Automatic Construction of a Legal Citation Graph from 100 Million Ukrainian Court Decisions: Large-Scale Extraction, Topological Analysis, and Ontology-Driven Clustering

arxiv: 2605.15362 · v1 · pith:2ZJ4HTA7new · submitted 2026-05-14 · 💻 cs.CL · cs.DL· cs.IR

Automatic Construction of a Legal Citation Graph from 100 Million Ukrainian Court Decisions: Large-Scale Extraction, Topological Analysis, and Ontology-Driven Clustering

Volodymyr Ovcharov This is my paper

Pith reviewed 2026-05-19 15:34 UTC · model grok-4.3

classification 💻 cs.CL cs.DLcs.IR

keywords legal citation graphUkrainian court decisionscommunity detectionlegal ontologycitation network analysislegislative predictionpower-law networks

0 comments p. Extension

pith:2ZJ4HTA7 Add to your LaTeX paper

What is a Pith Number?

\usepackage{pith}
\pithnumber{2ZJ4HTA7}

Prints a linked pith:2ZJ4HTA7 badge after your title and writes the identifier into PDF metadata. Compiles on arXiv with no extra files. Learn more

The pith

A citation graph from 100 million Ukrainian court decisions encodes legal domain boundaries without supervision and predicts future legislative importance with near-perfect accuracy.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper builds the first large-scale citation graph by scanning 99.5 million full-text Ukrainian court decisions with regular expressions to extract 502 million links of six types. The resulting network exhibits a power-law degree distribution and, through community detection on its co-citation projection, partitions itself into clusters that match the main legal domains such as civil, criminal, administrative, and commercial law. Citation statistics drawn from the same graph then forecast which articles will rank among the most cited in the future with an AUC of 0.9984, while temporal shifts in the network register events such as the 2022 invasion as measurable changes in citation entropy.

Core claim

By constructing a citation graph of 502 million edges from the complete registry of 99.5 million Ukrainian court decisions, the authors show that Louvain community detection applied to the co-citation projection recovers legal domain boundaries with modularity 0.44-0.55 and high temporal stability, while citation-derived features predict the top-1000 most important articles with AUC 0.9984 and detect legislative regime changes as phase transitions in network entropy.

What carries the argument

The co-citation projection of the citation graph together with Louvain community detection, which partitions the network into clusters that align with established legal domains without using any labeled training data.

If this is right

The citation network follows a power-law degree distribution with exponent 1.57, placing it near the EU Court of Justice in scale-free character.
Louvain communities on co-citations recover the four principal legal domains with modularity Q between 0.44 and 0.55 and normalized mutual information of 0.83-0.86 across time periods.
Citation features alone predict the top-1000 future high-impact articles with AUC 0.9984, outperforming a naive frequency baseline.
Temporal dynamics in the network identify legislative regime changes as phase transitions and register the 2022 invasion as a rise in citation entropy from 11.02 to 13.49.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The automatically derived legal ontology could be used as a dynamic layer in AI systems that retrieve and reason over case law without hand-crafted taxonomies.
The same extraction approach might be repeated on court archives from other countries to produce comparable jurisdiction-specific legal maps.
Entropy spikes or community reorganizations could serve as quantitative indicators for monitoring the emergence of new legal subfields after major societal events.

Load-bearing premise

Regular-expression patterns applied to full-text decisions correctly locate and classify all six types of citation links across the entire collection, and the 200-decision validation sample is representative of the remaining 99.5 million documents.

What would settle it

A larger manual audit of randomly sampled decisions that finds substantially lower precision than the reported 1.00 or that shows the detected communities fail to correspond to recognized legal fields when reviewed by domain experts would falsify the extraction and ontology claims.

Figures

Figures reproduced from arXiv: 2605.15362 by Volodymyr Ovcharov.

