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.
915
datasets available to search
ShareScore release 0.9.0
Dataset results
915 results for “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>
A biodiversity dataset graph: Biological Associations in TaxonWorks hash://sha256/e4a47c067d6c125da60c9a1b92b5eecdea539cb8666cd3aed99db347ae5b8ed0 hash://md5/686007de79cc2a49ab23fd3debe56e3f
<p>The intended use of this archive is to facilitate (meta-)analysis of Biological Associations captured in TaxonWorks [1]. TaxonWorks is an integrated web-based workbench for taxonomists and biodiversity scientists. It allows you to capture, organize, and enrich your data; share it with collaborators; and package it for analysis and publication. </p> <p>This dataset provides versioned snapshots of the TaxonWorks network as tracked by Preston [2,3,4] during 2024-05-07 using:</p> <pre><code>preston track -u https://sfg.taxonworks.org</code></pre> <p>. In addition, this dataset provides a processed version of the biological associations using the "preston tw-stream" command as generated by the following bash script:</p> <pre><code>#!/bin/bash # # Generates GloBI interaction JSON Lines from provided provenance log as generated by preston tw-stream. # /usr/local/bin/preston cat hash://sha256/c1b081afa6ea0f60570c24cca85c4d9acd91eeefe36b9cacd1fe53b6893ea154\ | /usr/local/bin/preston tw-stream </code></pre> <p><br>The script itself was executed using:</p> <pre><code>cat transform.sh | preston bash </code></pre> <p>The execution of this transform.sh script (with content id hash://sha256/6dfe3c4ebf877bed73aebbe88c7d388bf894c569578ed7b28ca68e57a6afe43b), as well as their results, is captured within this datasets also. A rdf/quads formatted machine readable version of the workflow execution description can be found via:</p> <pre><code>preston cat hash://sha256/e4a47c067d6c125da60c9a1b92b5eecdea539cb8666cd3aed99db347ae5b8ed0 </code></pre> <p>And, the resulting JSON Lines file has content id (or signature) hash://sha256/4c2b8642251ced5985660d63c565efa6e5a9bf3d12b3b0c0d9ac577905f5e897 and is also included as interactions.json to facilitate access. </p> <p>The first json record can be generated using:</p> <pre><code>preston cat hash://sha256/4c2b8642251ced5985660d63c565efa6e5a9bf3d12b3b0c0d9ac577905f5e897\ | head -n1\ | jq . </code></pre> <p>or, provided that the interactions.json has content id starting with hash://sha256/4c2b86...</p> <pre><code>cat interactions.json\ | head -n1\ | jq . </code></pre> <p>This produces the following (formatted) json object:</p> <pre><code>{<br> "http://www.w3.org/ns/prov#wasDerivedFrom": "hash://sha256/fdbf13dc5f3d9c5afbc03db62699e2ce2724c499b7d91d8b0bf31e39409b153a",<br> "http://www.w3.org/1999/02/22-rdf-syntax-ns#type": "application/vnd.taxonworks+json",<br> "referenceId": "https://sfg.taxonworks.org/api/v1/sources/213218",<br> "interactionId": "https://sfg.taxonworks.org/api/v1/biological_associations/227664",<br> "taxonRootsResolved": 2,<br> "referenceResolved": true,<br> "referenceCitation": "@article{213218,\n author = {Monzen, Kota},\n journal = {Annual Report of the Gakugei Faculty of the Iwate University},\n pages = {24-38},\n title = {Revision of the Japanese gall wasps with the descriptions of new genus, subgenus, species and subspecies (II). Cynipidae (Cynipinae) Hymenoptera.},\n volume = {6},\n year = {1954}\n}\n",<br> "interactionTypeId": "gid://taxon-works/BiologicalRelationship/69",<br> "interactionTypeName": "gall",<br> "sourceTaxonName": "Neuroterus hakonensis",<br> "sourceTaxonId": "gid://taxon-works/TaxonName/1174121",<br> "sourceTaxonRank": "species",<br> "sourceTaxonAuthorship": "Ashmead, 1904",<br> "sourceTaxonPath": "Root | Cynipidae | Neuroterus | Neuroterus hakonensis",<br> "sourceTaxonPathIds": "gid://taxon-works/TaxonName/623170 | gid://taxon-works/TaxonName/1170060 | gid://taxon-works/TaxonName/1170097 | gid://taxon-works/TaxonName/1174121",<br> "sourceTaxonPathNames": "nomenclatural rank | family | genus | species",<br> "targetTaxonName": "Quercus",<br> "targetTaxonId": "gid://taxon-works/TaxonName/1173543",<br> "targetTaxonRank": "genus",<br> "targetTaxonAuthorship": "",<br> "targetTaxonPath": "Root | Fagaceae | Quercus",<br> "targetTaxonPathIds": "gid://taxon-works/TaxonName/623170 | gid://taxon-works/TaxonName/1173542 | gid://taxon-works/TaxonName/1173543",<br> "targetTaxonPathNames": "nomenclatural rank | family | genus"<br>}<br></code></pre> <p>In this example, a claim is made that, according to https://sfg.taxonworks.org/api/v1/sources/213218 [6] Neuroterus hakonensis (a gall wasp) has a primary host in the genus of Quercus (oak tree). </p> <p>In total, 237,068 such claims can be found in the generated resource with alias interactions.json and content id starting with hash://sha256/4c2b86... .