03 / Experiments

Reproducibility before conclusions.

AegisAI emphasizes comparisons that can be inspected and repeated. The experimental program is being defined; no benchmark outcomes are reported here.

Comparison framework

One question. Four model definitions.

Implemented baselines do not imply completed comparative experiments.

Authorization models and implementation status
ModelAuthorization approachStatus
Model ARBACImplemented
Model BRBAC + ABACImplemented
Model CRBAC + deterministic contextual riskImplemented
Model DRBAC + ABAC + behavioral adaptive riskPlanned Research
Planned experimental program

Making the work repeatable.

01Planned Research

Scenario Generator

Future work: define repeatable access scenarios with documented identities, resource context and behavioral conditions.

02Planned Research

Attack Simulation

Future work: specify controlled abnormal-access and operational disruption scenarios with explicit ground truth.

03Planned Research

Benchmark Methodology

Future work: document model configurations, controlled conditions, measurement procedures and evaluation limitations.

04Planned Research

Datasets

Future work: publish dataset descriptions, provenance, access conditions and reproducible preparation steps when available.

Results

Experimental results in progress.

Comparative results and supporting artifacts will be published after evaluation. There are no published measurements on this portal yet.