Tilman Hinnerichs

Hello, I'm Tilman!

PhD Student | Computer Scientist | 2-Meter-Peter

I am a PhD student for Computer Science at Algorithmics Group and the STAR AI Lab at TU Delft with Sebastijan Dumancic and Neil Yorke-Smith, researching in the field of program synthesis, neuro-symbolic proving and reasoning, ontologies and fuzzy logics, and their application to bioinformatics.

Specifically, I try to combine these fields learning neuro-symbolic representations, thus making logical reasoning accessible to neural networks and ‘intuition’ accessible to reasoners.

One of my main projects is Herb.jl a program synthesis library written in Julia.

Previously, I was a Guest Researcher in the Bio-Ontology Research Group (BORG) at KAUST with Robert Hoehndorf and the Neural Data Science and Statistical Computing lab (NeurDS) with Nico Scherf at the MPI for CBS.

news

May 27, 2026	I gave an invited talk on neuro-symbolic program synthesis at the USCKI symposium “Neuro ∧ Symbolic ⇒ AI” in Utrecht, alongside invited talks by Luc de Raedt and Emile van Krieken.
May 13, 2026	My talk and poster on `Herb.jl` got accepted at JuliaCon Global 2026 in Mainz, Germany (Aug 10-15). Happy to meet you there!
Apr 2, 2026	I gave an invited talk on “Neural Program Synthesis: Learning to Program from Trees” as part of the “Graphs and Data” seminar, where Neil Yorke-Smith also gave a talk.
Feb 1, 2026	We published our extended paper “Herb.jl: A Unifying Program Synthesis Library” on ArXiv, showing how to use and customize Herb for your own program synthesis applications.
Nov 28, 2025	We presented `Herb.jl` at PLNL 2025. The Pluto notebook from the talk is available online.
Jul 1, 2025	We published our first Herb-related paper “Modelling Program Spaces in Program Synthesis with Constraints” on ArXiv, introducing our solver BART and a new notion of syntactic constraints.
Jan 3, 2025	Our project from the first BioHackathon MENA led to the paper “Generating unseen diseases patient data using ontology enhanced generative adversarial networks” by Chang Sun and Michel Dumontier, published in npj Digital Medicine.
Jul 10, 2024	We presented `Herb.jl` at JuliaCon 2024 with a poster session and a talk by Tilman Hinnerichs (video).
Mar 1, 2024	My talk on `Herb.jl` got accepted for JuliaCon 2024
Nov 16, 2023	I was awarded a scholarship by the German National Scholarship Foundation (Studienstiftung des Deutschen Volkes) for my PhD.
Jul 24, 2023	I attended the first European Summer School of AI (ESSAI) in Ljubljana, Slovenia
Apr 14, 2023	I gave a talk at the “Neural Causal and Symbolic AI” workshop organized by the TAILOR network
Feb 7, 2023	I participated in the first BioHackathon MENA at CBRC, KAUST

selected publications

2026

Herb.jl: A Unifying Program Synthesis Library

Tilman Hinnerichs, Reuben Gardos Reid, Jaap Jong, and 6 more authors

arXiv preprint arXiv:2510.09726, 2026

Bib

@article{hinnerichs2026herb,
  title = {Herb.jl: A Unifying Program Synthesis Library},
  author = {Hinnerichs, Tilman and Reid, Reuben Gardos and de Jong, Jaap and Swinkels, Bart and Wochner, Pamela and Filat, Nicolae and Magurescu, Tudor and Hanou, Issa and Dumancic, Sebastijan},
  journal = {arXiv preprint arXiv:2510.09726},
  year = {2026}
}

2025

NeSy

Declarative Design of Neural Predicates in Neuro-Symbolic Systems

Tilman Hinnerichs, Robin Manhaeve, Giuseppe Marra, and 1 more author

2025

PS/CP

Modelling Program Spaces in Program Synthesis with Constraints

Tilman Hinnerichs, Bart Swinkels, Jaap Jong, and 4 more authors

arXiv preprint arXiv:2508.00005, 2025

Bib

@article{hinnerichs2025modelling,
  title = {Modelling Program Spaces in Program Synthesis with Constraints},
  author = {Hinnerichs, Tilman and Swinkels, Bart and de Jong, Jaap and Reid, Reuben Gardos and Magirescu, Tudor and Yorke-Smith, Neil and Dumancic, Sebastijan},
  journal = {arXiv preprint arXiv:2508.00005},
  year = {2025}
}

2021

BioInf

DTI-Voodoo: machine learning over interaction networks and ontology-based background knowledge predicts drug–target interactions

Tilman Hinnerichs, and Robert Hoehndorf

Bioinformatics, 2021

Bib

@article{hinnerichs2021dti,
  title = {DTI-Voodoo: machine learning over interaction networks and ontology-based background knowledge predicts drug--target interactions},
  author = {Hinnerichs, Tilman and Hoehndorf, Robert},
  journal = {Bioinformatics},
  year = {2021},
  publisher = {Oxford University Press}
}