**Figure 2.** Figure 2: Citations extracted vs. decisions processed per year. The near-linear relationship con [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

**Figure 3.** Figure 3: Log-log degree distribution of legislation articles by citation count. The dashed line shows [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

**Figure 4.** Figure 4: Annual citation volume (2007–2025). Vertical lines mark major legislative events: 2010 [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗

**Figure 5.** Figure 5: Stacked area chart of citation volume by type (top 3 types). Case references (red) [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗

**Figure 6.** Figure 6: Top-10 most-cited legislation articles. Criminal Code art. 185 (theft) leads with 3.3M [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗

**Figure 7.** Figure 7: Mean vs. median citation degree by type (log scale, excluding Supreme Court singleton). [PITH_FULL_IMAGE:figures/full_fig_p012_7.png] view at source ↗

read the original abstract

Half a billion citation edges extracted from 100.7 million Ukrainian court decisions reveal that judicial citation structure encodes legal domain boundaries without supervision and predicts future legislative importance with near-perfect accuracy. We construct the first large-scale citation graph from the complete EDRSR registry (99.5 million full texts, 1.1 TB), extracting 502 million citation links across six types via regex on commodity hardware in approximately 5 hours, with precision of 1.00 on a 200-decision validation sample (95% Wilson CI: [0.982, 1.000]). Three principal findings emerge. (1) The degree distribution follows a power law (alpha = 1.57 +/- 0.008), placing the Ukrainian court network near the EU Court of Justice and below the US Supreme Court, with hub articles cited by millions of decisions. (2) Louvain community detection on the co-citation projection recovers legal domain boundaries (civil, criminal, administrative, commercial) with modularity Q = 0.44-0.55 and temporal stability (NMI = 0.83-0.86 across periods), constituting an automatically constructed legal ontology grounded in judicial practice. (3) Citation features predict top-1000 articles with AUC = 0.9984, substantially outperforming a naive frequency baseline (P@1000 = 0.655); temporal dynamics detect legislative regime changes as phase transitions and the 2022 invasion as a citation entropy spike (H: 11.02 -> 13.49) with emergent wartime legislation nodes. The citation-derived ontology is operationalized as the domain layer of a workflow memory system for LLM-assisted legal analysis, connecting to the ontology-controlled paradigm. The extraction pipeline, analysis code, and aggregated statistics are released as open data.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper builds a genuinely large citation graph from a complete national court registry and shows it can surface legal domains and predict legislative importance, but the regex extraction rests on thin validation.

read the letter

The main thing to know is that this extracts 502 million citation edges from 99.5 million Ukrainian court decisions and then uses the graph to recover domain boundaries via clustering and to forecast which articles will rank high in importance. The scale and the open release of the pipeline plus aggregated stats are the clearest strengths here. Standard network tools—power-law fitting with alpha around 1.57, Louvain on the co-citation projection giving Q of 0.44-0.55 and stable NMI, plus AUC 0.9984 on top-1000 prediction—produce results that line up with the claims and beat a simple frequency baseline. The temporal entropy spike tied to the 2022 invasion is a concrete observation that adds some external grounding. Releasing code and data on commodity hardware in five hours is also practical and helpful for follow-up work. The soft spot is the extraction itself. Precision reaches 1.00 on a 200-decision sample, yet there is no recall number, no description of how the sample was drawn, and no error analysis for the formatting and abbreviation variations common in Ukrainian decisions. If the regex systematically misses or misclassifies links, the power-law, communities, and prediction numbers all sit on the same uncertain base. The prediction task would also benefit from explicit temporal splits to rule out leakage. This is for people in legal informatics or citation-network studies who need large real-world graphs rather than for general machine-learning audiences. A reader who wants fresh data or an example of unsupervised ontology recovery from practice will get usable material. It deserves a serious referee because the dataset is new at this scale and the results are checkable once the validation details are filled in. Send it out, but ask the authors to strengthen the extraction evaluation and confirm the temporal handling before acceptance.

Referee Report

2 major / 2 minor

Summary. The paper constructs the first large-scale citation graph from the complete EDRSR registry of 99.5 million Ukrainian court decisions (1.1 TB), extracting 502 million citation links across six types via regex patterns on commodity hardware. It reports a power-law degree distribution (alpha = 1.57 +/- 0.008), uses Louvain community detection on the co-citation projection to recover legal domains (civil, criminal, administrative, commercial) with modularity Q = 0.44-0.55 and temporal stability (NMI = 0.83-0.86), and shows citation features predict the top-1000 most important articles with AUC = 0.9984 (outperforming a frequency baseline). The work releases the extraction pipeline, analysis code, and aggregated statistics as open data and operationalizes the resulting ontology as a domain layer for LLM-assisted legal analysis.

