Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
219
datasets available to search
ShareScore release 0.7.1
Dataset results
219 results for “knowledge graphs”
MUHAI Benchmark : Task 1 (Short story generation with Knowledge Graphs)
<p><strong>Meaning and Understanding in Human-Centric AI (MUHAI) Benchmark<br> Task 1 (Short story generation with Knowledge Graphs and Language Models)</strong> </p> <p>The dataset can be used to test understandability of text generated through the combination of knowledge graphs and language models without using knowledge graph embeddings.<br> <br> The task here is to generate 5-sentence stories from a set of <em>subject-predicate-object</em> triples that are extracted from a knowledge graph. Two steps need to be performed:</p> <p>1. Language model fine-tuning (SVO triple extraction + model fine-tuning)<br> 2. Story generation (knowledge enrichment + text generation) <br> <br> The submission includes the following data:</p> <ol> <li>Original ROC stories corpus (100 stories)</li> <li>ROC stories encoded with relevant triples (extracted through SpaCy, 2 versions, with and without coreference resolution)</li> <li>Stories generated by the pre-trained model (GPT2-simple)</li> <li>Stories generated by the fine-tuned model (DICE + ConceptNet + DBpedia )</li> <li>Stories generated by the fine-tuned model (DICE + ConceptNet + DBpedia + WordNet )</li> <li>Stories generated by the GPT-2-keyword-generation (an open-source software that uses GPT-2 to generate text pertaining to the specified keywords)</li> <li>Model results</li> <li>Evaluation metrics description</li> <li>User-evaluation questionnaire </li> </ol> <p>Code : https://github.com/kmitd/muhai-dice_story</p>
The OREGANO knowledge graph for computational drug repurposing
<p>The files here are data files from the OREGANO project, which consists of building a holistic knowledge graph on drugs, including natural compounds. Here is the list of files:</p><p> </p><p>- OREGANO_V2.tsv : The triplet file used for link prediction. 3 columns : Subjet ; Predicate ; Object</p><p>- oreganov2.1_metadata_complet.ttl : The OREGANO knowledge graph in turtle format with the names and cross-references of the various integrated entities.</p><p> </p><p>The following files contain the cross-references of OREGANO entities according to their type. They are all organised as follows: the external sources are the titles of the columns and each line begins with the identifier of the entity in OREGANO :</p><p>- TARGET.tsv: Cross-reference table of the 22,096 targets.<br>- PHENOTYPES.tsv: Cross-reference table of the 11,605 phenotypes.<br>- DISEASES.tsv: Cross-reference table of the 18,333 diseases.<br>- PATHWAYS.tsv: Cross-reference table of the 2,129 pathways.<br>- GENES.tsv: Cross-reference table of the 35,794 genes.<br>- COMPOUND.tsv: Cross-reference table of the 90,868 compounds.<br>- INDICATIONS.tsv: Cross-reference table of the 2,714 indications.<br>- SIDE_EFFECT.tsv: Cross-reference table of the 6,060 side-effects.<br>- ACTIVITY.tsv: Names of the 78 activities.<br>- EFFECT.tsv: Names of the 171 effects.</p><p>The OREGANO knowledge graph is composed of 11 types of nodes and 19 types of links. The current version of the graph contains 88,937 nodes and 824,231 links.</p><p>A SPARQL endpoint has been provided to enable users to retrieve and explore the knowledge graph at <a href="http://91.121.148.199:8889/bigdata/#query">OREGANO SPARQL endpoint</a> .</p><p> </p><p>The integration files and the knowledge graph are available on the GitHub of the OREGANO project in the Integration folder: <a href="https://gitub.u-bordeaux.fr/erias/oregano">Gitub repository</a> .</p>
Event-QA: A Dataset for Event-Centric Question Answering over Knowledge Graphs
<p>Event-QA dataset contains 1000 semantic queries and the corresponding verbalisations for EventKG - a recently proposed event-centric knowledge graph containing over 970 thousand events.</p>
The OREGANO knowledge graph for computational drug repurposing
<p>The files here are data files from the OREGANO project, which consists of building a holistic knowledge graph on drugs, including natural compounds. Here is the list of files:</p><p> </p><p>- OREGANO_V2.tsv : The triplet file used for link prediction. 3 columns : Subjet ; Predicate ; Object</p><p>- oreganov2.1_metadata_complet.ttl : The OREGANO knowledge graph in turtle format with the names and cross-references of the various integrated entities.</p><p> </p><p>The following files contain the cross-references of OREGANO entities according to their type. They are all organised as follows: the external sources are the titles of the columns and each line begins with the identifier of the entity in OREGANO :</p><p>- TARGET.tsv: Cross-reference table of the 22,096 targets.<br>- PHENOTYPES.tsv: Cross-reference table of the 11,605 phenotypes.<br>- DISEASES.tsv: Cross-reference table of the 18,333 diseases.<br>- PATHWAYS.tsv: Cross-reference table of the 2,129 pathways.<br>- GENES.tsv: Cross-reference table of the 35,794 genes.<br>- COMPOUND.tsv: Cross-reference table of the 90,868 compounds.<br>- INDICATIONS.tsv: Cross-reference table of the 2,714 indications.<br>- SIDE_EFFECT.tsv: Cross-reference table of the 6,060 side-effects.<br>- ACTIVITY.tsv: Names of the 78 activities.<br>- EFFECT.tsv: Names of the 171 effects.</p><p>The OREGANO knowledge graph is composed of 11 types of nodes and 19 types of links. The current version of the graph contains 88,937 nodes and 824,231 links.</p><p>A SPARQL endpoint has been provided to enable users to retrieve and explore the knowledge graph at <a href="http://91.121.148.199:8889/bigdata/#query">OREGANO SPARQL endpoint</a> .</p><p> </p><p>The integration files and the knowledge graph are available on the GitHub of the OREGANO project in the Integration folder: <a href="https://gitub.u-bordeaux.fr/erias/oregano">Gitub repository</a> .</p><p> </p>