</p> <p>In addition, the archive preston.tar.gz to allow for batch download. The archive contains three types of files: index files, provenance logs and data files. In addition, index files have been individually included in this dataset publication to facilitate remote access. Index files provide a way to links provenance files in time to establish a versioning mechanism. Provenance files describe how, when, what and where the TaxonWorks content was retrieved. For more information, please visit https://preston.guoda.bio or https://doi.org/10.5281/zenodo.1410543 . </p> <p>To retrieve and verify the downloaded TaxonWorks biodiversity dataset graph, download preston.tar.gz. Then, extract the archive into a "data" folder. Alternatively, you can use the preston[2] command-line tool to "clone" this dataset using:</p> <pre><code>java -jar preston.jar clone --remote https://zenodo.org/record/11151783/files </code></pre> <p>After that, verify the index of the archive by reproducing the following provenance log history:</p> <pre><code> java -jar preston.jar history --log tsv</code></pre> <p>to be:</p> <pre><code>hash://sha256/e4a47c067d6c125da60c9a1b92b5eecdea539cb8666cd3aed99db347ae5b8ed0 http://www.w3.org/ns/prov#wasDerivedFrom hash://sha256/c1b081afa6ea0f60570c24cca85c4d9acd91eeefe36b9cacd1fe53b6893ea154 </code><br><code>hash://sha256/c1b081afa6ea0f60570c24cca85c4d9acd91eeefe36b9cacd1fe53b6893ea154 http://www.w3.org/ns/prov#wasDerivedFrom hash://sha256/a4d651aac5220487835e6178511886e98b845b2d98cb7c5447fb2b042e0654d2hash://sha256/a4d651aac5220487835e6178511886e98b845b2d98cb7c5447fb2b042e0654d2 http://www.w3.org/ns/prov#wasDerivedFrom hash://sha256/ab7550368905e7c919e70a306efbb97719a1edbba2cfe4c4515f635ebc0be4bb hash://sha256/a4d651aac5220487835e6178511886e98b845b2d98cb7c5447fb2b042e0654d2 http://www.w3.org/ns/prov#wasDerivedFrom hash://sha256/ab7550368905e7c919e70a306efbb97719a1edbba2cfe4c4515f635ebc0be4bb</code><br><code>hash://sha256/ab7550368905e7c919e70a306efbb97719a1edbba2cfe4c4515f635ebc0be4bb http://www.w3.org/ns/prov#wasDerivedFrom hash://sha256/ff5e709305e593c87711e897b6341b94e775e2f312aa6d4ae5ed6120babd6f5e urn:uuid:0659a54f-b713-4f86-a917-5be166a14110 http://purl.org/pav/hasVersion hash://sha256/ff5e709305e593c87711e897b6341b94e775e2f312aa6d4ae5ed6120babd6f5e </code></pre> <p><br>To check the integrity of the extracted archive, confirm that each line produce by the command "preston verify" produces lines as shown below, with each line including "CONTENT_PRESENT_VALID_HASH". Depending on hardware capacity, this may take a while.</p> <pre><code>java -jar preston.jar verify</code></pre> <p>Note that a copy of the java program "preston", preston.jar, is included in this publication. The program runs on java 8+ virtual machine using "java -jar preston.jar", or in short "preston". </p> <p>Files in this data publication:</p> <p>--- start of file descriptions ---</p> <p>-- description of archive and its contents (a rendition of this file) --<br>README</p> <p>-- biological associations indexed from TaxonWorks expressed in a GloBI [5] compatible JSON Lines file --<br>interactions.json</p> <p>-- first 10 biological associations indexed from TaxonWorks expressed in a GloBI [5] compatible JSON Lines file --<br>interactions-10.json</p> <p>-- executable java jar containing preston [2,3,4] v0.8.5-SNAPSHOT. --<br>preston.jar</p> <p>-- preston archive containing TaxonWorks data files, associated provenance logs and a provenance index --<br>preston.tar.gz</p> <p>-- individual provenance index files --</p> <p>1fed32bf78298d7ecc3d9f36d106f1d7d7773a8b9a5e47af6632f36c1f82adb5<br>29306c5c144c3d7fd21be344d8b6b554b6f6efa3b8f8f5c0b27cdf0e88785652<br>2a5de79372318317a382ea9a2cef069780b852b01210ef59e06b640a3539cb5a<br>d31ff1ef1dea88c5952181a4f30e7ea7862873aa5f66430451275aa6d08d329e<br>deb84d69224af488da585186f88cafc58e978db5f9897de624cc9b02c0c83742<br>e9c34683f1e826f68f841f3419bd5ee9c0fa18be04713a6fd3364f226c7c5f2f<br>f98d36a9dc7bd833c93b3b61130865628f7bc2f7bb0920e95afcd16fba3dc6a8<br>ffb41d48979ceb964fbfbeb68cb60b584b759950087fdcc012521b866249bc39</p> <p>--- end of file descriptions ---</p> <p>This work is funded in part by grant NSF OAC 1839201, NSF DBI 1901932, NSF DBI 1901926, and NSF DBI 2102006 from the National Science Foundation.<br> </p>
Graphs of data for elderly individuals (aged 60 to 120 years) with syphilis in Brazil
<p>A set of data graphs containing informations on elderly people with syphilis in Brazil, aged between 60-120 years with syphilis in Brazil, aged between 60-120 years and contains spreadsheet results of trend analysis of acquired syphilis, by regions of Brazil, in the period 2010-2020, referring to the article entitled "<strong>ACQUIRED SYPHILIS IN OLDER PEOPLE IN BRAZIL FROM 2010-2020".<br><br><br></strong>The dataset used to plot the graphs can be found at: <a href="https://doi.org/10.5281/zenodo.10086131">https://doi.org/10.5281/zenodo.10086131</a></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>
CauseNet: Towards a Causality Graph Extracted from the Web
