Accelerating specimen identification through Virtual Collections while promoting collaboration
Melanie de Leeuw · From Cabinets to Clouds - Bridging the gap between collection- and database management
Summaries and key moments are machine-produced. Check the recording for the speaker’s exact words.
The short versionVirtual reference collections with DOIs let experts identify and curate dispersed digital specimens remotely instead of shipping material or travelling.
What this was about
Melanie de Leeuw showed how the DiSSCo platform Disscover is turning from a passive lookup tool into an active workspace through Virtual Reference Collections (VRCs), funded by TETTRIs. VRCs pool digital specimens from multiple institutions into one virtual drawer. Three use cases were given: remote curation (e.g. Naturalis hawk moths identified only to family), expert-validated machine-learning training sets, and virtual field guides for rare taxa. Each VRC gets a DOI, aggregates across institutions, keeps two-way specimen–collection relationships, and supports expert annotation. Screenshots of the acceptance environment showed public browsing and ORCID-authenticated creation.
Why it matters. It turns existing digitised specimens into a collaborative identification resource and suggests a route to citable ML training data. It also raises the unresolved question of feeding annotations back to physical collections.
In the room
- Taxonomy is stuck in a cycle: physical distance limits experts, and digitised data stuck in silos doesn't help.
- VRCs pool digital specimens from institutions such as Naturalis, London, Paris and Oslo into one workspace.
- Use cases: curation at a distance, ML training datasets, and virtual field guides for rare or fragile taxa.
- Four pillars: a DOI per VRC, cross-institutional aggregation, dynamic entity relationships and active remote curation.
- Browsing is open without login; creating a VRC requires ORCID login and an appropriate role.
- Next steps: user-group involvement, VRC modification while preserving stability, and an annotation validation framework.
Notable moments
Transcript
Automatically generated captions can contain mistakes, especially in names and technical terms. Times are relative to the room recording.
Read the transcript ↓
Related talks.
Robots Rapidly Writing Records: Machine Annotations at Scale with DiSSCo
DiSSCo can now run adapted machine annotation services over whole datasets, pointing to an image-to-verified-data digitisation pipeline.
Experience of collaborative management of the World Auchenorrhyncha Database in TaxonWorks
TaxonWorks turned a personal taxonomic database into a collaborative, well-connected resource feeding Catalogue of Life, GBIF and iNaturalist.
Approaches to the Automated Tracking of Taxonomic Research Activities
TETTRIX aims to build a machine-run, regularly updated picture of who is doing taxonomy on what and where, and is asking the community which sources it is missing.