of automationVSManufacturing precision

Solution Approach 1:

The system replaces manual expert analysis (mechanical human process) with automated machine learning models that process application behavior data, achieving both high automation extent and maintained precision through algorithmic analysis of file execution patterns

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system implements feedback mechanisms where the machine learning models continuously learn from observed application behavior and file execution patterns, improving policy generation accuracy through iterative optimization based on actual system performance

Inventive Principle:
Principle #23Feedback

3Productivity

If all files in an application are allowed to execute, then ease of operation is maintained and security policy configuration is simple, but computing resources are wasted and security threats increase due to execution of unnecessary files

Engineering Contradiction:
Improveresource efficiencyVSAvoidapplication execution simplicity
Core Design Contradiction:
ProductivityVSEase of operation

Solution Approach 1:

The system extracts and identifies unnecessary files from applications using machine learning analysis of file execution patterns, separating required files from unnecessary ones to enable selective blocking while maintaining application functionality and improving resource efficiency

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3640826B1Utilizing heuristic and machine learning models to generate a mandatory access control policy for an application
Publication Date: 2021.04.21 ACCENTURE GLOBAL SOLUTIONS LTD
  • EP3640826B1 patent drawingFigure 1A
  • EP3640826B1 patent drawingFigure 1B
  • EP3640826B1 patent drawingFigure 1C

AI summary

A device processes, with a model, an application to identify a set of file paths with process identifiers. The device identifies patterns associated with the set of file paths with process identifiers, and determines positions of random elements in each file path of the set of file paths with process identifiers. The device processes the patterns and the positions of the random elements to train a machine learning model, and utilizes the machine learning model to generate a first set of rules to identify files required for execution of the application, and a second set of rules to identify files not required for execution of the application. The device generates a mandatory access control policy based on the first set of rules and the second set of rules, and provides the mandatory access control policy to be implemented by an operating system of a client device.