Significance. If the graph extraction holds at scale, the manuscript delivers a substantial contribution to computational legal studies and network science by providing the largest national judicial citation network to date. The unsupervised recovery of domain boundaries from citation structure and the near-perfect predictive performance for legislative importance demonstrate that citation topology encodes meaningful legal information. The explicit release of code, pipeline, and data is a clear strength that enables reproducibility and downstream use in ontology-controlled LLM workflows.

major comments (2)

[Validation subsection (likely §4)] Validation subsection (likely §4): The reported precision of 1.00 (Wilson CI [0.982, 1.000]) on a 200-decision sample supports extraction quality for the six citation types, but the manuscript provides neither recall nor any description of how the sample was drawn (random, stratified by court type/year/region, or otherwise). Ukrainian decisions exhibit substantial variation in formatting and phrasing; without recall measurement or error analysis on edge cases, the 502 million extracted edges may systematically under- or over-count links, directly affecting the power-law fit in §5, the Louvain communities (Q and NMI) in §6, and the AUC=0.9984 prediction results in §7.
[Prediction experiment (likely §7)] Prediction experiment (likely §7): The claim that citation features predict future legislative importance with AUC = 0.9984 requires explicit temporal train-test splits (e.g., training on decisions up to year T and testing on later periods). The current description does not specify such splits or how the top-1000 target articles are defined temporally, which is load-bearing for the interpretation of phase transitions and the 2022 invasion entropy spike as predictive signals rather than contemporaneous correlations.

minor comments (2)

[Abstract] The abstract states extraction took 'approximately 5 hours' but should specify exact hardware configuration and parallelization details to allow independent reproduction.
[Figures] Figure captions and legends for the degree-distribution plot and community visualizations should explicitly label axes, color mappings, and any resolution parameters used in Louvain.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their thorough review and valuable feedback on our manuscript. The comments highlight important aspects of validation and experimental design that we will address to strengthen the paper. Below we provide point-by-point responses to the major comments.

read point-by-point responses

Referee: Validation subsection (likely §4): The reported precision of 1.00 (Wilson CI [0.982, 1.000]) on a 200-decision sample supports extraction quality for the six citation types, but the manuscript provides neither recall nor any description of how the sample was drawn (random, stratified by court type/year/region, or otherwise). Ukrainian decisions exhibit substantial variation in formatting and phrasing; without recall measurement or error analysis on edge cases, the 502 million extracted edges may systematically under- or over-count links, directly affecting the power-law fit in §5, the Louvain communities (Q and NMI) in §6, and the AUC=0.9984 prediction results in §7.

Authors: We agree that a more complete validation including recall and sampling details would be beneficial. We will revise the validation subsection to describe the sample as having been drawn via stratified random sampling according to court type and decision year. We will also include a discussion of potential missed citations through manual inspection of edge cases and analyze how any under-counting could influence the power-law fit, community metrics, and prediction performance. This will provide a more balanced view of the extraction quality. revision: yes
Referee: Prediction experiment (likely §7): The claim that citation features predict future legislative importance with AUC = 0.9984 requires explicit temporal train-test splits (e.g., training on decisions up to year T and testing on later periods). The current description does not specify such splits or how the top-1000 target articles are defined temporally, which is load-bearing for the interpretation of phase transitions and the 2022 invasion entropy spike as predictive signals rather than contemporaneous correlations.

Authors: We thank the referee for this observation on the temporal aspects of the prediction task. We will update the description in the prediction experiment section to explicitly state the temporal train-test splits employed, with training on data up to year T and testing on subsequent periods. We will also clarify how the top-1000 articles are identified based on their importance in the post-training time window. These details will support the interpretation of the results as predictive, including the identification of phase transitions and the entropy changes associated with the 2022 invasion. revision: yes

Circularity Check

0 steps flagged

No circularity: results follow directly from regex extraction and standard graph algorithms

full rationale

The paper extracts citation edges via regex patterns applied to the full corpus, validates precision on a 200-decision sample, then applies power-law fitting, Louvain community detection, and citation-feature prediction using established methods. No step defines a quantity in terms of itself, renames a fitted parameter as a prediction, or relies on a self-citation chain for a uniqueness claim. All reported quantities (degree exponent, modularity Q, AUC) are computed outputs from the observed graph rather than tautological re-expressions of inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The work rests primarily on the assumption that regex can reliably extract citations and on standard network-science methods; no new physical entities or heavily fitted parameters are introduced.

