Pith. sign in

REVIEW 3 major objections 6 minor 39 references

An Ontology for Representing Curriculum and Learning Material

T0 review · 3 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The Curriculum KG Ontology is a formal OWL 2 DL schema for densely interlinking educational, learning, and training materials, materialized into a queryable knowledge graph and validated against educator-posed competency questions.

desk verdict A solid modular ontology and materialized KG for educational materials, undercut by validation evidence that does not actually validate. read the letter →

arxiv 2506.05751 v1 pith:PZDYM4GE submitted 2025-06-06 cs.CY cs.AI

Antrea Christou , Chris Davis Jaldi , Joseph Zalewski , Hande Küçük McGinty , Pascal Hitzler , Cogan Shimizu This is my paper
classification cs.CYcs.AI
keywords EducationOntologyKnowledgeGraphOWLCurriculumLearningPathPersonaCompetencyQuestions
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper aims to solve the fragmentation of online learning materials by proposing the Curriculum KG Ontology, a formal OWL 2 DL schema that models curricula as structured arrangements of modules, learning paths, audience personas, and referenced media. The authors build a materialized knowledge graph from real curriculum data for an open-knowledge-network education gateway, then test it against competency questions posed by educators, encoded as SPARQL queries. If the approach holds, it gives learners a seamless cross-platform progression through materials and gives AI agents a queryable, explainable structure for finding educational content.

What carries the argument

The load-bearing object is the schema itself: the Curriculum KG Ontology in OWL 2 DL, assembled from three modules—Persona, Learning Path, and Module—following a modular design methodology that reuses ontology design patterns for agent roles, ordered sequences, and explicit typing of controlled vocabularies. The argument is carried by the materialization pipeline, which reads tabular curriculum data and emits RDF triples, and by the evaluation loop, in which competency questions are encoded as SPARQL queries and executed over the resulting graph to show that the ontology supports the promised kinds of discovery. The key identities that make the schema expressive are the strict cardinality constraints on learning steps (each has at most one next and one previous step, and each refers to exactly one module) and the separation of a Module from the Media it references.

What would settle it

Manually recompute the answers to one of the published queries, say which topics have the most associated media, by tallying the rows of the source tabular data by hand; if the SPARQL result over the materialized graph disagrees with that manual count, the pipeline is not faithfully representing the data. Alternatively, take a new curriculum dataset not used by the authors, run it through the public pipeline, and have educators who have never seen the ontology pose questions; a large share of unanswered questions would refute the claim that the schema adequately models curricula.

Watch

Extended reading notes

Core claim

The paper claims that the Curriculum KG Ontology gives a sufficient and flexible schema for densely interlinking educational material: a Curriculum is a titled collection of Modules, each Module covers Topics and references Media, and LearningPaths order those Modules into sequences that are scoped by a Curriculum and determined by a Persona. Personas are audience archetypes (developers, executives, graduate students, and so on), and each determines exactly one learning path, so different audiences get different ordered routes through the same material. The authors materialize the CurrKG from real curriculum data—including an open curriculum for knowledge graphs and related educational datasets—and validate it by translating a set of competency questions from educators into SPARQL queries that run over the graph. They report that the queries return meaningful results whenever the underlying data is present, which they take as evidence that the ontology captures the required semantics, hierarchies, and properties.

Load-bearing premise

The argument hinges on the assumption that the competency questions obtained from educators are genuinely representative of what users need, and that answering them via the authors' own SPARQL translations is a real test of adequacy; the paper reports neither the elicitation protocol nor any pass/fail scoring.

