MEHRAN  HOSSEINI

I'm a postdoctoral researcher at King's College London's Informatics Department and a founder of NaPolleon, an AI-powered survey and prediction market backed by Oxford University Innovation. My research focuses on developing mathematical foundations for safe and verifiable AI systems, with particular emphasis on scalable verification tools and novel deep learning architectures that are more capable and amenable to formal verification.

At the KCL, I lead the research on safety verification of autonomous neural agents as part of the EU Horizon's REXASI-PRO project, collaborating with Nicola Paoletti and DDV Lab. Before joining KCL, I was at Imperial College London's Safe Artificial Intelligence Lab (SAIL) collaborating with Alessio Lomuscio on scalable verification of recurrent neural nets. Prior to this, I completed my DPhil in computer science at Oxford University, where I solved the universal termination problem for linear loop programs. At Oxford, I was lucky to have James Worrell and Joël Ouaknine as my supervisors.

                 

Avatar

Selected Publications (* denotes first/co-first author)

Gymnasium's Lunar Lander

LTL Verification of Memoryful Neural Agents

Mehran Hosseini*, Alessio Lomuscio, Nicola Paoletti - AAMAS 2025 - PDF coming soon.

Certified City Tranjectories

Certified Guidance for Planning with Deep Generative Models

Francesca Cairoli, Francesco Giacomarra, Mehran Hosseini, Nicola Paoletti - AAMAS 2025 - PDF coming soon.

Attention Mechanism

You Need to Pay Better Attention: Rethinking the Mathematics of Attention Mechanism

Mehran Hosseini*, Peyman Hosseini - arXiv Preprint 2024 - PDF.

VRCP Coverage

Verifiably Robust Conformal Prediction

Linus Jeary, Tom Kuipers, Mehran Hosseini, Nicola Paoletti - NeurIPS 2024 - PDF.

GeoConv Coordinate Channel

GeoPos: A Minimal Positional Encoding for Enhanced Fine-Grained Details in Image Synthesis Using Convolutional Neural Networks

Mehran Hosseini*, Peyman Hosseini - WACV 2025 - PDF.

Bayesian Neural Network

Tight Verification of Probabilistic Robustness in Bayesian Neural Networks

Ben Batten, Mehran Hosseini*, Alessio Lomuscio - AISTATS 2024 - PDF.

Gymnasium's Lunar Lander

Bounded and Unbounded Verification of RNN-Based Agents in Non-deterministic Environments

Mehran Hosseini*, Alessio Lomuscio - AAMAS 2023 - PDF.

Polyhedron Constraint

On Termination and Divergence of Linear Programs

Mehran Hosseini* - University of Oxford 2021 - PDF.

Polyhedron Constraint

Termination of Linear Loops over the Integers

Mehran Hosseini*, Joël Ouaknine, James Worrell - ICALP 2019 - PDF.

Fibonacci Spiral

Effective Divergence Analysis for Linear Recurrence Sequences

Shaull Almagor, Brynmor Chapman, Mehran Hosseini*, Joël Ouaknine, James Worrell - CONCUR 2018 - PDF.

Fibonacci Spiral

Isomorphism Classes of Doche-Icart-Kohel Curves over Finite Fields

Mehran Hosseini*, Reza Rezaeian Farashahi - Finite Fields and Their Applications 2016 - PDF.

Tools & Models

Vern, Verifier of Neural Nets

Vern, Verifier of Neural Nets

Vern is verifier for neural networks which supports memroyful models and LTL specifications.

Hummingbird Language Model

Hummingbird Language Model

Hummingbird is a proof-of-concept small & efficient language model based on Efficient Attention, trained on only 15B token.

Teaching, Leadership, & Service

Imperial College London

Imperial College London

Service

  • Organiser: VAS group's weekly seminars
Oxford University

University of Oxford

Teaching

  • Computational Complexity (Graduate)
  • Foundations of Computer Science (Graduate)

Leadership & Service

  • President of Oxford University Iranian Society (2018-2020)
  • President of Linacre College's tennis and pool societies (2018-2020)
  • Undergraduate Admission, Trinity College, Oxford
Isfahan University of Technology

Isfahan University of Technology

Lecturing

  • Calculus I Programming with Maple (Undergraduate)

Teaching

  • Cryptography (Graduate)
  • Computational Geometry (Graduate)
  • Programming with Maple (Graduate)
  • Programming with Maple (Undergraduate)
  • Mathematical Analysis I & II (Undergraduate)