axioms (1)

domain assumption Regular expressions applied to full-text decisions can extract citation links of six types with near-perfect precision
The entire pipeline begins with regex extraction on the 1.1 TB corpus.

pith-pipeline@v0.9.0 · 5879 in / 1256 out tokens · 59302 ms · 2026-05-19T15:34:09.322183+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Louvain community detection on the co-citation projection recovers legal domain boundaries (civil, criminal, administrative, commercial) with modularity Q = 0.44-0.55
IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

The degree distribution follows a power law (alpha = 1.57 +/- 0.008)

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

22 extracted references · 22 canonical work pages · 1 internal anchor

[1]

Blondel, Jean-Loup Guillaume, Renaud Lambiotte, and Etienne Lefebvre

Vincent D. Blondel, Jean-Loup Guillaume, Renaud Lambiotte, and Etienne Lefebvre. Fast unfolding of communities in large networks.Journal of Statistical Mechanics: Theory and Experiment, 2008(10):P10008, 2008. doi: 10.1088/1742-5468/2008/10/P10008

work page doi:10.1088/1742-5468/2008/10/p10008 2008
[2]

Bommarito, Daniel Martin Katz, and Eric M

Michael J. Bommarito, Daniel Martin Katz, and Eric M. Detterman. Lexnlp: Natural language processing and information extraction for legal and regulatory texts.Research Handbook on Big Data Law, 2018

work page 2018
[3]

LEGAL - BERT : The Muppets straight out of Law School

Ilias Chalkidis, Marios Fergadiotis, Prodromos Malakasiotis, Nikolaos Aletras, and Ion An- droutsopoulos. LEGAL-BERT: The muppets straight out of law school. InFindings of EMNLP, pages 2898–2904, 2020. doi: 10.18653/v1/2020.findings-emnlp.261. URLhttps: //aclanthology.org/2020.findings-emnlp.261/

work page doi:10.18653/v1/2020.findings-emnlp.261 2020
[4]

Aaron Clauset, Cosma Rohilla Shalizi, and Mark E. J. Newman. Power-law distributions in empirical data.SIAM Review, 51(4):661–703, 2009. doi: 10.1137/070710111. URLhttps: //arxiv.org/abs/0706.1062

work page internal anchor Pith review Pith/arXiv arXiv doi:10.1137/070710111 2009
[5]

Measuring law over time: A network analytical framework with an application to statutes and regulations in the united states and germany.Frontiers in Physics, 9:658463,

Corinna Coupette, Janis Beckedorf, Dirk Hartung, Michael Bommarito, and Daniel Martin Katz. Measuring law over time: A network analytical framework with an application to statutes and regulations in the united states and germany.Frontiers in Physics, 9:658463,

work page
[6]

URLhttps://doi.org/10.3389/fphy.2021.658463

doi: 10.3389/fphy.2021.658463. URLhttps://doi.org/10.3389/fphy.2021.658463

work page doi:10.3389/fphy.2021.658463 2021
[7]

Fowler, Timothy R

James H. Fowler, Timothy R. Johnson, James F. Spriggs, Sangick Jeon, and Paul J. Wahlbeck. Network analysis and the law: Measuring the legal importance of precedents at the U.S. Supreme Court.Political Analysis, 15(3):324–346, 2007. doi: 10.1093/pan/mpm011

work page doi:10.1093/pan/mpm011 2007
[8]

Using citation analysis techniques for computer-assisted legal research in conti- nental jurisdictions.Graduate thesis, University of Edinburgh, 2009

Anton Geist. Using citation analysis techniques for computer-assisted legal research in conti- nental jurisdictions.Graduate thesis, University of Edinburgh, 2009

work page 2009
[9]

Thomas R. Gruber. A translation approach to portable ontology specifications.Knowledge Acquisition, 5(2):199–220, 1993. doi: 10.1006/knac.1993.1008. 1Dataset:https://huggingface.co/datasets/overthelex/ukrainian-legal-citation-graph; source code: https://github.com/overthelex/SecondLayer. 14

work page doi:10.1006/knac.1993.1008 1993
[10]

The network of french legal codes

Pierre Mazzega, Danièle Bourcier, and Romain Boulet. The network of french legal codes. Proceedings of the 12th International Conference on Artificial Intelligence and Law, pages 236–237, 2009. doi: 10.1145/1568234.1568271

work page doi:10.1145/1568234.1568271 2009
[11]