Editorial extensions

If this is right

  • Learning materials scattered across platforms can be cross-linked through shared topics, curricula, and media references, so a learner can move through a learning path without hunting for each resource.
  • The ontology encodes prerequisite and ordering information, so a persona's next step can be derived by query: the current learning step points to its successor via the next-step relation.
  • The materialization pipeline degrades gracefully: when a field is missing, it skips the corresponding triple but keeps the rest of the row, so partial data still yields a usable graph.
  • Because controlled vocabularies like persona type, level, and audience are explicitly typed individuals, the set of allowed values can be extended without restructuring the class hierarchy.
  • The graph's queryability makes it a ready retrieval target for explainable AI systems: an agent can issue SPARQL queries, get structured answers, and cite the exact graph statements behind them.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The validation is only as strong as the competency questions: the paper does not report how they were collected or scored, so an independent check would be to have educators who have never seen the ontology write questions and see whether the graph answers them.
  • A natural next test is to run the same ontology and pipeline over curricula outside computer science; if the three modules remain sufficient for, say, biology or history, the framework's generality is supported, otherwise new modules or properties would be needed.
  • The paper's implicit position is that module-level granularity is the right level for interlinking educational material; a comparison with approaches that link inside documents could show whether that granularity best supports learning.
  • If the graph is used as a retrieval target for AI tutors, the assumption is that SPARQL-retrieved modules are genuinely useful next steps; a user study measuring learning outcomes after following the graph's recommendations would test that.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper presents the Curriculum KG Ontology, an OWL 2 DL ontology for representing curricula, personas, learning paths, learning steps, modules, media, events, and topics, and reports a materialized RDF graph for the Proto-OKN EduGate use case. The ontology is developed with Modular Ontology Modeling and reuses the AgentRole and Sequence ontology design patterns. The central claim is that the ontology and materialized graph are validated by competency questions sourced from domain experts and educators and expressed as SPARQL queries over the graph.

Significance. If the validation claim were fully supported, the paper would be a useful ontology-engineering contribution: the modular design and explicit reuse of ODPs are methodologically sound, the materialization pipeline is described concretely, and the open repository makes the artifacts available. The persona-centered learning-path modeling is a plausible and reusable design choice for educational knowledge graphs. The paper's significance is currently undercut, however, by the lack of reproducible evaluation evidence and by inconsistencies between the SPARQL queries and the axiomatization.

major comments (3)
  1. [Section 3.2, CQ1-CQ5] The SPARQL queries displayed in Section 3.2 use property names that do not appear in the axiomatization of Section 2.3. CQ1 queries edu-ont:hasLearningSteps, while axioms 7 and 13 describe the property as hLS/hasLearningStep; CQ2 queries edu-ont:hasAuthor, while axiom 3 uses assumesAuthorship; CQ3 applies edu-ont:coversTopic to Media, while axiom 14 only states Module ⊑ ∃coversTopic.Topic; CQ4 uses edu-ont:belongsToCategory, while axiom 17 uses belongsTo. Since the validation claim in the abstract rests on these queries exercising the ontology, the manuscript must either align the queries with the declared vocabulary or explicitly document any additional properties in the ontology file. As written, the reader cannot tell whether the queries exercise the ontology's axioms or traverse ad-hoc triples.
  2. [Section 3.2, evaluation results] No query results, counts, bindings, or pass/fail criteria are reported for any competency question. The text asserts that executing the SPARQL queries "helped confirm that the ontology sufficiently captured the semantics..." but does not show that the queries returned non-empty answers or correct answers. This is load-bearing: the abstract's claim that the materialized graph is "validated" cannot be assessed from the paper. Adding a table of per-CQ results (e.g., number of result rows and sample answers) is necessary.
  3. [Section 3.2, competency question provenance] The CQs are described as sourced from domain experts and educators, but no elicitation protocol, number or affiliation of experts, or list of the full CQ set is given; the complete set is only pointed to in a repository. Moreover, the SPARQL translations are written in the ontology's own vocabulary by the authors, making the evaluation largely self-referential. I am not arguing that this is circularity by construction, but as reported the external grounding of the CQs is unverifiable. The authors should include the full CQ list and describe how each CQ was obtained and how the SPARQL translation was checked.
minor comments (6)
  1. [Figure 2 and Section 2.3] Figure 2 caption says "yellow boxes indicate the classes" while Section 2.3 says "Gold boxes represent classes"; please unify the terminology and check the legend against the actual diagram.
  2. [Section 2.3, axiom (26)] Axiom (26), "Topic ⊑ =1asString xsd:String", mixes the DL object-property notation with a datatype property; this should be written consistently (e.g., Topic ⊑ =1 asString.xsd:string).
  3. [Section 3.2] The text says the rest of the CQs are available in the git repository, but the URL is not given in that section; include the full CQ list in the paper or an appendix, and cite the repository explicitly at the point of reference.
  4. [Section 4] There is a typo in "Developed by the Accocication for Computing Machinery"; it should be "Association for Computing Machinery".
  5. [References] Reference [8] is cited with "To appear, available from http://www.semantic-web-journal.net"; this is a published article and should be updated to its final bibliographic details.
  6. [Section 2.3, axioms 9-10] The global cardinality restriction ⊤ ⊑ ≤1 hNLS.⊤ is stronger than the prose "exactly one or no next learning step" if the intended reading is per learning path; clarify whether the restriction is global or scoped.