Dataset for Creating a Scholarly Knowledge Graph from Survey Article Tables
<ul> <li><strong>Selected papers.csv</strong><br> This file lists all selected survey papers used to create the knowledge graph. The file contains paper titles, table references (of the tables that are extracted), sources and the reference to the survey paper (either a DOI, or a full textual reference)<br> </li> <li><strong>ORKG comparisons.csv</strong><br> All comparisons imported in the ORKG are listed in this file. Per survey paper, multiple tables could be extracted, and therefore multiple comparisons are created. The files lists the internal IDs and the URL to the comparisons. <br> </li> <li><strong>Ingested papers.csv</strong><br> A full list of all individual papers extracted from the survey articles. The file contains paper titles and their respective URL in the ORKG. </li> </ul>
Node2Vec model - Czech Wikidata (knowledge graph / concepts / l80 / rw40)
<p>Node2Vec embedding model trained on Czech wikidata (from October 2020) concepts using gensim implementation of Word2Vec with the following parameters for random walks:</p> <ul> <li>length of walk = 80</li> <li>number of random walks = 40</li> </ul>
Node2Vec model - Czech Wikidata (knowledge graph / concepts / l40 / rw10)
<p>Node2Vec embedding model trained on Czech wikidata (from October 2020) concepts using gensim implementation of Word2Vec with the following parameters for random walks:</p> <ul> <li>length of walk = 40</li> <li>number of random walks = 10</li> </ul>
Node2Vec model - Czech Wikidata (knowledge graph / labels / l80 / rw40)
<p>Node2Vec embedding model trained on Czech wikidata (from October 2020) labels using gensim implementation of Word2Vec with the following parameters for random walks:</p> <ul> <li>length of walk = 80</li> <li>number of random walks = 40</li> </ul>
Node2Vec model - Czech Wikidata (knowledge graph / labels / l160 / rw40)
<p>Node2Vec embedding model trained on Czech wikidata (from October 2020) labels using gensim implementation of Word2Vec with the following parameters for random walks:</p> <ul> <li>length of walk = 160</li> <li>number of random walks = 40</li> </ul>
Supplementary data files for manuscript titled "From spreadsheet lab data templates to knowledge graphs: A FAIR data journey in the domain of AMR research"
<div>This data repository contains all the necessary supplementary files for the manuscript titled "<strong>From spreadsheet lab data templates to knowledge graphs: A FAIR data journey in the domain of AMR research.</strong>"</div> <div> </div> <div>The repository is a copy of the <a href="https://github.com/IMI-COMBINE/template2graphs">GitHub page</a> with the source code used to generate the graph and additional files required for the Lab Data Template.</div> <div> </div> <div>Below we provide a brief overview of the data files in the `additional folder` and their underlying purpose:</div> <div> <ul> <li>The <strong>Data Survey</strong> collects relevant project and data set information to set up a Data Management Plan. It can serve as an input for Lab Data Template development.</li> <li>The <strong>Lab Data Templates</strong> facilitate the collection of AMR research data (in vivo and in vitro) in several sub-tables. The Excel format is compatible with upload procedures into the data repository 'grit' and serves as input for a knowledge graph workflow.</li> <li>The <strong>Data dictionary</strong> is connected to the Lab Data Templates and ensures harmonized data entries. In addition, the dictionaries collect metadata beyond the content of the Lab Data Template (e.g. bacterial strain information or compound information) and link to ontologies where possible.</li> <li>The <strong>FAIR assessments</strong> have been used as a primer for improving the template. This report is generated using the FAIR-DSM model.</li> </ul> </div> <div>The templates have been used during the IMI2 GNA NOW project to collect information and have been improved according to FAIR standards in collaboration with the IMI FAIRplus project ("post FAIRification").</div>
Link-prediction on Biomedical Knowledge Graphs
<p>Release of code and experimental data from the paper <em>Towards Linking Graph Topology to Model Performance for Biomedical Knowledge Graph Completion </em>(<em>Machine Learning for Life and Material Sciences</em> workshop @ ICML2024) and <a href="https://arxiv.org/abs/2409.04103" rel="nofollow">The Role of Graph Topology in the Performance of Biomedical Knowledge Graph Completion Models</a>.