<p>Causal knowledge is seen as one of the key ingredients to advance artificial intelligence. Yet, few knowledge bases comprise causal knowledge to date, possibly due to significant efforts required for validation. Notwithstanding this challenge, we compile CauseNet, a large-scale knowledge base of <em>claimed </em>causal relations between causal concepts. By extraction from different semi- and unstructured web sources, we collect more than 11 million causal relations with an estimated extraction precision of 83% and construct the first large-scale and open-domain causality graph. We analyze the graph to gain insights about causal beliefs expressed on the web and we demonstrate its benefits in basic causal question answering. Future work may use the graph for causal reasoning, computational argumentation, multi-hop question answering, and more.</p> <p>When using the data, please make sure to refer to it as follows:</p> <pre><code>@inproceedings{heindorf2020causenet, author = {Stefan Heindorf and Yan Scholten and Henning Wachsmuth and Axel-Cyrille Ngonga Ngomo and Martin Potthast}, title = {CauseNet: Towards a Causality Graph Extracted from the Web}, booktitle = {{CIKM}}, pages = {3023--3030}, publisher = {{ACM}}, year = {2020} }</code></pre>
Random Simple Undirected Weighted Graphs
<div> <div> <div> <div> <div> <div dir="auto"> <div> <div> <h3>Dataset Description</h3> <p>We introduce a dataset consisting of <em><strong>over 60 flow matrices</strong></em> representing <strong>simple</strong>, <strong>undirected</strong>, <strong>weighted</strong> <strong>graphs</strong>. This dataset is designed to support empirical studies in graph algorithms, clustering, and network analysis.</p> <p>Each graph is characterized by</p> <ul> <li> <p><strong>Order (|V|):</strong><br><span><span>{20, 50, 100, 300, 500, 700, 800, 900, 1000, 2000, 3000}</span></span></p> </li> <li> <p><strong>Density:</strong><br>For each graph order, 10 instances are generated with edge densities from the following set:<br><span><span>{0.05, 0.1, 0.25, 0.5, 0.75, 0.9, 1.0}</span></span></p> </li> <li> <p><strong>Clustered Structure:</strong><br>Nodes are partitioned into predefined clusters, where intra-cluster edges have significantly higher weights, while all inter-cluster edges are uniformly weighted (weight = 1). This structure simulates modular graphs commonly encountered in real-world networks.</p> </li> <li> <p><strong>Number of Clusters:</strong><br>The number of clusters varies according to graph order:<br><span><span>{5, 5, 10, 30, 50, 70, 80, 90, 100, 400, 500}</span></span></p> </li> </ul> <p>This dataset enables systematic testing of algorithms under varying structural conditions, including <strong>scale</strong>, <strong>sparsity</strong>, and <strong>community strength.</strong></p> </div> </div> </div> </div> </div> </div> </div> </div> <div> <div> <div> <div> <div> <div dir="auto"> <div> <div> <p> </p> </div> </div> </div> </div> </div> </div> </div> </div>
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>
Lehigh University Benchmark (LUBM): Evolving Graph Simulation
<p>The Lehigh University Benchmark (LUBM) generates benchmark datasets containing people working at universities [1]. We use the Data Generator v1.7 to generate 10 versions of a graph containing 100 universities [2].<br> Thus, all versions are of similar size, but we emulate modifications by generating different vertex identifiers, i.e., each version is considered a timestamped graph. Each graph contains about 2.1 M vertices and 13 M edges.<br> Over all versions, the mean degree is 6.7 (+- 0.1), the mean in-degree is 6.8 (+- 0.1), and the mean out-degree is 5.1 (+- 0.1).</p> <p>1. <a href="https://dblp.uni-trier.de/pid/80/5390.html">Yuanbo Guo</a>, <a href="https://dblp.uni-trier.de/pid/48/6834.html">Zhengxiang Pan</a>, <a href="https://dblp.uni-trier.de/pid/94/1154.html">Jeff Heflin</a>: LUBM: A benchmark for OWL knowledge base systems. <a href="https://dblp.uni-trier.de/db/journals/ws/ws3.html#GuoPH05">J. Web Semant. 3(2-3)</a>: 158-182 (2005)</p> <p>2. <a href="https://dblp.uni-trier.de/pid/222/6353.html">Till Blume</a>, <a href="https://dblp.uni-trier.de/pid/r/DavidRicherby.html">David Richerby</a>, <a href="https://dblp.uni-trier.de/pid/06/2380.html">Ansgar Scherp</a>: Incremental and Parallel Computation of Structural Graph Summaries for Evolving Graphs. <a href="https://dblp.uni-trier.de/db/conf/cikm/cikm2020.html#BlumeRS20">CIKM 2020</a>: 75-84</p>
Berlin Sparql Benchmark (BSBM): Evolving Graph Simulation