Circularity Check

0 steps flagged · score 1.0 of 10

No circular derivation: the ontology and materialized graph are presented as direct modeling artifacts, and the CQ/SPARQL mismatches are reproducibility and correctness concerns rather than tautological reductions.

full rationale

The paper's central derivation is an ontology design plus a materialized KG and a competency-question validation. There is no fitted parameter renamed as a prediction and no equation-level reduction of the claimed result to its inputs. The ontology axioms in Section 2.3 are stated directly from domain modeling decisions (Persona, LearningPath, Module, Curriculum, etc.), with methodology choices (MOMo, ODPs from MODL, OPLa) cited from the authors' own prior work; these self-citations are procedural and do not smuggle in the validation result or force the ontology's content. The CQ validation in Section 3.2 is claimed to use competency questions 'sourced from domain experts and educators'; if taken at face value, this is external grounding, and the translation to SPARQL is the authors' operationalization. The absence of per-query result counts and the apparent mismatches between query predicates (hasLearningSteps, hasAuthor, coversTopic on Media, belongsToCategory) and the axiomatization (hLS, assumesAuthorship, coversTopic on Module, belongsTo) are serious reproducibility and correctness issues, but they do not make the derivation circular: they make the reported validation unverifiable, not tautological. Accordingly, no circular step can be exhibited with a quote-and-reduction, and the score is low. The self-citations are not load-bearing for the paper's main claim that the ontology can structure and interlink educational material.

Assumptions & free parameters 0 free parameters · 7 assumptions · 5 invented entities

No numerical free parameters are fitted; design choices such as the list of Personas (Developers, Instructors, Analysts, Executives, Graduate Students) and Category taxonomy (Foundation, Survey, Methodology, Standard) are hand-selected modeling decisions, not fitted values. The ontology relies on standard W3C semantics and imported design patterns, with several ad hoc cardinality constraints.

assumptions (7)
  • standard math OWL 2 DL and RDF/SPARQL W3C standards provide the formal semantics.
    The ontology is asserted in OWL 2 DL and queried via SPARQL; these standards are assumed throughout Sections 2 and 3.
  • domain assumption AgentRole ODP from the MODL library is appropriate for modeling Persona and Author roles.
    Section 2.2 states the Persona module is based on the AgentRole pattern; this choice is imported from prior literature without further justification.
  • domain assumption Sequence and Semantic Trajectory ODPs are appropriate for modeling LearningPaths.
    Section 2.2 and 2.3 state the LearningPath is based on the Sequence and Semantic Trajectory patterns; this is an adopted modeling assumption.
  • ad hoc to paper Every Curriculum must have at least one Module (axiom 2).
    This is a stated modeling decision in Section 2.3 (Curriculum) that 'a curriculum is pointless... if there are no modules'; it is not forced by external standards.
  • ad hoc to paper Every LearningStep refers to exactly one Module (axiom 13).
    Section 2.3 (LearningStep) imposes a strict cardinality; this is an author-defined constraint.
  • ad hoc to paper Every Persona determines exactly one LearningPath (axiom 24).
    Section 2.3 (Persona) imposes strict cardinality 'as each Persona is very granular'; this is a domain modeling choice, not independently validated.
  • domain assumption subeventOf is a subproperty of po-feature and has meronymous transitivity (from Winston part-whole relations).
    Section 2.3 (Event) imports this property from reference [35]; it is a domain assumption about part-whole relations.
invented entities (5)
  • Persona class
    purpose: Represents audience-specific learner types (developer, instructor, executive, etc.) that determine learning paths.
    Introduced as a central modeling construct in Section 2.3; no external falsifiable handle exists for this class.
  • LearningPath class
    purpose: Represents a dynamic sequence of LearningSteps tailored to a Persona.
    Introduced as a core class based on Sequence/Semantic Trajectory patterns; it is an ontology modeling construct.
  • LearningStep class
    purpose: Represents an individual step in a LearningPath, linked to exactly one Module.
    New class defined in Section 2.3 with strict cardinality; no independent empirical evidence.
  • Module class
    purpose: Represents a formal educational unit covering topics and referencing Media.
    Defined as distinct from Media in Section 2.3; modeling construct with no external validation.
  • Media class
    purpose: Represents documents, videos, and other resources referenced by Modules and Events.
    A general class for resources; not a new physical entity, but a modeling construct.