</p> <div> <div>Knowledge Graph Completion has been increasingly adopted as a useful method for several tasks in biomedical research, like drug repurposing or drug-target identification. To that end, a variety of datasets and Knowledge Graph Embedding models has been proposed over the years. However, little is known about the properties that render a dataset useful for a given task and, even though theoretical properties of Knowledge Graph Embedding models are well understood, their practical utility in this field remains controversial. We conduct a comprehensive investigation into the topological properties of publicly available biomedical Knowledge Graphs and establish links to the accuracy observed in real-world applications. By releasing all model predictions we invite the community to build upon our work and continue improving the understanding of these crucial applications.</div> <div> </div> <div>Experiments were conducted on six datasets: five from the biomedical domain (<a href="../records/268568">Hetionet</a>, <a href="https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/IXA7BM">PrimeKG</a>, <a href="../records/4077338">PharmKG</a>, <a href="../records/5361324">OpenBioLink2020 HQ</a>, <a href="../records/7011027">PharMeBINet</a>) and one trivia KG (<a href="https://aclanthology.org/W15-4007.pdf">FB15k-237</a>). All datasets were randomly split into training, validation and test set (80% / 10% / 10%; in the case of PharMeBINet, 99.3% / 0.35% / 0.35% to mitigate the increased inference cost on the larger dataset).</div> <div>On each dataset, five different KGE models were compared: <a href="https://dl.acm.org/doi/10.5555/2999792.2999923">TransE</a>, <a href="https://arxiv.org/abs/1412.6575">DistMult</a>, <a href="https://arxiv.org/abs/1902.10197">RotatE</a>, <a href="https://arxiv.org/abs/2209.08271">TripleRE</a>, <a href="https://dl.acm.org/doi/10.5555/3504035.3504256">ConvE</a>. Hyperparameters were tuned on the validation split (see final train configurations in <code>train/scripts</code>). We release results for tail predictions on the test split. In particular, each test query <code>(h,r,?)</code> is scored against all entities in the KG and we compute the rank of the score of the correct completion <code>(h,r,t)</code> , after masking out scores of other <code>(h,r,t')</code> triples contained in the graph.</div> <div>Note: the ranks provided are computed as the average between the optimistic and pessimistic ranks of triple scores.</div> <div> </div> <div>Inside <code>experimental_data.zip</code>, the following files are provided.</div> <div> <ul> <li><code>datasets/{dataset}</code>: a folder for each dataset, containing <ul> <li><code>{dataset}_preprocessing.ipynb</code>: a Jupyter notebook for downloading and preprocessing the datasets. In particular, this generates the custom label->ID mapping for entities and relations, and the numerical tensor of <code>(h_ID,r_ID,t_ID)</code> triples for all edges in the graph, which can be used to compute graph topological metrics (e.g., using <a href="https://github.com/graphcore-research/kg-topology-toolbox">kg-topology-toolbox</a>) and compare them with the edge prediction accuracy.</li> <li><code>test_ranks.csv</code>: csv table with columns <code>["h", "r", "t"]</code> specifying the head, relation, tail IDs of the test triples, and columns <code>["DistMult", "TransE", "RotatE", "TripleRE", "ConvE"]</code> with the rank of the ground-truth tail in the ordered list of predictions made by the five KGE models;</li> <li><code>entity_dict.csv</code>: list of entity labels, ordered by entity ID (as generated in the preprocessing notebook);</li> <li><code>relation_dict.csv</code>: list of relation labels, ordered by relation ID (as generated in the preprocessing notebook).</li> </ul> </li> <li><code>train</code>: code to reproduce training (and validation) of the five KGE models, using the <a href="https://github.com/graphcore-research/bess-kge">BESS-KGE</a> distribution framework. <ul> <li><code>train/scripts</code>: executable scripts, with specifications of the final hyperparameters for all models and datasets.</li> </ul> </li> <li><code>notebooks</code>: Jupyter notebooks for data analysis and generation of all the figures in the paper.</li> </ul> <p>The separate <code>top_100_tail_predictions.zip</code> archive contains, for each of the test queries in the corresponding <code>test_ranks.csv</code> table, the IDs of the top-100 tail predictions made by each of the five KGE models, ordered by decreasing likelihood. The predictions are released in a <code>.npz</code> archive of numpy arrays (one array of shape <code>(n_test_triples, 100)</code> for each of the KGE models). </p> </div> </div>
NeSy4VRD: A Multifaceted Resource for Neurosymbolic AI Research using Knowledge Graphs in Visual Relationship Detection
<p><strong>NeSy4VRD</strong></p> <p>NeSy4VRD is a multifaceted, multipurpose resource designed to foster neurosymbolic AI (NeSy) research, particularly NeSy research using Semantic Web technologies such as OWL ontologies, OWL-based knowledge graphs and OWL-based reasoning as symbolic components. The NeSy4VRD research resource pertains to the <em>computer vision</em> field of AI and, within that field, to the application tasks of <em>visual relationship detection (VRD) and scene graph generation</em>.</p> <p>Whilst the core motivation of the NeSy4VRD research resource is to foster computer vision-based NeSy research using Semantic Web technologies such as OWL ontologies and OWL-based knowledge graphs, AI researchers can readily use NeSy4VRD to either: 1) pursue computer vision-based NeSy research without involving Semantic Web technologies as symbolic components, or 2) pursue computer vision research without NeSy (i.e. pursue research that focuses purely on deep learning alone, without involving symbolic components of any kind). This is the sense in which we describe NeSy4VRD as being <em>multipurpose</em>: it can readily be used by diverse groups of computer vision-based AI researchers with diverse interests and objectives.