<p>The Berlin SPARQL Benchmark (BSBM) is a suite of benchmarks built around an e-commerce use case [1]. We generated 21 versions of the dataset with different scale factors. The first dataset, with a scale factor of 100, contains about 7,000 vertices and 75,000 edges. We generated versions with scale factors between 2,000 and 40,000 in steps of 2,000. The largest dataset contains about 1.3 M vertices and 13 M edges. For our experiments in [2], we first use the different versions ordered from smallest to largest (version 0 to 20) to simulate a growing graph database. Subsequently, we reverse the order to emulate a shrinking graph database. Over all versions, the mean degree is 8.1 (+- 0.5), the mean in-degree is 4.6 (+- 0.3), and the mean out-degree is 9.8 (+- 0.2).</p> <p>1. <a href="https://dblp.uni-trier.de/pid/b/ChristianBizer.html">Christian Bizer</a>, <a href="https://dblp.uni-trier.de/pid/47/7466.html">Andreas Schultz</a>: The Berlin SPARQL Benchmark. <a href="https://dblp.uni-trier.de/db/journals/ijswis/ijswis5.html#BizerS09">Int. J. Semantic Web Inf. Syst. 5(2)</a>: 1-24 (2009)</p> <p>2. <a href="https://dblp.uni-trier.de/pid/222/6353.html">Till Blume</a>, <a href="https://dblp.uni-trier.de/pid/r/DavidRicherby.html">David Richerby</a>, <a href="https://dblp.uni-trier.de/pid/06/2380.html">Ansgar Scherp</a>: Incremental and Parallel Computation of Structural Graph Summaries for Evolving Graphs. <a href="https://dblp.uni-trier.de/db/conf/cikm/cikm2020.html#BlumeRS20">CIKM 2020</a>: 75-84</p>
Graph Data: Hydrological impact of widespread afforestation in Great Britain using a large ensemble of modelled scenarios
<p>Data used for creating the figures in the paper: Hydrological impact of widespread afforestation in Great Britain using a large ensemble of modelled scenarios.</p> <p>It contains the flow exceedances (as mm day<sup>-1</sup>), flow duration slope, median elasticity and runoff ratio for the different afforestation scenarios. Also included is the information on the changes of broadleaf afforestation. </p> <p>If you have any questions, please email marcus.buechel@ouce.ox.ac.uk.</p>
Expert judgements for evaluating deduplication of OpenAIRE Research Graph
<p>Expert judgments used to evaluate the deduplication algorithm used for constructing the OpenAIRE Research Graph.</p> <p>Each expert assigned each group with one of the following predetermined classes (that also indicate whether contained entities are equivalent or not):</p> <ul> <li>AMBIGUOUS: At least one DOIs is invalid (no metadata are available) (N/A)</li> <li>DELETED-DUPLICATES: DOIs once pointing to the same research object, currently deleted. (TRUE)</li> <li>MULTI-PUBLISHED: Article published in more than one locations (full or abstract) (TRUE)</li> <li>VERSIONS: Multiple versions of the same research object (e.g. pre-prints, post-prints etc). (TRUE)</li> <li>ERRONEOUS: Unrelated set of objects. (FALSE)</li> <li>PAPER-EXTENSIONS: Extended version of a conference paper in a journal. (FALSE)</li> <li>PART-OF-A-GROUP: Multiple parts of the same research object (e.g. multi-part publication, photos of the same collection etc). (FALSE)</li> <li>SUPPLEMENTARY: A publication and its supplementary material (including errata). (FALSE)</li> </ul> <p>The following are provided:</p> <ul> <li>OpenAIRE identifier</li> <li>Judgement</li> <li>Count of numbers in group</li> <li>DOIs in group</li> </ul>
Challenges in Migrating Imperative Deep Learning Programs to Graph Execution: An Empirical Study
<p>Efficiency is essential to support responsiveness w.r.t. ever-growing datasets, especially for Deep Learning (DL) systems. DL frameworks have traditionally embraced deferred execution-style DL code that supports symbolic, graph-based Deep Neural Network (DNN) computation. While scalable, such development tends to produce DL code that is error-prone, non-intuitive, and difficult to debug. Consequently, more natural, less error-prone imperative DL frameworks encouraging eager execution have emerged but at the expense of run-time performance. While hybrid approaches aim for the "best of both worlds," the challenges in applying them in the real world are largely unknown. We conduct a data-driven analysis of challenges—and resultant bugs—involved in writing reliable yet performant imperative DL code by studying 250 open-source projects, consisting of 19.7 MLOC, along with 470 and 446 manually examined code patches and bug reports, respectively. The results indicate that hybridization: (i) is prone to API misuse, (ii) can result in performance degradation—the opposite of its intention, and (iii) has limited application due to execution mode incompatibility. We put forth several recommendations, best practices, and anti-patterns for effectively hybridizing imperative DL code, potentially benefiting DL practitioners, API designers, tool developers, and educators.</p>
Filtered AMR Graphs Bank
<p>The filtered AMR graphs bank is the output of the <a href="http://github.com/polifonia-project/Polifonia-Knowledge-Extractor">Polifonia Knowledge Extractor</a> pipeline.</p> <p>The Polifonia Mini Textual Corpus is pre-processed through coreference resolution and minimal rule-based post-OCR correction, then it is given as input to <a href="https://github.com/SapienzaNLP/spring">SPRING</a> to obtain AMR graphs.</p> <p>The filtered AMR graphs bank encompasses the sentence-AMR graph pairs extrapolated from our initial AMR graphs bank corresponding to AMR2Text-generated sentences associated with a positive (>0) <a href="https://github.com/google-research/bleurt">BLEURT</a> score.</p> <p>Full description at: <a href="https://github.com/polifonia-project/Polifonia-Knowledge-Extractor">https://github.com/polifonia-project/Polifonia-Knowledge-Extractor</a></p>
Wikipedia time-series graph
<p>Wikipedia temporal graph.</p> <p>The dataset is based on two Wikipedia SQL dumps: (1) English language articles and (2) user visit counts per page per hour (aka pagecounts). The original datasets are publicly available on the Wikimedia website.</p> <p>Static graph structure is extracted from English language Wikipedia articles. Redirects are removed. Before building the Wikipedia graph we introduce thresholds on the minimum number of visits per hour and maximum in-degree. We remove the pages that have less than 500 visits per hour at least once during the specified period. Besides, we remove the nodes (pages) with in-degree higher than 8 000 to build a more meaningful initial graph. After cleaning, the graph contains 116 016 nodes (out of total 4 856 639 pages), 6 573 475 edges. The graph can be imported in two ways: (1) using edges.csv and vertices.csv or (2) using enwiki-20150403-graph.gt file that can be opened with open source Python library Graph-Tool.</p> <p>Time-series data contains users' visit counts from 02:00, 23 September 2014 until 23:00, 30 April 2015. The total number of hours is 5278. The data is stored in two formats: CSV and H5. CSV file contains data in the following format [page_id :: count_views :: layer], where layer represents an hour. In H5 file, each layer corresponds to an hour as well.</p>
HEroBM: a deep equivariant graph neural network for high-fidelity backmapping from coarse-grained to all-atom structures
<p><span>Molecular simulations play a pivotal role in chemistry, biology, and material sciences, enabling the</span><br><span>study of complex dynamic properties within systems. Coarse-grained (CG) techniques have emerged</span><br><span>as indispensable tools in this domain, facilitating the sampling of large-scale systems and extending</span><br><span>simulation timescales by simplifying system representation. However, CG approaches involve a trade-</span><br><span>off: they sacrifice atomistic details that may be crucial for understanding the underlying processes.</span><br><span>To address this challenge, a recommended strategy is to identify key CG conformations and employ</span><br><span>backmapping methods to retrieve atomistic coordinates. Currently, rule-based methods often yield</span><br><span>suboptimal geometries and rely on energy relaxation, resulting in less-than-optimal outcomes. In</span><br><span>contrast, machine learning techniques offer higher accuracy but may lack transferability between</span><br><span>systems or be tied to specific CG mappings. In this study, we present HEroBM, a dynamic and scalable</span><br><span>method that utilizes deep equivariant graph neural networks and a hierarchical approach to achieve</span><br><span>high-resolution backmapping. HEroBM is capable of handling any type of CG mapping, providing a</span><br><span>versatile and efficient protocol for reconstructing atomistic structures with high accuracy. Grounded</span><br><span>in local principles, HEroBM spans the entire chemical space and can be applied across systems of</span><br><span>varying composition and sizes. We demonstrate the versatility of our framework through a range of</span><br><span>biological systems, including a complex real-case scenario. Here, our end-to-end backmapping approach</span><br><span>accurately generates atomistic coordinates for a G protein-coupled receptor bound to an organic small</span><br><span>molecule within a cholesterol/phospholipid bilayer. The high-fidelity HEroBM backmapping enables</span><br><span>researchers to effortlessly transition between CG and all-atom simulations, opening unprecedented</span><br><span>avenues for molecular investigations.</span></p>
SCG Dataset from Graph Neural Networks in Supply Chain Analytics and Optimization: Concepts, Perspectives, Dataset and Benchmarks
