I am a mathematician (math/CS PhD, physics MSc) working in information theory and discrete stochastic processes. I am also interested in theoretical properties of large language models and quantitative linguistics.
I work at the IPI PAN, with the Statistical Analysis and Modeling Group, but occasionally I collaborate also with the Linguistic Engineering Group.

Here is a sample of my interests: (in English) (po polsku)
If you are interested in a PhD under my supervision, please read this.

My email address is ldebowsk@ipipan.waw.pl and my phone number is (+48) 22 3800 553.

Selected publications:

• Book "Information Theory Meets Power Laws: Stochastic Processes and Language Models" at Wiley (2021); it received the prize of the Committee on Informatics of the Polish Academy of Sciences.
• Article "A Refutation of Finite-State Language Models through Zipf’s Law for Factual Knowledge" in Entropy (2021).
• Article "Can Menzerath’s law be a criterion of complexity in communication?" in PLOS ONE (2021) with Iván G. Torre and Antoni Hernández-Fernández.
• Article "Universal Coding and Prediction on Ergodic Random Points" in the Bulletin of Symbolic Logic (2022) with Tomasz Steifer.
• Article "Universal Densities Exist for Every Finite Reference Measure" in the IEEE Transactions on Information Theory (2023). Erratum.

Work-related items:

• profiles: ORCID, arXiv, GitHub, DBLP, Google Scholar, ResearchGate, LinkedIn, Twitter
• slides, teaching, CV (contains the complete publication list)
• PhD students: Tomasz Steifer (defended 29.09.2020, co-supervised by Dariusz Kalociński)

Nerd's interest:

• Jak się wzbogacić prawie na pewno? (How to get rich almost surely?)
• Charty zostały... czyli o generowaniu wierszy sylabicznych. (On automatic generation of rhymed poems.)
• Phonetic transcription and computational poetry.
• The Chaos by Gerard Nolst Trenité, transcribed into IPA symbols by me.
• The second version of my Perl script for extracting genealogical trees from GEDCOM databases (Ancestris, Gramps, Family Tree Maker) and typesetting them in Latex using the genealogytree.sty package: ZIP archive, which contains the gedcom2latex.pl script and an example of usage.