</p> <p>The NeSy4VRD research resource in its entirety is distributed across two locations: Zenodo and GitHub.</p> <p> </p> <p><strong>NeSy4VRD on Zenodo: the NeSy4VRD dataset package</strong></p> <p>This entry on Zenodo hosts the <em>NeSy4VRD dataset package</em>, which includes the <em>NeSy4VRD dataset</em> and its companion <em>NeSy4VRD ontology</em>, an OWL ontology called VRD-World.</p> <p>The <em>NeSy4VRD dataset</em> consists of an image dataset with associated visual relationship annotations. The images of the <em>NeSy4VRD dataset</em> are the same as those that were once publicly available as part of the <a href="https://cs.stanford.edu/people/ranjaykrishna/vrd/">VRD</a> dataset. The NeSy4VRD visual relationship annotations are a highly customised and quality-improved version of the original VRD visual relationship annotations. The <em>NeSy4VRD dataset</em> is designed for computer vision-based research that involves detecting objects in images and predicting relationships between ordered pairs of those objects. A visual relationship for an image of the <em>NeSy4VRD dataset</em> has the form <'subject', 'predicate', 'object'>, where the 'subject' and 'object' are two objects in the image, and the 'predicate' describes some relation between them. Both the 'subject' and 'object' objects are specified in terms of bounding boxes and object classes. For example, representative annotated visual relationships are <'person', 'ride', 'horse'>, <'hat', 'on', 'teddy bear'> and <'cat', 'under', 'pillow'>.</p> <p>Visual relationship detection is pursued as a computer vision application task in its own right, and as a building block capability for the broader application task of scene graph generation. Scene graph generation, in turn, is commonly used as a precursor to a variety of enriched, downstream visual understanding and reasoning application tasks, such as image captioning, visual question answering, image retrieval, image generation and multimedia event processing.</p> <p>The <em>NeSy4VRD ontology</em>, VRD-World, is a rich, well-aligned, companion OWL ontology engineered specifically for use with the <em>NeSy4VRD dataset.</em> It directly describes the domain of the <em>NeSy4VRD dataset</em>, as reflected in the NeSy4VRD visual relationship annotations. More specifically, all of the object classes that feature in the NeSy4VRD visual relationship annotations have corresponding classes within the VRD-World OWL class hierarchy, and all of the predicates that feature in the NeSy4VRD visual relationship annotations have corresponding properties within the VRD-World OWL object property hierarchy. The rich structure of the VRD-World class hierarchy and the rich characteristics and relationships of the VRD-World object properties together give the VRD-World OWL ontology rich inference semantics. These provide ample opportunity for OWL reasoning to be meaningfully exercised and exploited in NeSy research that uses OWL ontologies and OWL-based knowledge graphs as symbolic components. There is also ample potential for NeSy researchers to explore supplementing the OWL reasoning capabilities afforded by the VRD-World ontology with Datalog rules and reasoning.</p> <p>Use of the <em>NeSy4VRD ontology</em>, VRD-World, in conjunction with the <em>NeSy4VRD dataset </em>is, of course, purely optional, however. Computer vision AI researchers who have no interest in NeSy, or NeSy researchers who have no interest in OWL ontologies and OWL-based knowledge graphs, can ignore the <em>NeSy4VRD ontology</em> and use the <em>NeSy4VRD dataset </em>by itself.</p> <p>All computer vision-based AI research user groups can, if they wish, also avail themselves of the other components of the NeSy4VRD research resource available on GitHub.</p> <p> </p> <p><strong>NeSy4VRD on GitHub: open source infrastructure supporting extensibility, and sample code</strong></p> <p>The NeSy4VRD research resource incorporates additional components that are companions to the <em>NeSy4VRD dataset package</em> here on Zenodo. These companion components are available at <a href="https://github.com/djherron/NeSy4VRD/">NeSy4VRD on GitHub</a>. These companion components consist of:</p> <ul> <li>comprehensive open source Python-based infrastructure supporting the extensibility of the NeSy4VRD visual relationship annotations (and, thereby, the extensibility of the <em>NeSy4VRD ontology</em>, VRD-World, as well)</li> <li>open source Python sample code showing how one can work with the NeSy4VRD visual relationship annotations in conjunction with the <em>NeSy4VRD ontology</em>, VRD-World, and RDF knowledge graphs.</li> </ul> <p>The NeSy4VRD infrastructure supporting extensibility consists of:</p> <ul> <li>open source Python code for conducting deep and comprehensive analyses of the <em>NeSy4VRD dataset</em> (the VRD images and their associated NeSy4VRD visual relationship annotations)</li> <li>an open source, custom-designed <em>NeSy4VRD protocol</em> for specifying visual relationship annotation customisation instructions declaratively, in text files</li> <li>an open source, custom-designed <em>NeSy4VRD workflow, </em>implemented using Python scripts and modules, for applying small or large volumes of customisations or extensions to the NeSy4VRD visual relationship annotations in a configurable, managed, automated and repeatable process.