<p><strong>Abstract:</strong> Graph Neural Networks (GNNs) have recently gained traction in transportation, bioinformatics, language and image processing, but research on their application to supply chain management remains limited. Supply chains are inherently graph-like, making them ideal for GNN methodologies, which can optimize and solve complex problems. The barriers include a lack of proper conceptual foundations, familiarity with graph applications in SCM, and real-world benchmark datasets for GNN-based supply chain research. To address this, we discuss and connect supply chains with graph structures for effective GNN application, providing detailed formulations, examples, mathematical definitions, and task guidelines. Additionally, we present a multi-perspective real-world benchmark dataset from a leading FMCG company in Bangladesh, focusing on supply chain planning. We discuss various supply chain tasks using GNNs and benchmark several state-of-the-art models on homogeneous and heterogeneous graphs across six supply chain analytics tasks. Our analysis shows that GNN-based models consistently outperform statistical ML and other deep learning models by around 10-30% in regression, 10-30% in classification and detection tasks, and 15-40% in anomaly detection tasks on designated metrics. With this work, we lay the groundwork for solving supply chain problems using GNNs, supported by conceptual discussions, methodological insights, and a comprehensive dataset.</p>
Speculative Automated Refactoring of Imperative Deep Learning Programs to Graph Execution
<p>Efficiency is essential to support ever-growing datasets, especially for Deep Learning (DL) systems. DL frameworks have traditionally embraced deferred execution-style DL code---supporting symbolic, graph-based Deep Neural Network (DNN) computation. While scalable, such development is error-prone, non-intuitive, and difficult to debug. Consequently, more natural, imperative DL frameworks encouraging eager execution have emerged but at the expense of run-time performance. Though hybrid approaches aim for the "best of both worlds," using them effectively requires subtle considerations. Our key insight is that, while DL programs typically execute sequentially, hybridizing imperative DL code resembles parallelizing sequential code in traditional systems. Inspired by this, we present an automated refactoring approach that assists developers in determining which otherwise eagerly-executed imperative DL functions could be effectively and efficiently executed as graphs. The approach features novel static imperative tensor and side-effect analyses for Python. Due to its inherent dynamism, analyzing Python may be unsound; however, the conservative approach leverages a speculative (keyword-based) analysis for resolving difficult cases that informs developers of any assumptions made. The approach is: (i) implemented as a plug-in to the PyDev Eclipse IDE that integrates the WALA Ariadne analysis framework and (ii) evaluated on nineteen DL projects consisting of 132 KLOC. The results show that 326 of 766 candidate functions (42.56%) were refactorable, and an average relative speedup of 2.16x on performance tests was observed with negligible differences in model accuracy. The results indicate that the approach is useful in optimizing imperative DL code to its full potential.</p>
OpenAIRE Graph: Dataset for research communities and initiatives
<p>This dataset contains metadata records of the OpenAIRE Graph relevant for the research communities and initiatives collaborating with OpenAIRE and with a public Community Gateway on <a href="https://connect.openaire.eu">OpenAIRE CONNECT</a> as of July 2025.</p> <p>Each file is a tar archive containing gzip files with one json per line. Each json is compliant to the schema available at <a href="https://doi.org/10.5281/zenodo.14891476" target="_blank" rel="noopener">https://doi.org/10.5281/zenodo.14891476</a></p> <table style="width: 100%; height: 822.939px;"> <tbody> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;"><strong>Research community name</strong></td> <td style="width: 38.9414%; height: 19.5938px;"><strong>File name</strong></td> <td style="width: 23.1067%; height: 19.5938px;"><strong>URL to the gateway</strong></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">Argo France</td> <td style="width: 38.9414%; height: 19.5938px;">argo-france.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://argo-france.openaire.eu/">https://argo-france.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">Aurora University Alliance</td> <td style="width: 38.9414%; height: 19.5938px;">aurora.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://aurora.openaire.eu">https://aurora.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">Transport Research (EC projects BE OPEN and SciLake)</td> <td style="width: 38.9414%; height: 19.5938px;">beopen.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://beopen.openaire.eu">https://beopen.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">CIVICA Alliance</td> <td style="width: 38.9414%; height: 19.5938px;">civica.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://civica.openaire.eu" target="_blank" rel="noopener">https://civica.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">COVID 19</td> <td style="width: 38.9414%; height: 19.5938px;">covid-19.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://covid-19.openaire.eu">https://covid-19.