Museum records often contain errors or blurry information. Many objects in databases are named under generic terms that provide little insight into the specific nature or function of an object. Moreover, some of these designations reflect colonial or outdated ways of thinking in which the objects were collected and documented.

This challenge addresses this problem and explores several methods for identifying generic, or incorrect object names in museum databases, as well as suggesting more accurate and context-specific terms to rename these objects.

The goal is not to automate cataloguing, but to support museum professionals in reviewing and improving metadata through evidence-based recommendations.

Research questions:

• Which object names occur particularly often as (too) generic, ambiguous, outdated, or uncertain designations?
• Can image similarity searches help identify comparable objects in other collections? Are there other AI tools that can efficiently support renaming efforts?
• Which museum or community databases are most useful for suggesting more specific or alternative object names?
• ...

Data sources:

We will work with a sample dataset from the Ethnographic Museum at the University of Zurich. Possible comparative databases:

• Rietberg Museum Zürich: https://rietberg.ch/sammlungen
• Pitt Rivers Museum: https://www.prm.ox.ac.uk/collections-online#/search
• Museum der Kulturen Basel: https://onlinecollection.mkb.ch/#/query/e65ddcbe-62de-4949-878e-c2f759c5ad46
• Mapping Philippine Material Culture: https://philippinestudies.uk/mapping/
• Yukon First Nations: https://searching-for-our-heritage.service.yukon.ca/
• Amutat database (Alutiiq): https://alutiiqmuseum.org/collection/index.php/browse/amutatObjects
• ...

Examples:

• Bambusgitarre → Röhrenzither (generic/eurocentric designation)
• Figur → Figur: Gottheitsstatue (generic designation)
• Körbli → Korb, klein (improved discoverability through standardization)

Who we need:

Programmers, developers, data scientists, museum professionals, cataloguers, metadata specialists, community heritage experts, researchers, and anyone interested in AI, cultural heritage, decolonising collections, and improving access to museum data.

What we bring:

7,000 question marks, a century of cataloguing history, plenty of metadata puzzles, and hopefully fewer "Figures" by Saturday afternoon