how reviews work

0 comments
Cite this review

Pith. "Pith review of An Ontology for Representing Curriculum and Learning Material." pith.science (2026). https://pith.science/paper/PZDYM4GE

@misc{pith2026250605751,
  author       = {Pith},
  title        = {Pith review of: An Ontology for Representing Curriculum and Learning Material},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PZDYM4GE}},
  note         = {Machine review of arXiv:2506.05751}
}
read the original abstract

Educational, learning, and training materials have become extremely commonplace across the Internet. Yet, they frequently remain disconnected from each other, fall into platform silos, and so on. One way to overcome this is to provide a mechanism to integrate the material and provide cross-links across topics. In this paper, we present the Curriculum KG Ontology, which we use as a framework for the dense interlinking of educational materials, by first starting with organizational and broad pedagogical principles. We provide a materialized graph for the Prototype Open Knowledge Network use-case, and validate it using competency questions sourced from domain experts and educators.

Figures

Figures reproduced from arXiv: 2506.05751 by the authors.

Figure 1
Figure 1. Schematic comparison between the Agent Role ontol [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. The schema diagram for the overall Curriculum KG Ontology. The yellow boxes indicate the classes, the purple ones [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

39 extracted references · 36 canonical work pages

  1. [1]

    Abu-Salih and S

    B. Abu-Salih and S. Alotaibi. A systematic lit- erature review of knowledge graph construc- tion and application in education.Heliyon, 10(3), 2024

    work page 2024

  2. [2]

    Computer sci- ence curricula 2023: Curriculum guidelines for undergraduate degree programs in com- puter science, 2023

    ACM, IEEE-CS, and AAAI. Computer sci- ence curricula 2023: Curriculum guidelines for undergraduate degree programs in com- puter science, 2023. Accessed April 12, 2025

    work page 2023

  3. [3]

    Amini, S

    R. Amini, S. S. Norouzi, P. Hitzler, and R. Amini. Towards complex ontology align- ment using large language models. In S. Tiwari, B. Villazón-Terrazas, F. Ortiz- Rodríguez, and S. Sahri, editors,Knowledge Graphs and Semantic Web - 6th International Conference, KGSWC 2024, Paris, France, December 11-13, 2024, Proceedings, volume 15459 ofLecture Notes in C...

    work page 2024

  4. [4]

    https://odpa.github.io/

    Association for ontology design patterns. https://odpa.github.io/. Ac- cessed: 10 May 2025

    work page 2025

  5. [5]

    N. Y . Ayadi, S. Bernard, R. Bossy, M. Courtin, B. G. H. Happi, P. Larmande, F. Michel, C. Nedellec, C. Roussey, and C. Faron. A unified approach to publish semantic anno- tations of agricultural documents as knowl- edge graphs.Smart Agricultural Technology, 8:100484, 2024

    work page 2024

  6. [6]

    Beckett, T

    D. Beckett, T. Berners-Lee, E. Prud’hommeaux, and G. Carothers, feb 2014. W3C recommendation

    work page 2014

  7. [7]

    Blomqvist, K

    E. Blomqvist, K. Hammar, and V . Presutti. Engineering Ontologies with Patterns – The eXtreme Design Methodology. In P. Hit- zler, A. Gangemi, K. Janowicz, A. Krisnadhi, and V . Presutti, editors,Ontology Engineering with Ontology Design Patterns – F oundations and Applications, volume 25 ofStudies on the Semantic Web, pages 23–50. IOS Press, 2016

    work page 2016

  8. [8]

    Calvanese, B

    D. Calvanese, B. Cogrel, S. Komla- Ebri, R. Kontchakov, D. Lanti, M. Rezk, M. Rodriguez-Muro, and G. XIao. Ontop: Answering SPARQL queries over relational databases.Semantic Web, 2017. To appear, available from http://www.semantic-web- journal.net

    work page 2017

Show all 39 references
  1. [9]

