Nathanaël Fijalkow

Nathanaël Fijalkow is a Researcher at CNRS in LaBRI (Bordeaux) and a Research Fellow at The Alan Turing Institute in London. His primary research fields include games, machine learning, automata theory, and dynamical systems, with a focus on synthesizing programs from logical specifications and probabilistic models.

Research Interests span program synthesis (programming by example), controller synthesis (temporal logic specifications), games on graphs (parity/mean payoff games), probabilistic automata (bounded ambiguity), and invariants for linear dynamical systems. He bridges formal methods with machine learning through projects like DeepSynth.

Scientific Contributions include:

• Undecidability results for probabilistic automata
• Advances in parity game algorithms (quasi-polynomial lower bounds)
• Foundations of probabilistic modal logics
• Efficient synthesis techniques using SMT solvers and distributional learning

Supervision involves guiding postdocs and PhD students such as Guillaume Lagarde, Antonio Casares, and Pierre Ohlmann. He has secured grants like the Momentum DeepSynth project (2019-2021), aiming to merge formal methods with ML for program synthesis.