Esquivel, Ludvig Lizana, and Martin Rosvall

Atieh Mirshahvalad, Argimiro V. Esquivel, Ludvig Lizana, and Martin Rosvall. Dynamics of interacting information waves in networks.Physical Review E, 89:012809, 2014. doi: 10.1103/ PhysRevE.89.012809

work page 2014
[12]

Emergence of network effects and predictability in the judicial system.Scientific Reports, 11: 2740, 2021

Enys Monés, Piotr Sapiezynski, Simon Thordal, Henrik Palmer Olsen, and Sune Lehmann. Emergence of network effects and predictability in the judicial system.Scientific Reports, 11: 2740, 2021. doi: 10.1038/s41598-021-82430-x

work page doi:10.1038/s41598-021-82430-x 2021
[13]

Mark E. J. Newman and Michelle Girvan. Finding and evaluating community structure in networks.Physical Review E, 69(2):026113, 2004. doi: 10.1103/PhysRevE.69.026113

work page doi:10.1103/physreve.69.026113 2004
[14]

From ontology-controlled systems to oversight-controlled systems: A domain constitution for edit-trace rlhf.Cybernetics and Systems Analysis, 2026

Vladimir Ovcharov. From ontology-controlled systems to oversight-controlled systems: A domain constitution for edit-trace rlhf.Cybernetics and Systems Analysis, 2026. Submitted

work page 2026
[15]

Workflow memory for long-horizon agentic composition: Architecture, dual-mode retrieval, and retrieval-correction signal.arXiv preprint, 2026

Vladimir Ovcharov. Workflow memory for long-horizon agentic composition: Architecture, dual-mode retrieval, and retrieval-correction signal.arXiv preprint, 2026

work page 2026
[16]

Alexander V. Palagin. Architecture of ontology-controlled computer systems.Cybernetics and Systems Analysis, 42(2):254–264, 2006. doi: 10.1007/s10559-006-0061-z

work page doi:10.1007/s10559-006-0061-z 2006
[17]

Palagin, Serhiy L

Alexander V. Palagin, Serhiy L. Kryvyi, and Mykola G. Petrenko. On the automation of the process of extracting knowledge from natural language texts. InNatural and Artificial Intelligence, International Book Series, Sofia, 2012. ITHEA

work page 2012
[18]

Distribu- tional semantic modeling: A revised technique to train term/word vector space models ap- plying the ontology-related approach.Problems in Programming, (2–3):341–351, 2020

Oleksandr Palagin, Vitalii Velychko, Kyrylo Malakhov, and Oleksandr Shchurov. Distribu- tional semantic modeling: A revised technique to train term/word vector space models ap- plying the ontology-related approach.Problems in Programming, (2–3):341–351, 2020. doi: 10.15407/pp2020.02-03.341. URLhttps://arxiv.org/abs/2003.03350

work page doi:10.15407/pp2020.02-03.341 2020
[19]

OntoChat- GPT information system: Ontology-driven structured prompts for ChatGPT meta-learning

Oleksandr Palagin, Vladislav Kaverinskiy, Anna Litvin, and Kyrylo Malakhov. OntoChat- GPT information system: Ontology-driven structured prompts for ChatGPT meta-learning. International Journal of Computing, 22(2):170–183, 2023. doi: 10.47839/ijc.22.2.3086. URL https://arxiv.org/abs/2307.05082

work page doi:10.47839/ijc.22.2.3086 2023
[20]

EDRSR: Unified State Register of Court Decisions of Ukraine.https://reyestr.court.gov.ua/, 2024

State Judicial Administration of Ukraine. EDRSR: Unified State Register of Court Decisions of Ukraine.https://reyestr.court.gov.ua/, 2024. Accessed: 2026-05-13

work page 2024
[21]

Legislation of Ukraine — Verkhovna Rada of Ukraine.https: //zakon.rada.gov.ua/, 2024

Verkhovna Rada of Ukraine. Legislation of Ukraine — Verkhovna Rada of Ukraine.https: //zakon.rada.gov.ua/, 2024. Accessed: 2026-05-13

work page 2024
[22]

