Samuel Waugh's N8 CIR Summer Internship leads to Development of new AI Tool for Art Contextualisation

This summer, Durham University Computer Science student Samuel Waugh completed a research internship with the N8 Centre of Excellence in Computationally Intensive Research (N8 CIR) under my supervision. Over the course of the project, Samuel designed and developed ArteFact, an innovative web-based tool that helps art historians and cultural researchers uncover meaningful connections between paintings and scholarly writing.

At the core of ArteFact is a domain-adapted CLIP model, trained on an open-access art history corpus. This enables the system to embed sentences from scholarly texts and recommend the most relevant passages in response to a user-uploaded painting. The platform integrates a range of features, including:

• Interactive image cropping for targeted analysis
• Saliency map generation to highlight important visual regions
• Contextualised sentence viewing showing extracted text in its original source
• Instant access to open access works linked to the findings
• Citation export for proper scholarly acknowledgement

The project was shaped through co-design workshops with Durham’s Institute of Medieval and Early Modern Studies (IMEMS). Feedback from humanities scholars ensured that the tool is not only technically sophisticated but also genuinely useful to its intended audience.

What impressed me most was how quickly Samuel was able to translate a complex research idea into a usable system. He showed real creativity, technical skill, and an ability to think carefully about end users – qualities that will serve him very well in his career.

A demo version of ArteFact is now available to explore online: https://huggingface.co/spaces/samwaugh/ArteFact

Code at: https://github.com/sammwaughh/ArteFact