Selected work

Representative Papers

Conjugacy languages in free inverse monoids

Aug 10, 2026

This study investigates the computational complexity and formal language-theoretic properties of the language of shortest representatives of conjugacy classes in free inverse monoids. It extends the notion of conjugacy languages to this algebraic setting by introducing a new equivalence relation, UConj, which distinguishes between trivial and nontrivial conjugacy classes. Employing techniques from formal language theory, automata theory, and combinatorial group theory, the authors prove that for rank at least two, this language is neither context-free nor co-context-free, whereas in the one-generator case it is context-free. Furthermore, they provide an explicit context-free grammar for the geodesic language of trivial conjugacy classes. These results are subsequently generalized to hyperbolic groups, right-angled Artin groups, and virtually abelian groups.

0 citationsRead paper

The Post Correspondence Problem for free groups is undecidable

Jul 15, 2026

This study addresses the decidability of the Post Correspondence Problem (PCP) over free groups when one of the two homomorphisms is injective. By reducing the halting problem for cyclic tag systems to the triviality problem for equalizers of free group homomorphisms and employing constructions based on finite partial deterministic inverse automata, the authors prove that PCP remains undecidable even under this restriction. This result resolves a long-standing open question posed by Stallings in 1984, establishing for the first time that injectivity of one homomorphism does not render the problem decidable. As a corollary, it follows that there is no general algorithm to compute the rank of such equalizers, nor to effectively construct bases or associated automata for fixed subgroups of virtual endomorphisms.

0 citationsRead paper

A language-theoretic approach to study the density of subsets in free groups

Mar 30, 2026

This study investigates the natural density of subsets in free groups, with a focus on rational subsets and finitely generated subgroups. By introducing the notion of relative density from formal language theory and integrating it with irreducible shifts of finite type and the language of freely reduced words, the authors develop a linguistic framework to analyze density behavior. Their main contributions include the first language-theoretic proof that automorphism orbits have natural density zero, a complete characterization of rational subsets possessing positive density, and the demonstration that a subgroup has positive density if and only if it has finite index. Furthermore, in non-convergent cases, they establish weak convergence properties and prove convergence of the average of upper and lower densities.

0 citationsRead paper

PartisanLens: A Multilingual Dataset of Hyperpartisan and Conspiratorial Immigration Narratives in European Media

Jan 07, 2026arXiv.org

This study addresses the paucity of systematic analyses of extreme partisanship and the “great replacement” conspiracy theory in European multilingual political discourse, as existing research has predominantly focused on English-language contexts. The authors construct the first multilingual dataset comprising 1,617 news headlines in Spanish, Italian, and Portuguese, annotated across multiple dimensions of political rhetoric. For the first time, they integrate socioeconomic and ideological profiles to guide large language models (LLMs) in simulating human annotation behavior from diverse political standpoints. By combining human-annotated benchmarks with automated classification, the project establishes a robust baseline to evaluate the capabilities and limitations of LLMs in detecting inflammatory narratives. The dataset and evaluation framework are publicly released to advance research on political discourse in European linguistic contexts.

0 citationsRead paper

Taggus: An Automated Pipeline for the Extraction of Characters' Social Networks from Portuguese Fiction Literature

Aug 05, 2025

To address the poor performance of character identification and social relation extraction in Portuguese fictional literature under low-resource conditions, this paper proposes Taggus—a fully automated, end-to-end NLP pipeline. Taggus integrates part-of-speech tagging, lightweight named entity recognition, and multi-stage heuristic rules, requiring neither large-scale annotated corpora nor large language models. Its key innovations include a Portuguese literary text–specific coreference resolution mechanism and an interaction detection strategy, both tailored to the linguistic and narrative conventions of the genre. Evaluated on a manually annotated Portuguese fiction corpus, Taggus achieves 94.1% F1 for character identification and 75.9% F1 for interaction detection—surpassing the prior state of the art by 50.7 and 22.3 percentage points, respectively. By offering a reusable, high-accuracy, and dependency-light processing paradigm, Taggus advances deep structural analysis of literary texts in under-resourced languages.