Determining authority of dutch case law

Radboud Winkels, Jelle de Ruyter, and Henryk Kroese. Determining authority of dutch case law. InLegal Knowledge and Information Systems (JURIX 2011), pages 103–112, 2011. doi: 10.3233/978-1-60750-981-3-103. 15

work page doi:10.3233/978-1-60750-981-3-103 2011

[1] [1]

Blondel, Jean-Loup Guillaume, Renaud Lambiotte, and Etienne Lefebvre

Vincent D. Blondel, Jean-Loup Guillaume, Renaud Lambiotte, and Etienne Lefebvre. Fast unfolding of communities in large networks.Journal of Statistical Mechanics: Theory and Experiment, 2008(10):P10008, 2008. doi: 10.1088/1742-5468/2008/10/P10008

work page doi:10.1088/1742-5468/2008/10/p10008 2008

[2] [2]

Bommarito, Daniel Martin Katz, and Eric M

Michael J. Bommarito, Daniel Martin Katz, and Eric M. Detterman. Lexnlp: Natural language processing and information extraction for legal and regulatory texts.Research Handbook on Big Data Law, 2018

work page 2018

[3] [3]

LEGAL - BERT : The Muppets straight out of Law School

Ilias Chalkidis, Marios Fergadiotis, Prodromos Malakasiotis, Nikolaos Aletras, and Ion An- droutsopoulos. LEGAL-BERT: The muppets straight out of law school. InFindings of EMNLP, pages 2898–2904, 2020. doi: 10.18653/v1/2020.findings-emnlp.261. URLhttps: //aclanthology.org/2020.findings-emnlp.261/

work page doi:10.18653/v1/2020.findings-emnlp.261 2020

[4] [4]

Aaron Clauset, Cosma Rohilla Shalizi, and Mark E. J. Newman. Power-law distributions in empirical data.SIAM Review, 51(4):661–703, 2009. doi: 10.1137/070710111. URLhttps: //arxiv.org/abs/0706.1062

work page internal anchor Pith review Pith/arXiv arXiv doi:10.1137/070710111 2009

[5] [5]

Measuring law over time: A network analytical framework with an application to statutes and regulations in the united states and germany.Frontiers in Physics, 9:658463,

Corinna Coupette, Janis Beckedorf, Dirk Hartung, Michael Bommarito, and Daniel Martin Katz. Measuring law over time: A network analytical framework with an application to statutes and regulations in the united states and germany.Frontiers in Physics, 9:658463,

work page

[6] [6]

URLhttps://doi.org/10.3389/fphy.2021.658463

doi: 10.3389/fphy.2021.658463. URLhttps://doi.org/10.3389/fphy.2021.658463

work page doi:10.3389/fphy.2021.658463 2021

[7] [7]

Fowler, Timothy R

James H. Fowler, Timothy R. Johnson, James F. Spriggs, Sangick Jeon, and Paul J. Wahlbeck. Network analysis and the law: Measuring the legal importance of precedents at the U.S. Supreme Court.Political Analysis, 15(3):324–346, 2007. doi: 10.1093/pan/mpm011

work page doi:10.1093/pan/mpm011 2007

[8] [8]

Using citation analysis techniques for computer-assisted legal research in conti- nental jurisdictions.Graduate thesis, University of Edinburgh, 2009

Anton Geist. Using citation analysis techniques for computer-assisted legal research in conti- nental jurisdictions.Graduate thesis, University of Edinburgh, 2009

work page 2009

[9] [9]

Thomas R. Gruber. A translation approach to portable ontology specifications.Knowledge Acquisition, 5(2):199–220, 1993. doi: 10.1006/knac.1993.1008. 1Dataset:https://huggingface.co/datasets/overthelex/ukrainian-legal-citation-graph; source code: https://github.com/overthelex/SecondLayer. 14

work page doi:10.1006/knac.1993.1008 1993

[10] [10]

The network of french legal codes

Pierre Mazzega, Danièle Bourcier, and Romain Boulet. The network of french legal codes. Proceedings of the 12th International Conference on Artificial Intelligence and Law, pages 236–237, 2009. doi: 10.1145/1568234.1568271

work page doi:10.1145/1568234.1568271 2009

[11] [11]

Esquivel, Ludvig Lizana, and Martin Rosvall

Atieh Mirshahvalad, Argimiro V. Esquivel, Ludvig Lizana, and Martin Rosvall. Dynamics of interacting information waves in networks.Physical Review E, 89:012809, 2014. doi: 10.1103/ PhysRevE.89.012809

work page 2014

[12] [12]

