
(Updated) 1st Feb. 2025
BRIEF REPORT ON THE DISSCO MACHINE ANNOTATION SERVICES HACKATHON.
From 24 to 26 March 2025, fifteen Biohackers from the DiSSCo partners Naturalis Biodiversity Center, Royal Museum for Central Africa, Museum für Naturkunde Berlin, Natural History Museum London, Complutense University of Madrid, Senckenberg Society for Nature Research and Nature Research Centre Vilnius met at the Senckenberg Research Institute Frankfurt/Main to integrate new machine annotation services (MAS) for DiSSCo’s curation and annotation platform DiSSCover.

Three teams focused on the adoption of services for DiSSCo involving (i) taxamorph, a machine learning-based tool to highlight morphologically significant features of biological specimen, (ii) quality annotation of AI-generated herbarium sheets parts combining the tools LeafMachine and INDEED, and (iii) mapping of habitat descriptions to controlled vocabularies using Large Language Models based on Mistral AI and OntoGPT.

In addition, the Machine Annotation Developer documentation was updated and extended (https://dissco.github.io/mas-
The meeting showed convincingly, how DiSSCo’s e-service portfolio can be significantly extended by virtue of open standards – including notably the Open Digital Specimen data model – adopted by its design.

Event Dates: March 24th -26th, 2025
Event Address: Senckenberg Research Institute, Senckenberganalage 25, Frankfurt am Main, Germany
THIS INFORMATION IS ABOUT A PAST EVENT
REGISTRATION IS NO LONGER AVAILABLE
More info: rajapreethi.rajendran@senckenberg.de
Do you want to know more about the technical side of DiSSCo? DiSSCo puts different technical knowledge platforms at the scientific community’s disposal:
DiSSCoTech: Get the latest technical posts about the design of DiSSCo’s Infrastructure
DiSSCo Labs: A preview of experimental services and demonstrators by the DiSSCo community
DiSSCo GitHub: Code hosting for DiSSCo software, version control and collaboration
DiSSCo Modelling Framework: A WikiBase tool that is configured to create an abstraction of the DiSSCo data model