Thomas Flinkow

About

Hi! I'm a fourth-year PhD student in the Principles of Programming group at Maynooth University in Ireland, working in the area of formal verification of machine learning on the Modular AI Verification and Visualisation (MAIVV) project.

Currently, I'm investigating ways to obtain correct-by-construction machine-learnt models via property-driven training.

I'm supervised by Rosemary Monahan and Barak A. Pearlmutter, and my work is supported by Taighde Éireann - Research Ireland under grant 20/FFP-P/8853 and by the ADAPT Research Ireland Centre for AI Driven Digital Content Technology at Maynooth University under grant 13/RC/2106_P2.

Preprints

Publications currently in submission or in preparation for submission.

Jun 2025
pdf code
A General Framework for Property-Driven Machine Learning

Thomas Flinkow, Marco Casadio, Colin Kessler, Rosemary Monahan, Ekaterina Komendantskaya
arXiv: arXiv:2505.00466

@misc{flinkowGeneralFrameworkPropertyDriven2025,
  title         = {A {{General Framework}} for {{Property-Driven Machine Learning}}},
  author        = {Flinkow, Thomas and Casadio, Marco and Kessler, Colin and Grīnbergs, Rosemary and Komendantskaya, Ekaterina},
  year          = {2025},
  month         = jun,
  number        = {arXiv:2505.00466},
  eprint        = {2505.00466},
  primaryclass  = {cs},
  publisher     = {arXiv},
  doi           = {10.48550/arXiv.2505.00466},
  urldate       = {2025-06-25},
  archiveprefix = {arXiv},
}

Publications

A full list is available on dblp, Google Scholar, and ORCiD.

Sep 2025
pdf code
Comparing differentiable logics for learning with logical constraints

Thomas Flinkow, Barak A. Pearlmutter, Rosemary Monahan
In Science of Computer Programming, 244, 103280.
doi: 10.1016/j.scico.2025.103280

@article{flinkowComparingDifferentiableLogics2025,
  title         = {Comparing Differentiable Logics for Learning with Logical Constraints},
  author        = {Flinkow, Thomas and Pearlmutter, Barak A. and Monahan, Rosemary},
  year          = {2025},
  month         = sep,
  journal       = {Science of Computer Programming},
  volume        = {244},
  pages         = {103280},
  issn          = {0167-6423},
  doi           = {10.1016/j.scico.2025.103280},
}
iFM'24
pdf slides poster
Towards Correct-by-Construction Machine-Learnt Models

Thomas Flinkow, Barak A. Pearlmutter, Rosemary Monahan
In PhD Symposium at 19th International Conference on Integrated Formal Methods (iFM), Manchester, UK.

@inproceedings{flinkowCorrectbyConstructionMachineLearntModels2024,
  title         = {Towards {{Correct-by-Construction Machine-Learnt Models}}},
  booktitle     = {Proceedings of the {{PhD Symposium}} at the 19th {{International Conference}} on Integrated {{Formal Methods}} in {{Manchester}} 2024 ({{iFM}} 2024), {{Manchester}}, {{United Kingdom}}, {{November}} 12, 2024},
  author        = {Flinkow, Thomas and Pearlmutter, Barak A. and Monahan, Rosemary},
  year          = {2024},
  month         = nov,
  series        = {{{CEUR Workshop Proceedings}}},
  volume        = {3860},
  pages         = {23--29},
  publisher     = {CEUR-WS.org},
}
TAP'24
pdf
Cyclone: A New Tool for Verifying/Testing Graph-Based Structures

Hao Wu, Thomas Flinkow, Dominique Méry
In Tests and Proofs: 18th International Conference (TAP), Milan, IT.
DOI: 10.1007/978-3-031-72044-4_6

@inproceedings{wuCycloneNewTool2024,
  title      = {Cyclone: {{A New Tool}} for~{{Verifying}}/{{Testing Graph-Based Structures}}: {{Tool Paper}}},
  shorttitle = {Cyclone},
  booktitle  = {Tests and {{Proofs}}: 18th {{International Conference}}, {{TAP}} 2024, {{Milan}}, {{Italy}}, {{September}} 9--10, 2024, {{Proceedings}}},
  author     = {Wu, Hao and Flinkow, Thomas and M{\'e}ry, Dominique},
  editor     = {Huisman, Marieke and Howar, Falk},
  year       = {2024},
  month      = sep,
  series     = {Lecture Notes in Computer Science},
  volume     = {15153},
  pages      = {107--124},
  publisher  = {Springer-Verlag},
  address    = {Berlin, Heidelberg},
  doi        = {10.1007/978-3-031-72044-4_6},
  isbn       = {978-3-031-72043-7},
}

Talks & Posters

Most slides and posters can be found above alongside their respective publication. Below are a few additional presentations given independently or at different venues.

SAIV'25
poster
Comparing Differentiable Logics for Learning with Logical Constraints

At 8th International Symposium on AI Verification (SAIV), Zagreb, HR.

Sep 2024
slides
Property-driven Machine Learning with Differentiable Logics

Heriot-Watt University Edinburgh, UK.

Aug 2024
slides
Differentiable Logics for Machine Learning with Logical Constraints in Practice

IT-University of Copenhagen, DK.

Academic Service

Co-Organiser:
PC Member:
Journal Reviewer: