More and more knowledge graphs are constructed for private use, e.g., the Amazon Product Graph [1] or the Fashion Knowledge Graph by Zalando 1 ,or public use, e.g., DBpedia 2 or Wikidata 3 . While techniques to automatically construct KGs from existing Web objects exist (e.g., scraping Web tables), there is still room for improvement. So far, constructing knowledge graphs was considered an engineering task, however, more scientifically robust methods keep on emerging. These methods were widely questioned for their verbosity, low performance or difficulty of use, while the data sources' variety and complexity cause further syntax and semantic interoperability issues.

Declarative methods (mapping languages) for describing rules to construct knowledge graphs and approaches to execute those rules keep on emerging. Nevertheless constructing knowledge graphs is still not a straightforward task because several existing challenges remain and yet the barriers to construct knowledge graphs are not lowered enough to be easily and broadly adopted by industry. These reasons and the vastly populated knowledge graph construction W3C Community Group 4 show that there are still open questions that require further investigation to come up with groundbreaking solutions.

Addressing challenges related to knowledge graphs construction requires well-founded research, including the investigation of concepts and development of tools as well as methods for their evaluation. R2RML was recommended in 2012 by W3C, and since then, different extensions, alternatives and implementations were proposed [2,3,4]. Certain approaches followed the ETL-like paradigm, e.g., SDM-RDFizer [5], RocketRML [6], and FunMap [7], while Fifth International Workshop On Knowledge Graph Construction Co-located with the ESWC 2024, 27th May 2024, Crete, Greece Envelope david.chaves@upm.es (D. Chaves-Fraga); anastasia.dimou@kuleuven.be (A. Dimou); ana.iglesiasm@upm.es (A. Iglesias-Molina); umutcan.serles@sti2.at (U. Serles); dylan.van.assche@ugent.be (D. V. Assche) others the query-answering paradigm, e.g., Ultrawrap [8], Morph-RDB [9] and Ontop [10]. Besides R2RML-based extensions, alternatives were proposed, e.g., SPARQL-Generate [11] and ShExML [12], as well as methods to perform data transformations while constructing knowledge graphs, e.g., FnO [13] and FunUL [14].

The fifth edition of the knowledge graph construction workshop5 has a special focus on time on novel techniques, frameworks, architectures, and tools for the new extensions of RML such as RDF Collections and Containers, and RDF-Star support and the 2023 release of the RDF Mapping Language (RML) [15] in general. It also included:

• Keynote. The workshop includes the keynote from Lionel Tailhardat (Orange): "Anomaly Detection For Telco Companies: Challenges And Opportunities In Knowledge Graph Construction" • The Second Knowledge Graph Construction Challenge. The edition of this year's challenge has a double objective: benchmarking systems to (i) find which RDF graph construction system optimizes for metrics i.e. execution time, CPU and memory usage; and (ii) how compliant are they with the 2023 revision of RML and its new modules.

The final goal of the event is to provide a venue for scientific discourse, systematic analysis and rigorous evaluation of languages, techniques and tools, as well as practical and applied experiences and lessons-learned for constructing knowledge graphs from academia and industry.

Eight papers were submitted. The reviews were open and public, and hosted at Open Review 6 . Each paper received at least three reviews from reviewers with different background and status. Each paper received a review from a senior, a junior and an industry researcher.

Five papers were accepted and one was conditionally accepted. Five of the accepted papers were long papers and one was a short paper. The following papers were accepted for publication and presented at the workshop:

• Not Everybody Speaks RDF: Knowledge Conversion between Different Data Representations [16]. • BURPing Through RML Test Cases [17].

Graphs [18]. • RML-view-to-CSV: A Proof-of-Concept Implementation for RML Logical Views [19].

• R2[RML]-ChatGPT Framework [20]. • Towards Self-Configuring Knowledge Graph Construction Pipelines using LLMs -A Case Study with RML [21].

During the workshop, the second edition of the Knowledge Graph Construction Challenge was organized with two different tracks: (i) conformance with the new RML modules, and (ii) performance of engines on the same hardware.

The first track around conformance with the new RML modules encouraged developers of RML engines to support the specifications of the new RML modules by evaluating their engines against 365 test cases provided by the maintainers of each RML module. RML-Core (238 test cases), which focus on the core parts of RDF generation, provides the biggest number of test cases, followed by RML-IO (67 test cases) to access various data sources and targets. Data transformations with FnO were also present through the RML-FNML (13 test cases) module. Newer modules e.g. RML-Star (18 test cases) for RDF-Star support and RML-CC (29 test cases) to generate RDFS Collections & Containers provided new challenges for existing engines as they impact the RDF generation process. We had 5 participating engines for the first track: RMLMapper [2], SDM-RDFizer [5], mapping-template [16], RPT/SANSA [22], and BURP [17].

The second track around performance was similar to the previous edition except that now each participant had access to a common hardware environment. This way, each engine had the same restrictions regarding CPU and RAM. Through this track, we wanted to not only focus on execution time but also resource consumption of each engine. This track consisted of 2 parts: (i) artificial data for analyzing specific parameters of the construction process e.g. joins, data size, mappings, and (ii) real-life data of the GTFS Madrid Benchmark to evaluate approaches in real use cases. We had 6 participating engines for the second track: mappingtemplate [16], FlexRML [23], RMLWeaver-js [24], RPT/Sansa [22], RMLStreamer [25], and RML-view-to-CSV+RMLStreamer [19].

Several participants also submitted a report of their participation in one or both tracks. The following reports are included in the proceedings:

• RMLStreamer supported by RML-view-to-CSV in the Performance Track of the KGCW Challenge 2024 [26]. • RMLWeaver-JS: An Algebraic Mapping Engine in the KGCW Challenge 2024 [24].

• Performance Results of FlexRML in the KGCW Challenge 2024 [27].

• Backwards or Forwards? [R2]RML Backwards Compatibility in RMLMapper [28].

• The Conformance of an RML Processor Built from Scratch to Validate RML Specifications and Test Cases [29]. • Results for Knowledge Graph Creation Challenge 2024: SDM-RDFizer [30].

• KGCW2024 Challenge Report: RDFProcessingToolkit [31].