</li> </ul> <p>The purpose behind providing comprehensive infrastructure to support extensibility of the NeSy4VRD visual relationship annotations is to make it easy for researchers to take the <em>NeSy4VRD dataset</em> in new directions, by further enriching the annotations, or by tailoring them to introduce new or more data conditions that better suit their particular research needs and interests. The option to use the NeSy4VRD extensibility infrastructure in this way applies equally well to each of the diverse potential NeSy4VRD user groups already mentioned.</p> <p>The NeSy4VRD extensibility infrastructure, however, may be of particular interest to NeSy researchers interested in using the <em>NeSy4VRD ontology</em>, VRD-World, in conjunction with the <em>NeSy4VRD dataset. </em>These researchers can of course tailor the VRD-World ontology if they wish without needing to modify or extend the NeSy4VRD visual relationship annotations in any way. But their degrees of freedom for doing so will be limited by the need to maintain alignment with the NeSy4VRD visual relationship annotations and the particular set of object classes and predicates to which they refer. If NeSy researchers want full freedom to tailor the VRD-World ontology, they may well need to tailor the NeSy4VRD visual relationship annotations first, in order that alignment be maintained.</p> <p>To illustrate our point, and to illustrate our vision of how the NeSy4VRD extensibility infrastructure can be used, let us consider a simple example. It is common in computer vision to distinguish between <em>thing</em> objects (that have well-defined shapes) and <em>stuff</em> objects (that are amorphous). Suppose a researcher wishes to have a greater number of <em>stuff</em> object classes with which to work. Water is such a <em>stuff</em> object. Many VRD images contain water but it is not currently one of the annotated object classes and hence is never referenced in any visual relationship annotations. So adding a <em>Water</em> class to the class hierarchy of the VRD-World ontology would be pointless because it would never acquire any instances (because an object detector would never detect any). However, our hypothetical researcher could choose to do the following:</p> <ul> <li>use the analysis functionality of the NeSy4VRD extensibility infrastructure to find images containing water (by, say, searching for images whose visual relationships refer to object classes such as 'boat', 'surfboard', 'sand', 'umbrella', etc.);</li> <li>use free image analysis software (such as GIMP, at gimp.org) to get bounding boxes for instances of water in these images;</li> <li>use the <em>NeSy4VRD protocol</em> to specify new visual relationships for these images that refer to the new 'water' objects (e.g. <'boat', 'on', 'water'>);</li> <li>use the <em>NeSy4VRD workflow</em> to introduce the new object class 'water' and to apply the specified new visual relationships to the sets of annotations for the affected images;</li> <li>introduce class Water to the class hierarchy of the VRD-World ontology (using, say, the free Protege ontology editor);</li> <li>continue experimenting, now with the added benefit of the additional <em>stuff</em> object class 'water';</li> <li>contribute the enriched set of NeSy4VRD visual relationship annotations, and the enriched companion VRD-World ontology, to research communities.</li> </ul> <p> </p> <p><strong>Information pertaining to the VRD dataset</strong></p> <p>Information about the original VRD dataset is available <a href="https://cs.stanford.edu/people/ranjaykrishna/vrd/">here</a>. </p> <p>Public availability of the VRD images (via information accessible from that location) ceased sometime in the latter part of 2021. We thank Dr. Ranjay Krishna, one of the principals associated with the VRD dataset, for granting us permission to re-establish the public availability of the VRD images as part of NeSy4VRD.</p> <p>The original VRD visual relationship annotations are still publicly available from that location. But our deep analysis of those annotations, driven by our desire to design a robust companion ontology, revealed them to be highly problematic in many ways that made credible ontology modelling infeasible. They were also found to be replete with all manner of errors. The NeSy4VRD visual relationship annotations are far superior and we recommend them over the original VRD annotations to anyone contemplating conducting research using the VRD images. The NeSy4VRD annotations also have the added benefit of the rich, well-aligned companion <em>NeSy4VRD ontology</em>, VRD-World, for those whose research requires such a companion ontology.</p> <p>Researchers wishing to use the original VRD dataset may still do so. They can access the VRD images here, from within the <em>NeSy4VRD dataset</em> on Zenodo, and access the VRD visual relationship annotations from the location in the link.</p> <p><em>A note of caution</em>: the <em>NeSy4VRD ontology</em>, VRD-World, is <em>not</em><strong> </strong>compatible with the original VRD visual relationship annotations and cannot be used in conjunction with them. The VRD-World ontology has been engineered in relation to the highly customised and quality-improved NeSy4VRD visual relationship annotations. The customisations that were applied include ones that introduced many new object classes, merged some of the existing object classes, introduced one new predicate, and changed several predicate names.