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">DARIAH EU</td> <td style="width: 38.9414%; height: 19.5938px;">dariah.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://dariah.openaire.eu">https://dariah.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">Digital Humanities and Cultural Heritage</td> <td style="width: 38.9414%; height: 19.5938px;">dh-ch.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://dh-ch.openaire.eu">https://dh-ch.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">Virtual Twins in Health (EC project EDITH)</td> <td style="width: 38.9414%; height: 19.5938px;">dth.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://dth.openaire.eu">https://dth.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">European Digital Innovation Hubs Network ADRIA</td> <td style="width: 38.9414%; height: 19.5938px;">edih-adria.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://edih-adria.openaire.eu">https://edih-adria.openaire.eu</a></td> </tr> <tr style="height: 39.1875px;"> <td style="width: 33.9453%; height: 39.1875px;">European Geothermal Research and Innovation Search Engine </td> <td style="width: 38.9414%; height: 39.1875px;">egrise.tar</td> <td style="width: 23.1067%; height: 39.1875px;"><a href="https://egrise.openaire.eu">https://egrise.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">ELIXIR Greece</td> <td style="width: 38.9414%; height: 19.5938px;">elixir-gr.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://elixir-gr.openaire.eu">https://elixir-gr.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">Energy Planning (EC project SciLake)</td> <td style="width: 38.9414%; height: 19.5938px;">energy-planning_1.tar, energy_planning_2.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://energy-planning.openaire.eu/">https://energy-planning.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">Energy Research (EC project Enermaps)</td> <td style="width: 38.9414%; height: 19.5938px;">enermaps.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://enermaps.openaire.eu">https://enermaps.openaire.eu</a></td> </tr> <tr style="height: 39.1875px;"> <td style="width: 33.9453%; height: 39.1875px;">European University for Smart Urban Coastal Sustainability</td> <td style="width: 38.9414%; height: 39.1875px;">eu-conexus.tar</td> <td style="width: 23.1067%; height: 39.1875px;"><a href="https://eu-conexus.openaire.eu/">https://eu-conexus.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">European University of Technology+</td> <td style="width: 38.9414%; height: 19.5938px;">eut.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://eut.openaire.eu/">https://eut.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">EUTOPIA Alliance</td> <td style="width: 38.9414%; height: 19.5938px;">eutopia.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://eutopia.openaire.eu/">https://eutopia.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">FORTHEM Alliance</td> <td style="width: 38.9414%; height: 19.5938px;">forthem.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://forthem.openaire.eu">https://forthem.openaire.eu</a></td> </tr> <tr> <td style="width: 33.9453%;">[NEW] GoTriple </td> <td style="width: 38.9414%;">gotriple_1.tar, gotriple_2.tar</td> <td style="width: 23.1067%;"><a href="https://gotriple.openaire.eu/">https://gotriple.openaire.eu/</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">Heritage Science (EC project IPERION HS)</td> <td style="width: 38.9414%; height: 19.5938px;">heritage-science.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://heritage-science.openaire.eu/">https://heritage-science.openaire.eu</a></td> </tr> <tr style="height: 39.1875px;"> <td style="width: 33.9453%; height: 39.1875px;">Institut national de recherche en informatique et en automatique (EC project GraspOS)</td> <td style="width: 38.9414%; height: 39.1875px;">inria.tar</td> <td style="width: 23.1067%; height: 39.1875px;"><a href="https://inria.openaire.eu" target="_blank" rel="noopener">https://inria.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">IPERION HS</td> <td style="width: 38.9414%; height: 19.5938px;">iperionhs.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://iperionhs.openaire.eu">https://iperionhs.openaire.eu</a></td> </tr> <tr style="height: 39.1875px;"> <td style="width: 33.9453%; height: 39.1875px;">Knowmad Institut</td> <td style="width: 38.9414%; height: 39.1875px;">knowmad_1.tar, knowmad_2.tar, knowmad_3.tar, knowmad_4.tar, knowmad_5.tar, knowmad_6.tar, knowmad_7.tar, knowmad_8.tar, knowmad_9.tar</td> <td style="width: 23.1067%; height: 39.1875px;"><a href="https://knowmad.openaire.eu/">https://knowmad.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">LifeWathc ERIC</td> <td style="width: 38.9414%; height: 19.5938px;">lifewatch-eric.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://lifewatch-eric.openaire.eu/">https://lifewatch-eric.