Agent-Based Semantic Role Mining for Intelligent Access Control in Multi-Domain Collaborative Applications of Smart...

Area of Science:

• Computer Science
• Artificial Intelligence
• Information Security

Background:

• Traditional Role-Based Access Control (RBAC) struggles with dynamic, multi-domain smart city environments due to limitations in adapting to changing user, task, and resource information.
• Existing RBAC models lack semantically meaningful business roles, hindering effective access decisions in complex collaborative settings.

Purpose of the Study:

• To propose an Intelligent Role-based Access Control (I-RBAC) model for dynamic multi-domain smart city environments.
• To leverage intelligent software agents and semantic ontologies for enhanced access control capabilities.

Main Methods:

• Developed a core I-RBAC ontology using Standard Occupational Classification (SOC) semantic business roles.
• Employed intelligent agents with word embedding and bidirectional LSTM deep neural networks for semantic role mining.
• Automated organizational ontology population from unstructured text policies and matched it with the core I-RBAC ontology.

Main Results:

• Achieved accurate extraction of semantically meaningful business roles.
• Demonstrated high accuracy in deriving RDF triples (Subject, Predicate, Object) from organizational text policies.
• Validated the model's effectiveness through large-scale collaboration case scenarios across five multi-domain organizations.

Conclusions:

• The I-RBAC model effectively addresses the challenges of access control in dynamic, multi-domain smart city environments.
• Intelligent agents and semantic role mining significantly improve the accuracy and adaptability of access control systems.
• The proposed ontology-driven approach provides a robust framework for unified business role extraction and access management.