Skip to content
AN
☰ Menu
Independent Research Initiative
AIADEEB INTELLIGENCE

Researching how intelligence becomes action.

Adeeb Intelligence is an independent research initiative led by Adeeb Noor. We study how intelligent systems turn evidence into consequential decisions — and build methods that make those decisions measurable, auditable and human-accountable.

Mission →Research thrusts →Systems →Work with us →
Mission

Make consequential intelligence measurable, auditable and accountable.

Our work starts where a model output becomes an allocation, clinical alert, engineering action or institutional decision. We focus on the gap between what a system scores and what it actually changes.

Reveal hidden differencesDevelop evaluation methods that expose action-level divergence hidden by aggregate performance.
Strengthen evidence integrityTrace what evidence a system used, what it missed and when that changes a decision.
Translate research into useTurn methods into reproducible tools, public systems, applied studies and responsibly protected IP.
Research thrusts

Three connected questions.

The domains differ; the research logic is the same: evaluate the decision chain, not only the model output.

01 · DECISION INTELLIGENCE

AI Evaluation & Decision Systems

Question: When do apparently equivalent systems act differently?

Selection identity · execution provenance · evaluator robustness · consequential evaluation.Explore RIDI →
02 · BIOMEDICAL EVIDENCE

Biomedical Knowledge & Drug Safety

Question: How does incomplete or misrepresented evidence alter clinical prioritisation?

Drug identity · DDI completeness · clinical relevance · negative controls · evidence grading.Open ANTI-DDI →
03 · HUMAN JUDGMENT

Generative AI, Judgment & Education

Question: How can AI scale learning and design without erasing context, evidence or human judgment?

Defensible judgment · fidelity debt · cultural calibration · computing education.Teaching systems →
How we work

Measure. Trace. Stress. Translate.

We use a simple research discipline across domains.

01Measure the consequential difference

Look beyond average performance to the items, evidence or people whose outcomes actually change.

02Trace the decision state

Preserve provenance, assumptions, dependencies and evidence so the result can be inspected.

03Stress the claim

Change the condition, boundary or dependency and observe whether the decision still holds.

04Translate the finding

Release the result as a paper, reproducible artifact, public system, institutional pilot or protected invention.

Research systems

Ideas are stronger when they can be inspected.

Selected systems and artifacts that make the research operational.

Abstract visual for decision intelligence research
OPEN RESEARCH

RIDI

Audits selection change and the identity–utility frontier when ranking systems become decisions.

Interactive demo →
Clinical AI visual for HealthX
IP TRANSLATION

HealthX

Governed AI middleware for auditable, policy-aware clinical-IP workflows.

Institutional IP reviewPublic research page →
MIYAR workforce intelligence interface
IP TRANSLATION

MIYAR

Governed workforce intelligence with occupational-taxonomy compliance and auditable decision trails.

Institutional IP reviewPublic prototype →
iSCARB teaching system interface
OPEN EDUCATION RESEARCH

iSCARB

A course-level system for defensible engineering judgment in the generative-AI era.

Explore the system →
People

Led by Adeeb Noor.

Professor of Data Science & AI at King Abdulaziz University. Projects bring together students, co-authors, domain experts and institutional collaborators according to the research question.

Independent initiativeAdeeb Intelligence is a personal research initiative led by Adeeb Noor and is not presented as an official organizational unit of King Abdulaziz University.
Work with Adeeb Intelligence

Bring a hard question and evidence worth testing.

We are most interested in collaborations where the problem, evidence source and expected contribution are explicit.

Researchers & facultyJoint studies, replication, methodological extensions and cross-institutional validation.
StudentsTheses and projects in AI evaluation, biomedical informatics, computing education and research software.
Institutions & industryReal decision problems, domain validation, data access and applied research pilots.
When you contact us, include:the question you want to test · the evidence or data you can access · the contribution you can make.

Research should survive outside the paper.

Explore the research record, reproduce an artifact, or propose a collaboration.

Research agenda →Publications →Contact →
Privacy & analytics preferences