    Coursera — Degrees, Certificates, & Free Online Courses.https://www

    Coursera. Coursera — Degrees, Certificates, & Free Online Courses.https://www. coursera.org/. Accessed: 2025-04-10. 11

    2025

  2. [10]

    https://github.com/kastle-lab/curriculum-kg

    Curriculum knowledge graph repository. https://github.com/kastle-lab/curriculum-kg

  3. [11]

    Eberhart, C

    A. Eberhart, C. Shimizu, S. Chowdhury, M. K. Sarker, and P. Hitzler. Expressibility of owl axioms with patterns. InThe Semantic Web: 18th International Conference, ESWC 2021, Virtual Event, June 6–10, 2021, Pro- ceedings, page 230–245, Berlin, Heidelberg,

    2021

  4. [12]

    A. S. Foundation.https://jena. apache.org/documentation/ fuseki2/. accessed: 10 May 2025

    2025

  5. [13]

    Grévisse, R

    C. Grévisse, R. Manrique, O. Marino, and S. Rothkugel.Knowledge Graph-Based Teacher Support for Learning Material Au- thoring, pages 177–191. Springer, 10 2018

    2018

  6. [14]

    Hammar and V

    K. Hammar and V . Presutti.Template-based content ODP instantiation, pages 1–13. IOS Press, 12 2017

    2017

  7. [15]

    Hitzler, A

    P. Hitzler, A. Gangemi, K. Janowicz, A. Kris- nadhi, and V . Presutti, editors.Ontology En- gineering with Ontology Design Patterns – F oundations and Applications, volume 25 of Studies on the Semantic Web. IOS Press, 2016

    2016

  8. [16]

    Hitzler, A

    P. Hitzler, A. Gangemi, K. Janowicz, A. A. Krisnadhi, and V . Presutti. Towards a sim- ple but useful ontology design pattern repre- sentation language. In E. Blomqvist, Ó. Cor- cho, M. Horridge, D. Carral, and R. Hoekstra, editors,Proceedings of the 8th Workshop on Ontology D...

    2017

  9. [17]

    Hitzler, M

    P. Hitzler, M. Krötzsch, B. Parsia, P. Patel- Schneider, and S. Rudolph, editors.OWL 2 Web Ontology Language: Primer (Second Edition). W3C Recommenda- tion, 11 December 2012. Available at http://www.w3.org/TR/owl2-primer/

    2012

  10. [18]

    Hitzler, M

    P. Hitzler, M. Krötzsch, and S. Rudolph. F oundations of Semantic Web Technologies. Chapman & Hall/CRC, 2010

    2010

  11. [19]

    Hogan, E

    A. Hogan, E. Blomqvist, M. Cochez, C. D’amato, G. D. Melo, C. Gutierrez, S. Kir- rane, J. E. L. Gayo, R. Navigli, S. Neu- maier, A.-C. N. Ngomo, A. Polleres, S. M. Rashid, A. Rula, L. Schmelzeisen, J. Sequeda, S. Staab, and A. Zimmermann. Knowledge graphs.ACM Comput. Surv., 54...

    2021

  12. [20]

    Ilkou, H

    E. Ilkou, H. Abu-Rasheed, D. Chaves-Fraga, and E. Engelbrecht. Teaching knowledge graph for knowledge graphs education. Under review in the Semantic Web Journal, 2024

    2024

  13. [21]

    C. D. Jaldi, E. Ilkou, N. Schroeder, and C. Shimizu. Education in the era of neu- rosymbolic ai.Journal of Web Semantics, 85:100857, 2025

    2025

  14. [22]

    C. M. Keet and Z. C. Khan. On the roles of competency questions in ontology engineer- ing. In M. Alam, M. Rospocher, M. van Erp, L. Hollink, and G. A. Gesese, editors,Knowl- edge Engineering and Knowledge Manage- ment, pages 123–132, Cham, 2025. Springer Nature Switzerland

    2025

  15. [23]

    Lewis, E

    P. Lewis, E. Perez, A. Piktus, F. Petroni, V . Karpukhin, N. Goyal, H. Küttler, M. Lewis, W.-t. Yih, T. Rocktäschel, S. Riedel, and D. Kiela. Retrieval-augmented generation for knowledge-intensive nlp tasks. InProceed- ings of the 34th International Conference on Neural Inform...

