Using Wikidata and Metaphactory to Underpin an Integrated Flora of Canada
Notice bibliographique
Résumé
We are using Wikidata and Metaphactory to build an Integrated Flora of Canada (IFC). IFC will be integrated in two senses: First, it will draw on multiple existing flora (e.g. Flora of North America, Flora of Manitoba, etc.) for content. Second, it will be a portal to related resources such as annotations, specimens, literature, and sequence data. Background We had success using Semantic Media Wiki (SMW) as the platform for an on-line representation of the Flora of North America (FNA). We used Charaparser (Cui 2012) to extract plant structures (e.g. “stem”), characters (e.g. “external texture”), and character values (e.g. “glabrous”) from the semi-structured FNA treatments. We then loaded this data into SMW, which allows us to query for taxa based on their character traits, and enables a broad range of exploratory analysis, both for purposes of hypothesis generation, and also to provide support for or against specific scientific hypotheses. Migrating to Wikidata/Wikibase We decided to explore a migration from SMW to Wikibase for three main reasons: simplified workflow; triple level provenance; and sustainability. Simplified workflow: Our workflow for our FNA-based portal includes Natural Language Processing (NLP) of coarse-grained XML to get the fine-grained XML, transforming this XML for input into SMW, and a custom SMW skin for displaying the data. We consider the coarse-grained XML to be canonical. When it changes (because we find an error, or we improve our NLP), we have to re-run the transformation, and re-load the data, which is time-consuming. Ideally, our presentation would be based on API calls to the data itself, eliminating the need to transform and re-load after every change. Provenance: Wikidata's provenance model supports having multiple, conflicting assertions for the same character trait, which is something that inevitably happens when floristic data is integrated. Sustainability: Wikidata has strong support from the Wikimedia Foundation, while SMW is increasingly seen as a legacy system. Wikibase vs. Wikidata Wikidata, however, is not a suitable home for the Integrated Flora of Canada. It is built upon a relatively small number of community curated properties, while we have ~4500 properties for the Asteraceae family alone. The model we want to pursue is to use Wikidata for a small group of core properties (e.g. accepted name, parent taxon, etc.), and to use our own instance of Wikibase for the much larger number of specialized morphological properties (e.g. adaxial leaf colour, leaf external texture, etc.) Essentially, we will be running our own Wikidata, over which we would exercise full control. Miller (2018) decribes deploying this curation model in another domain. Metaphactory Metaphactory is a suite of middleware and front-end interfaces for authoring, managing, and querying knowledge graphs, including mechanisms for faceted search and geospatial visualizations. It is also the software (together with Blazegraph) behind the Wikidata Query Service. Metaphactory provides us with a SPARQL endpoint; a templating mechanism that allows each taxonomic treatment to be rendered via a collection of SPARQL queries; reasoning capabilities (via an underlying graph database) that permit the organization of over 42,000 morphological properties; and a variety of search and discovery tools. There are a number of ways in which Wikidata and Metaphactory can work together, and we are still exploring questions such as: Will provenance be managed via named graphs, or via the Wikidata snak model?; How will data flow between the two platforms? Etc. We will report on our findings to date, and invite collaboration with related Wikimedia-based projects.
Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.
Comment cette classification a été obtenuedéplier
Prédiction machine sur la base complète
Imitation des enseignantsNi prévalence calibrée, ni vérité terrain. Validation humaine à venir. Le volet Gemma est une étiquette directe du modèle pour chaque travail de la base, lue sur la notice réduite au titre. Le volet Codex est un classifieur appris des 10 348 étiquettes directes de Codex et calibré sur les taux pondérés de l'échantillon; les champs sans appui suffisant ne portent aucun appel Codex. Le mode candidate est l'union des deux volets; le consensus est leur intersection. Ces sorties portent le statut machine_predicted_unvalidated et ne sont pas des étiquettes humaines.
Scores du classifieur distillé par catégorie (deux têtes)
| Catégorie | Codex | Gemma |
|---|---|---|
| Métarecherche | 0,003 | 0,008 |
| Méta-épidémiologie (sens strict) | 0,001 | 0,001 |
| Méta-épidémiologie (sens large) | 0,001 | 0,001 |
| Bibliométrie | 0,010 | 0,009 |
| Études des sciences et des technologies | 0,008 | 0,004 |
| Communication savante | 0,009 | 0,005 |
| Science ouverte | 0,003 | 0,005 |
| Intégrité de la recherche | 0,001 | 0,002 |
| Charge utile insuffisante (le modèle a refusé de juger) | 0,017 | 0,003 |
Scores machine (provisoires)
Les deux têtes enseignantes du modèle étudiant, lues sur ce travail. Un score ordonne la base pour la relecture; il n'affirme jamais une catégorie, et le statut de validation accompagne chaque rangée tel quel.
Scores de référence d'un modèle non mature (critères de maturité non atteints, 7 itérations). Un score ordonne; il n'affirme jamais une catégorie.
score_only:v0-immature-baseline · tel quel depuis la passe de notation : score_only signifie que le nombre peut ordonner les travaux, et qu'aucune étiquette de catégorie n'en découleClassification
machine, non validéePrédiction automatique; un appel candidat d’une seule source (Gemma direct ou Codex distillé), pas un consensus.
Le détail, modèle par modèle et score par score, se trouve en fin de page sous « Comment cette classification a été obtenue ».