</p> <p>However, researchers can, if they wish, use the NeSy4VRD extensibility infrastructure (described above) to undertake their own customisation and quality-improvement exercise with respect to the original VRD visual relationship annotations. This is precisely how the NeSy4VRD visual relationship annotations were created in the first place. The primary intended use case of NeSy4VRD's extensibility infrastructure, however, is for researchers to use the NeSy4VRD visual relationship annotations as their starting point, and to take these annotations forward with onward customisations and extensions, as illustrated in the example use case given above.</p> <p> </p> <p> </p>
Results of KROWN: Knowledge Graph Construction Benchmark
<p>In this Zenodo repository we present the results of using KROWN to benchmark popular RDF Graph Materialization systems such as RMLMapper, RMLStreamer, Morph-KGC, SDM-RDFizer, and Ontop (in materialization mode). </p> <h1>What is KROWN 👑?</h1> <p>KROWN 👑 is a benchmark for materialization systems to construct Knowledge Graphs from (semi-)heterogeneous data sources using declarative mappings such as<a href="http://w3id.org/rml/portal"> RML</a>.</p> <p>Many benchmarks already exist for virtualization systems e.g.<a href="https://github.com/oeg-upm/gtfs-bench"> GTFS-Madrid-Bench</a>,<a href="https://ontop-vkg.org/npd-benchmark/"> NPD</a>,<a href="http://wbsg.informatik.uni-mannheim.de/bizer/berlinsparqlbenchmark/"> BSBM</a> which focus on complex queries with a single declarative mapping. However, materialization systems are unaffected by complex queries since their input is the dataset and the mappings to generate a Knowledge Graph. Some specialized datasets exist to benchmark specific limitations of materialization systems such as duplicated or empty values in datasets e.g.<a href="https://doi.org/10.57702/4c9ivpgs"> GENOMICS</a>, but they do not cover all aspects of materialization systems. Therefore, it is hard to compare materialization systems among each other in general which is where KROWN 👑 comes in! </p> <h1>Results</h1> <p>The raw results are available as ZIP archives, the analysis of the results are available in the spreadsheet <em>results.ods</em>.</p> <h2>Evaluation setup</h2> <p>We generated several scenarios using <a href="https://github.com/kg-construct/KROWN/tree/main/data-generator">KROWN’s data generator</a> and executed them 5 times with <a href="https://github.com/kg-construct/KROWN/tree/main/execution-framework">KROWN’s execution framework</a>. All experiments were performed on Ubuntu 22.04 LTS machines (Linux 5.15.0, x86_64) with each Intel(R) Xeon(R) CPU E5-2650 v2 @ 2.60GHz, 48 GB RAM memory, and 2 GB swap memory. The output of each materialization system was set to N-Triples.</p> <h2>Materialization systems</h2> <p>We selected the most popular maintained materialization systems for constructing RDF graphs for performing our experiments with KROWN:</p> <ul> <li> <p>RMLMapper</p> </li> <li> <p>RMLStreamer</p> </li> <li> <p>Morph-KGC</p> </li> <li> <p>SDM-RDFizer</p> </li> <li> <p>OntopM (Ontop in materialization mode)</p> </li> </ul> <p><strong>Note</strong>: KROWN is flexible and allows adding any other materialization system, see <a href="https://github.com/kg-construct/KROWN/tree/main/execution-framework">KROWN’s execution framework</a> documentation for more information.</p> <h2>Scenarios</h2> <p>We consider the following scenarios:</p> <ul> <li> <p>Raw data: number of rows, columns and cell size</p> </li> <li> <p>Duplicates & empty values: percentage of the data containing duplicates or empty values</p> </li> <li> <p>Mappings: Triples Maps (TM), Predicate Object Maps (POM), Named Graph Maps (NG).</p> </li> <li> <p>Joins: relations (1-N, N-1, N-M), conditions, and duplicates during joins</p> </li> </ul> <p><strong>Note</strong>: KROWN is flexible and allows adding any other scenario, see <a href="https://github.com/kg-construct/KROWN/tree/main/data-generator">KROWN’s data generator documentation</a> for more information.</p> <p>In the table below we list all parameter values we used to configure our scenarios:</p> <div> <table> <tbody> <tr> <td> <p><strong>Scenario</strong></p> </td> <td> <p><strong>Parameter values</strong></p> </td> </tr> <tr> <td> <p>Raw data: rows</p> </td> <td> <p>10K, 100K, 1M, 10M</p> </td> </tr> <tr> <td> <p>Raw data: columns</p> </td> <td> <p>1, 10, 20, 30</p> </td> </tr> <tr> <td> <p>Raw data: cell size</p> </td> <td> <p>500, 1K, 5K, 10K </p> </td> </tr> <tr> <td> <p>Duplicates: percentage</p> </td> <td> <p>0%, 25%, 50%, 75%, 100%</p> </td> </tr> <tr> <td> <p>Empty values: percentage</p> </td> <td> <p>0%, 25%, 50%, 75%, 100%</p> </td> </tr> <tr> <td> <p>Mappings: TMs + 5POMs</p> </td> <td> <p>1, 10, 20, 30 TMs</p> </td> </tr> <tr> <td> <p>Mappings: 20TMs + POMs</p> </td> <td> <p>1, 3, 5, 10 POMs</p> </td> </tr> <tr> <td> <p>Mappings: NG in SM</p> </td> <td> <p>1, 5, 10, 15 NGs</p> </td> </tr> <tr> <td> <p>Mappings: NG in POM</p> </td> <td> <p>1, 5, 10, 15 NGs</p> </td> </tr> <tr> <td> <p>Mappings: NG in SM/POM</p> </td> <td> <p>1/1, 5/5, 10/10, 15/15 NGs</p> </td> </tr> <tr> <td> <p>Joins: 1-N relations</p> </td> <td> <p>1-1, 1-5, 1-10, 1-15</p> </td> </tr> <tr> <td> <p>Joins: N-1 relations</p> </td> <td> <p>1-1, 5-1, 10-1, 15-1</p> </td> </tr> <tr> <td> <p>Joins: N-M relations </p> </td> <td> <p>3-3, 3-5, 5-3, 10-5, 5-10</p> </td> </tr> <tr> <td> <p>Joins: join conditions</p> </td> <td> <p>1, 5, 10, 15</p> </td> </tr> <tr> <td> <p>Joins: join duplicates</p> </td> <td> <p>0, 5, 10, 15</p> </td> </tr> </tbody> </table> </div> <h1> </h1>