0 citationsRead paper

Recent publications

Latest Papers

Conjugacy languages in free inverse monoids

Aug 10, 2026

This study investigates the computational complexity and formal language-theoretic properties of the language of shortest representatives of conjugacy classes in free inverse monoids. It extends the notion of conjugacy languages to this algebraic setting by introducing a new equivalence relation, UConj, which distinguishes between trivial and nontrivial conjugacy classes. Employing techniques from formal language theory, automata theory, and combinatorial group theory, the authors prove that for rank at least two, this language is neither context-free nor co-context-free, whereas in the one-generator case it is context-free. Furthermore, they provide an explicit context-free grammar for the geodesic language of trivial conjugacy classes. These results are subsequently generalized to hyperbolic groups, right-angled Artin groups, and virtually abelian groups.

0 citationsRead paper

The Post Correspondence Problem for free groups is undecidable

Jul 15, 2026

This study addresses the decidability of the Post Correspondence Problem (PCP) over free groups when one of the two homomorphisms is injective. By reducing the halting problem for cyclic tag systems to the triviality problem for equalizers of free group homomorphisms and employing constructions based on finite partial deterministic inverse automata, the authors prove that PCP remains undecidable even under this restriction. This result resolves a long-standing open question posed by Stallings in 1984, establishing for the first time that injectivity of one homomorphism does not render the problem decidable. As a corollary, it follows that there is no general algorithm to compute the rank of such equalizers, nor to effectively construct bases or associated automata for fixed subgroups of virtual endomorphisms.

0 citationsRead paper

A language-theoretic approach to study the density of subsets in free groups

Mar 30, 2026

This study investigates the natural density of subsets in free groups, with a focus on rational subsets and finitely generated subgroups. By introducing the notion of relative density from formal language theory and integrating it with irreducible shifts of finite type and the language of freely reduced words, the authors develop a linguistic framework to analyze density behavior. Their main contributions include the first language-theoretic proof that automorphism orbits have natural density zero, a complete characterization of rational subsets possessing positive density, and the demonstration that a subgroup has positive density if and only if it has finite index. Furthermore, in non-convergent cases, they establish weak convergence properties and prove convergence of the average of upper and lower densities.

0 citationsRead paper

PartisanLens: A Multilingual Dataset of Hyperpartisan and Conspiratorial Immigration Narratives in European Media

Jan 07, 2026arXiv.org

This study addresses the paucity of systematic analyses of extreme partisanship and the “great replacement” conspiracy theory in European multilingual political discourse, as existing research has predominantly focused on English-language contexts. The authors construct the first multilingual dataset comprising 1,617 news headlines in Spanish, Italian, and Portuguese, annotated across multiple dimensions of political rhetoric. For the first time, they integrate socioeconomic and ideological profiles to guide large language models (LLMs) in simulating human annotation behavior from diverse political standpoints. By combining human-annotated benchmarks with automated classification, the project establishes a robust baseline to evaluate the capabilities and limitations of LLMs in detecting inflammatory narratives. The dataset and evaluation framework are publicly released to advance research on political discourse in European linguistic contexts.

0 citationsRead paper

Taggus: An Automated Pipeline for the Extraction of Characters' Social Networks from Portuguese Fiction Literature

Aug 05, 2025

To address the poor performance of character identification and social relation extraction in Portuguese fictional literature under low-resource conditions, this paper proposes Taggus—a fully automated, end-to-end NLP pipeline. Taggus integrates part-of-speech tagging, lightweight named entity recognition, and multi-stage heuristic rules, requiring neither large-scale annotated corpora nor large language models. Its key innovations include a Portuguese literary text–specific coreference resolution mechanism and an interaction detection strategy, both tailored to the linguistic and narrative conventions of the genre. Evaluated on a manually annotated Portuguese fiction corpus, Taggus achieves 94.1% F1 for character identification and 75.9% F1 for interaction detection—surpassing the prior state of the art by 50.7 and 22.3 percentage points, respectively. By offering a reusable, high-accuracy, and dependency-light processing paradigm, Taggus advances deep structural analysis of literary texts in under-resourced languages.

0 citationsRead paper