Algorithms
Design and analysis of algorithms, with attention to structure, efficiency, and computational complexity.
Researcher at SLATE
Sam Urmian
I am a researcher at SLATE working at the intersection of algorithms and AI. My academic background is rooted in PhD studies in the Department of Informatics at the University of Bergen, where I am part of the Algorithms Group. My work brings together theoretical computer science, automated reasoning, and artificial intelligence. I am also involved in organizing the AI Olympiad in Norway.
I am currently working on synthetic data generation and federated learning, and I am also interested in model checking, automated theorem proving, and optimization algorithms. If you are interested in these topics, feel free to contact me.
- Reveal email
- Role
- Researcher at SLATE
- Affiliation
- Department of Informatics, Algorithms Group
- Also
- Director of NOKI
My Interests
Optimization
Exact and heuristic approaches to difficult optimization problems, especially in discrete and theoretical settings.
Automated Theorem Proving
Formal reasoning, proof search, and computational methods for automated deduction.
Machine Learning
Statistical learning, data-driven methods, and modern AI approaches including representation learning and neural models.
Combinatorial Problems
Graph problems, combinatorial optimization, and structurally rich problems that connect theory with applications.
Projects
Current
ASPIRE
On the SLATE ASPIRE project page, I am listed as a researcher on the project team working on synthetic data generation, differential privacy, and federated learning.
Open project pageEduTrust AI
EduTrust AI focuses on trust, responsibility, privacy, and the use of artificial intelligence in education through interdisciplinary collaboration.
Open project pageAI LEARN
AI LEARN is SLATE's national AI research centre initiative on hybrid intelligence and the interaction between humans and artificial intelligence.
Open project pageNOKI
NOKI, the Norwegian Olympiad in Artificial Intelligence, is another initiative I am involved in and publicly connected to as director.
Open project pagePast
AUTOPROVING
Automated Theorem Proving from the Mindset of Parameterized Complexity Theory (Project no. 288761).
Open project pageSoftware
Zygosity
Solver for computing twin-width using a heuristic approach.
Open repositoryTreeWidzard
An engine for tree-decomposition-based algorithms.
LUBP
Learning Unitary Branching Programs.
Open repository