Tree Annotation Vocabulary (TAV) - Knowledge Graph and Annotated Dataset
<p>This dataset contains all the files used in developing the Tree-KG, the knowledge graph to capture the tree annotations in the works of Vladimir Nabokov. </p> <p>In the Annotated Dataset folder, 6 spreadsheets in excel (.xlsx) format are provided. They are numbered. Note that annotated data are all in English as the consulted works are the English translations of the literary works of Nabokov.</p> <p>(1) contains the tree annotations from the novels originally written in Russian by Vladimir Nabokov.</p> <p>(2) contains the tree annotations from the novels originally written in English by Vladimir Nabokov.</p> <p>(3) contains the tree annotations from the short stories originally written in Russian and English by Vladimir Nabokov.</p> <p>(4) is the knowledge base (KB) developed to link the annotated trees to Wikidata and DBPedia.</p> <p>(5) is the benchmarking results of some entity recognition tools. It includes the relevant passages from Nabokov's novels that were used in the experiments as well as the prompts used in getting the results.</p> <p>(6) represents the complete bibliographic details of the works of Vladimir Nabokov (https://thenabokovian.org/abbreviations).</p> <p>In the Ontology Versions folder, four ontology (TAV) files in turtle (.ttl) format are provided. They are all numbered and dated to represent their different versions. Some sample SPARQL queries are provided in a .txt file. The KG was developed on Protégé. </p> <p>(1) contains the essential schema for the TAV vocabulary.</p> <p>(2) contains the schema for TAV vocabulary with links to external vocabularies (Schema.Org; Open Annotation, etc.). </p> <p>(3) contains the Tree-KG in so far it reflects data from three novels (Mary; King, Queen, Knave; Glory).</p> <p>(4) contains the entire Tree-KG based on all the works mentioned in the excel sheets (20 books).</p> <p>(5) contains some sample SPARQL queries (.txt) file.</p>
Deep learning models predicting gene functions and pathways using public DRKG knowledge graph and graph neural network
<p>The attached dataset contains pretrained link prediction models, as described in our paper 'Morphological Map of Under- and Over-Expression of Genes in Human Cells'.</p>
Dataset variants used in "Task-Driven Knowledge Graph Filtering Improves Prioritizing Drugs for Repurposing"
<p>This file contains all datasets and variants thereof used in the linked paper. We do not take credit for constructing the datasets, which has been done by the respective original authors (<a href="https://github.com/hetio/hetionet">https://github.com/hetio/hetionet</a>, <a href="https://github.com/gnn4dr/DRKG">https://github.com/gnn4dr/DRKG</a>). For our work we produced modified versions (called "subset" in the file) by applying our metapath based filtering approach. For validation purposed we also constructed ablation versions where one specific type of entities is missing (i.e. "nogene", "noside", etc).</p>
Medieval manuscripts and their migrations: Using SPARQL to investigate the research potential of an aggregated Knowledge Graph
<p>This dataset contains the <strong>SPARQL queries</strong> presented and discussed in our article published in <em>Digital Medievalist</em> 2022 (as a PDF file), together with the <strong>results of those queries</strong> as CSV files. The query and step numbering follows that given in the article.</p> <p>The queries can be run against the SPARQL endpoint for the <strong>Mapping Manuscript Migrations</strong> project: <a href="https://ldf.fi/mmm/sparql">https://ldf.fi/mmm/sparql</a></p> <p>The full <strong>Mapping Manuscript Migrations dataset </strong>can also be downloaded from the Zenodo repository and installed in your own triple store: <a href="https://zenodo.org/record/4440464">https://zenodo.org/record/4440464</a></p> <p>When copying and pasting these SPARQL queries into a SPARQL client like <a href="https://yasgui.triply.cc/">YASGUI</a>, please check that the line numbering has been copied over correctly. Copying from a PDF file can sometimes break a single long line into multiple separate lines, which will cause a SPARQL validation error.</p> <p>The CSV files contain the results of the queries when run against the Mapping Manuscript Migrations SPARQL endpoint as of 17 December 2021. Please note that Query 2, Step 2, produces no results, so a CSV file has not been provided.</p> <p>The<strong> Mapping Manuscript Migrations portal </strong>can be found at <a href="https://mappingmanuscriptmigrations.org/en/">https://mappingmanuscriptmigrations.org/en/ </a></p> <p>SPARQL tutorials are included in the project's <strong>GitHub documentation</strong>: <a href="https://mapping-manuscript-migrations.github.io/">https://mapping-manuscript-migrations.github.io/</a></p>
Supporting Online Toxicity Detection with Knowledge Graphs: Data