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">European Marine Science</td> <td style="width: 38.9414%; height: 19.5938px;">mes.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://mes.openaire.eu">https://mes.openaire.eu</a></td> </tr> <tr style="height: 39.1875px;"> <td style="width: 33.9453%; height: 39.1875px;">Atmospheric Research Community (EC project NEANIAS)</td> <td style="width: 38.9414%; height: 39.1875px;">neanias-atmospheric.tar</td> <td style="width: 23.1067%; height: 39.1875px;"><a href="https://neanias-atmospheric.openaire.eu/">https://neanias-atmospheric.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">Space Research Community (EC project NEANIAS)</td> <td style="width: 38.9414%; height: 19.5938px;">neanias-space.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://neanias-space.openaire.eu/">https://neanias-space.openaire.eu</a></td> </tr> <tr style="height: 39.1875px;"> <td style="width: 33.9453%; height: 39.1875px;">Underwater Research Community (EC project NEANIAS)</td> <td style="width: 38.9414%; height: 39.1875px;">neanias-underwater.tar</td> <td style="width: 23.1067%; height: 39.1875px;"><a href="https://neanias-underwater.openaire.eu/">https://neanias-underwater.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">Netherlands Research Portal</td> <td style="width: 38.9414%; height: 19.5938px;">netherlands_1.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://netherlands.openaire.eu/">https://netherlands.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">[NEW] Neuroscience<br>(EC project SciLake - former Neuroinformatics community is now a subcommunity of Neuroscience)</td> <td style="width: 38.9414%; height: 19.5938px;">neuroscience_1.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://neuroscience.openaire.eu">https://neuroscience.openaire.eu</a></td> </tr> <tr style="height: 39.1875px;"> <td style="width: 33.9453%; height: 39.1875px;">North American Studies</td> <td style="width: 38.9414%; height: 39.1875px;">north-american-studies.tar</td> <td style="width: 23.1067%; height: 39.1875px;"><a href="https://north-american-studies.openaire.eu">https://north-american-studies.openaire.eu</a></td> </tr> <tr style="height: 39.1875px;"> <td style="width: 33.9453%; height: 39.1875px;">Rural Digital Europe (EC project DESIRA)</td> <td style="width: 38.9414%; height: 39.1875px;">rural-digital-europe.tar</td> <td style="width: 23.1067%; height: 39.1875px;"><a href="https://rural-digital-europe.openaire.eu/">https://rural-digital-europe.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">Sustainable Development Solutions Network - Greece </td> <td style="width: 38.9414%; height: 19.5938px;">sdsn-gr.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://sdsn-gr.openaire.eu/">https://sdsn-gr.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">Technological University Network</td> <td style="width: 38.9414%; height: 19.5938px;">tunet.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://tunet.openaire.eu">https://tunet.openaire.eu</a></td> </tr> <tr> <td style="width: 33.9453%;">[NEW] UNITE! University Alliance</td> <td style="width: 38.9414%;">unite.tar</td> <td style="width: 23.1067%;"><a href="https://unite.openaire.eu">https://unite.openaire.eu</a></td> </tr> <tr style="height: 19.5938px;"> <td style="width: 33.9453%; height: 19.5938px;">University of the Arctic (UArctic) </td> <td style="width: 38.9414%; height: 19.5938px;">uarctic_1.tar, uarctic_2.tar</td> <td style="width: 23.1067%; height: 19.5938px;"><a href="https://uarctic.openaire.eu/">https://uarctic.openaire.eu</a></td> </tr> </tbody> </table>
Advection datasets from "Multi-scale rotation-equivariant graph neural networks for unsteady Eulerian fluid dynamics"
<p>Advection datasets from the paper:<br> Multi-scale rotation-equivariant graph neural networks for unsteady Eulerian fluid dynamics (https://doi.org/10.1063/5.0097679)</p> <p>The datasets are:<br> - AdvBox<br> - AdvInBox<br> - AdvTaylor<br> - AdvCircle<br> - AdvCircleAng<br> - AdvSquare<br> - AdvEllipseH<br> - AdvEllipseV<br> - AdvSpline<br> - AdvSquareAndCircle<br> - Adv3Circles</p> <p>Check the "README.txt" file for information on how the simulations are organised. The features of each dataset and how they were generated are explained in the journal publication.</p> <p> </p> <p>To cite these datasets, use the following reference:</p> <p>Mario Lino, Stathi Fotiadis, Anil A. Bharath, and Chris Cantwell. "Multi-scale rotation-equivariant graph neural networks for unsteady Eulerian fluid dynamics". Physics of Fluids, 34 (2022).</p> <pre><code>@article{lino2022multi, author = {Lino, Mario and Fotiadis, Stathi and Bharath, Anil A. and Cantwell, Chris}, title = {{Multi-scale rotation-equivariant graph neural networks for unsteady Eulerian fluid dynamics}}, journal = {Physics of Fluids}, volume = {34}, year = {2022}, url = {https://doi.org/10.1063/5.0097679}, }</code></pre> <p><br> </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.