Onboarding Project:

• Create a Meta Page
• Define the exploratory image domain: monuments
• Create dataset of items and image candidates
• Build a Relevance Model
• Build a Quality model

Q3 goal:

1. Retrieve lists of wikidata items T184734
2. Retrieve candidate images for wikidata items T184737
3. Filter candidate images by relevance T184738
4. Build a model for image quality T184739
5. Filter candidate images by quality T184740
6. Evaluate

Obervations - bottlenecks:

• 1 needs domain expertise for e.g. monuments and distribute process for reading dumps
• 4 speed-up with GPU, also pixels are not available internally