<p>This data repository contains the output files from the analysis of the paper "Supporting Online Toxicity Detection with Knowledge Graphs" presented at the International Conference on Web and Social Media 2022 (ICWSM-2022).</p> <p> </p> <p>The data contains annotations of gender and sexual orientation entities provided by the Gender and Sexual Orientation Ontology (https://bioportal.bioontology.org/ontologies/GSSO).</p> <p>We analyse demographic group samples from the Civil Comments Identities dataset (https://www.tensorflow.org/datasets/catalog/civil_comments).</p>
SeaLiT Knowledge Graphs - Maritime History Data in RDF using a CIDOC-CRM extension (SeaLiT Ontology)
<p><strong>SeaLiT Knowledge Graphs</strong> is an RDF dataset of maritime history data that has been transcribed (and then transformed) from original archival sources in the context of the <a href="http://www.sealitproject.eu/">SeaLiT Project</a> (Seafaring Lives in Transition, Mediterranean Maritime Labour and Shipping, 1850s-1920s). The underlying data model is the <a href="https://zenodo.org/record/5964240">SeaLiT Ontology</a>, an extension of the ISO standard <strong>CIDOC-CRM</strong> (ISO 21127:2014) for the modelling and integration of maritime history information. </p> <p>The knowledge graphs integrate data of totally 16 different types of archival sources:</p> <ul> <li>Crew Lists <ul> <li>Crew and displacement list (Roll)</li> <li>Crew List (Ruoli di Equipaggio)</li> <li>General Spanish Crew List</li> </ul> </li> <li>Registers / Lists <ul> <li>Students Register</li> <li>Civil Register</li> <li>Register of Maritime Personnel</li> <li>Register of Maritime Workers (Matricole della gente di mare)</li> <li>Sailors Register (Libro de registro de marineros)</li> <li>Naval Ship Register List</li> <li>Seagoing Personnel</li> <li>Lists of ships</li> </ul> </li> <li>Censuses <ul> <li>Census La Ciotat</li> <li>First National all-Russian Census of the Russian Empire</li> </ul> </li> <li>Payrolls <ul> <li>Payrolls of private archives and libraries in Greece</li> <li>Payrolls of Russian Steam Navigation and Trading Company</li> </ul> </li> <li>Employment records <ul> <li>Shipyards of Messageries Maritimes, La Ciotat</li> </ul> </li> </ul> <p>More information about the archival sources are available through the <a href="https://sealitproject.eu/dictionary-of-source-types-list">SeaLiT website</a>. Data exploration applications over these sources are also publicly available (<a href="https://catalogues.sealitproject.eu/">SeaLiT Catalogues</a>, <a href="http://rs.sealitproject.eu/">SeaLiT ResearchSpace</a>). </p> <p>Data from these archival sources has been transcribed in tabular form and then curated by historians of SeaLiT using the <a href="https://www.ics.forth.gr/isl/fast-cat">FAST CAT</a> system. The transcripts (records), together with the curated vocabulary terms and entity instances (ships, persons, locations, organizations), are then transformed to RDF using the SeaLiT Ontology as the target (domain) model. To this end, the corresponding schema mappings between the original schemata and the ontology were defined using the <a href="https://github.com/isl/x3ml">X3ML</a> mapping definition language, that were subsequently used for delivering the RDF datasets. </p> <p>More information about the FAST CAT system and the data transcription, curation and transformation processes can be found in the following paper:</p> <blockquote> <p>P. Fafalios, K. Petrakis, G. Samaritakis, K. Doerr, A. Kritsotaki, Y. Tzitzikas, M. Doerr, "FAST CAT: Collaborative Data Entry and Curation for Semantic Interoperability in Digital Humanities", ACM Journal on Computing and Cultural Heritage, 2021. <a href="https://doi.org/10.1145/3461460">https://doi.org/10.1145/3461460</a> [<a href="http://users.ics.forth.gr/~fafalios/files/pubs/fafaliosJOCCH2021.pdf">pdf</a>, <a href="http://users.ics.forth.gr/~fafalios/files/bibs/fafaliosJOCCH2021.bib">bib</a>]</p> </blockquote> <p>The RDF dataset is provided as a set of TriG files per record per archival source. For each record, the dataset provides: i) one trig file for the record's data (<em>records.trig</em>), ii) one trig file for the record's (curated) vocabulary terms (<em>vocabularies.trig</em>), and iii) four trig files for the record's (curated) entity instances (<em>ships.trig, persons.trig, persons.trig, organizations.trig</em>).</p> <p>We also provide the RDFS files of the used ontologies (SeaLiT Ontology verson 1.0, CIDOC-CRM version 7.1.1). </p>
Publication and Maintenance of Relational Data in Enterprise Knowledge Graphs Created (Files used in the experiments)
<p>This dataset contains two files created for the experiments presented in the article: Publication and Maintenance of RDB2RDF Views Externally Materialized in Enterprise Knowledge Graphs.</p> <p><strong>mapR2RML_MusicBrainz_completo.txt</strong><strong>:</strong> We created the R2RML mapping for translating MBD data into the Music Ontology vocabulary, which is used for publishing the LMB view. The LMB view was materialized using the D2RQ tool. It took 67 minutes to materialize the view with approximately 41.1 GB of NTriples. We also provided SPARQL endpoint for querying LMB View.</p> <p><strong>TriggersAndProcedures.txt</strong>: We created the triggers, procedures, and class in java to implement the rules required to compute and publish the changesets.</p> <p><strong>relationalViewDefinition.pdf</strong>: This document gives details about the process of creating the relational views used in the experiments.</p>
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.