Open to collaboration · Safe Autonomy · WMG, Warwick
I work on the measurement problem at the centre of AI safety assurance: how to tell whether a system's behaviour in simulation is evidence about its behaviour in the world — and how to make that claim defensible.
Coventry, United Kingdom
danial.safaei@warwick.ac.uk
Decisive Feature Fidelity · IEEE IV 2026
Same output, different reason. A real frame and its synthetic twin can produce identical outputs from the same frozen model while relying on different visual evidence. DFF makes that divergence measurable — so fidelity is judged by why the system decides, not how the images look.
Bibliometrics · auto-updated
Refreshed weekly from Google Scholar · Aug 2026
Featured · peer-reviewed
Supervision & partnership
PhD supervised by Prof. Siddartha Khastgir and Prof. Matthew Higgins, WMG. Previously co-supervised by Dr Xingyu Zhao.
Industrial research partnership with Siemens Digital Industries Software since February 2025.
Before Warwick: MSc Artificial Intelligence, University of Tehran (GPA 19.41/20).
Peer-reviewed work first, preprints separately. Author order as published. This list renders from a single data file — one edit updates the whole site.
Siemens Digital Industries Software
Industrial research partnership running alongside the doctorate, connecting simulation-fidelity research to production virtual-validation practice.
WMG, University of Warwick
Postgraduate Society, University of Warwick
Elected representative for postgraduate researchers on academic matters.
Safe Autonomy Research Group, WMG, University of Warwick
Doctoral research on safety assurance and trustworthy evaluation for autonomous systems.
University of Tehran
Research across deep learning, reinforcement and inverse reinforcement learning, computer vision, and graph-based methods.
WMG, University of Warwick, United Kingdom
University of Tehran, Iran
GPA 19.41 / 20.0.
University of Kashan, Iran
Live from the GitHub API — stars, language, and last activity update themselves.
I am glad to hear from people working on simulation-to-real transfer, validation methodology, or safety evidence for autonomous systems.