Emergence of network effects and predictability in the judicial system.Scientific Reports, 11: 2740, 2021

Enys Monés, Piotr Sapiezynski, Simon Thordal, Henrik Palmer Olsen, and Sune Lehmann. Emergence of network effects and predictability in the judicial system.Scientific Reports, 11: 2740, 2021. doi: 10.1038/s41598-021-82430-x

work page doi:10.1038/s41598-021-82430-x 2021

[13] [13]

Mark E. J. Newman and Michelle Girvan. Finding and evaluating community structure in networks.Physical Review E, 69(2):026113, 2004. doi: 10.1103/PhysRevE.69.026113

work page doi:10.1103/physreve.69.026113 2004

[14] [14]

From ontology-controlled systems to oversight-controlled systems: A domain constitution for edit-trace rlhf.Cybernetics and Systems Analysis, 2026

Vladimir Ovcharov. From ontology-controlled systems to oversight-controlled systems: A domain constitution for edit-trace rlhf.Cybernetics and Systems Analysis, 2026. Submitted

work page 2026

[15] [15]

Workflow memory for long-horizon agentic composition: Architecture, dual-mode retrieval, and retrieval-correction signal.arXiv preprint, 2026

Vladimir Ovcharov. Workflow memory for long-horizon agentic composition: Architecture, dual-mode retrieval, and retrieval-correction signal.arXiv preprint, 2026

work page 2026

[16] [16]

Alexander V. Palagin. Architecture of ontology-controlled computer systems.Cybernetics and Systems Analysis, 42(2):254–264, 2006. doi: 10.1007/s10559-006-0061-z

work page doi:10.1007/s10559-006-0061-z 2006

[17] [17]

Palagin, Serhiy L

Alexander V. Palagin, Serhiy L. Kryvyi, and Mykola G. Petrenko. On the automation of the process of extracting knowledge from natural language texts. InNatural and Artificial Intelligence, International Book Series, Sofia, 2012. ITHEA

work page 2012

[18] [18]

Distribu- tional semantic modeling: A revised technique to train term/word vector space models ap- plying the ontology-related approach.Problems in Programming, (2–3):341–351, 2020

Oleksandr Palagin, Vitalii Velychko, Kyrylo Malakhov, and Oleksandr Shchurov. Distribu- tional semantic modeling: A revised technique to train term/word vector space models ap- plying the ontology-related approach.Problems in Programming, (2–3):341–351, 2020. doi: 10.15407/pp2020.02-03.341. URLhttps://arxiv.org/abs/2003.03350

work page doi:10.15407/pp2020.02-03.341 2020

[19] [19]

OntoChat- GPT information system: Ontology-driven structured prompts for ChatGPT meta-learning

Oleksandr Palagin, Vladislav Kaverinskiy, Anna Litvin, and Kyrylo Malakhov. OntoChat- GPT information system: Ontology-driven structured prompts for ChatGPT meta-learning. International Journal of Computing, 22(2):170–183, 2023. doi: 10.47839/ijc.22.2.3086. URL https://arxiv.org/abs/2307.05082

work page doi:10.47839/ijc.22.2.3086 2023

[20] [20]

EDRSR: Unified State Register of Court Decisions of Ukraine.https://reyestr.court.gov.ua/, 2024

State Judicial Administration of Ukraine. EDRSR: Unified State Register of Court Decisions of Ukraine.https://reyestr.court.gov.ua/, 2024. Accessed: 2026-05-13

work page 2024

[21] [21]

Legislation of Ukraine — Verkhovna Rada of Ukraine.https: //zakon.rada.gov.ua/, 2024

Verkhovna Rada of Ukraine. Legislation of Ukraine — Verkhovna Rada of Ukraine.https: //zakon.rada.gov.ua/, 2024. Accessed: 2026-05-13

work page 2024

[22] [22]

Determining authority of dutch case law

Radboud Winkels, Jelle de Ruyter, and Henryk Kroese. Determining authority of dutch case law. InLegal Knowledge and Information Systems (JURIX 2011), pages 103–112, 2011. doi: 10.3233/978-1-60750-981-3-103. 15

work page doi:10.3233/978-1-60750-981-3-103 2011