    2020

  16. [24]

    Medium — Read and Write Sto- ries.https://medium.com/

    Medium. Medium — Read and Write Sto- ries.https://medium.com/. Accessed: 2025-04-10

    2025

  17. [25]

    Motik, P

    B. Motik, P. F. Patel-Schneider, and B. Cuenca Grau, editors.OWL 2 Web Ontology Language: Direct Semantics. W3C Recommendation, 27 October 2009. Available at http://www.w3.org/TR/owl2- direct-semantics/

    2009

  18. [26]

    N. Noy, Y . Gao, A. Jain, A. Narayanan, A. Patterson, and J. Taylor. Industry-scale knowledge graphs: lessons and challenges. Commun. ACM, 62(8):36–43, July 2019

    2019

  19. [27]

    Proto-OKN — NSF Proto Open Knowledge Networks.https://www

    Proto-OKN. Proto-OKN — NSF Proto Open Knowledge Networks.https://www. proto-okn.net/. Accessed: 2025-04-10

    2025

  20. [28]

    RDFLib — Python library for working with RDF.https://rdflib

    RDFLib Team. RDFLib — Python library for working with RDF.https://rdflib. readthedocs.io/. Accessed: 2025-04- 11

    2025

  21. [29]

    Schreiber, Y

    G. Schreiber, Y . Raimond, F. Manola, E. Miller, and B. McBride, jun 2014. 12

    2014

  22. [30]

    Seaborne and E

    A. Seaborne and E. Prud’hommeaux. SPARQL query language for RDF, 2008. http://www.w3.org/TR/rdf-sparql-query/

    2008

  23. [31]

    Shimizu, K

    C. Shimizu, K. Hammar, and P. Hitzler. Modular ontology modeling.Semantic Web, 14(3):459–489, 2023

    2023

  24. [32]

    Shimizu, Q

    C. Shimizu, Q. Hirt, and P. Hitzler. Modl: A modular ontology design library. In WOP@ISWC, 2019

    2019

  25. [33]

    Shimizu and P

    C. Shimizu and P. Hitzler. Accelerating knowledge graph and ontology engineering with large language models.J. Web Semant., 85:100862, 2025

    2025

  26. [34]

    Shimizu, P

    C. Shimizu, P. Hitzler, Q. Hirt, D. Rehberger, S. G. Estrecha, C. Foley, A. M. Sheill, W. Hawthorne, J. Mixter, E. Watrall, R. Carty, and D. Tarr. The enslaved ontology: Peoples of the historic slave trade.Journal of Web Se- mantics, 63:100567, 2020

    2020

  27. [35]

    Shimizu, P

    C. Shimizu, P. Hitzler, and C. Paul. Ontol- ogy design patterns for winston’s taxonomy of part-whole relations. InEmerging Topics in Semantic Technologies, pages 119–129. IOS Press, 2018

    2018

  28. [36]

    Singh, A

    A. Singh, A. Ehtesham, S. Kumar, and T. T. Khoei. Agentic retrieval-augmented genera- tion: A survey on agentic rag.arXiv preprint arXiv:2501.09136, 2025

    2025 arXiv

  29. [37]

    Tahsin, Y

    R. Tahsin, Y . Li, M. S. Abolhasani, and F. Ameri. Generation of semantic knowl- edge graphs from maintenance work orders data.J. Maintenance Engineering, 11(2):45– 60, 2024

    2024

  30. [38]

    L. Wang, C. Ma, X. Feng, Z. Zhang, H. Yang, J. Zhang, Z. Chen, J. Tang, X. Chen, Y . Lin, et al. A survey on large language model based autonomous agents.Frontiers of Computer Science, 18(6):186345, 2024

    2024

  31. [39]

    Zalewski, L

    J. Zalewski, L. Zhou, C. Shimizu, and P. Hit- zler. Ink browser - the interactive knowledge browser. In O. Seneviratne, C. Pesquita, J. Se- queda, and L. Etcheverry, editors,Proceed- ings of the ISWC 2021 Posters, Demos and In- dustry Tracks: From Novel Ideas to Industrial Pra...

    2021

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.