Search by Similarity: Explore Collections with Embeddings

This second meeting is dedicated to embeddings, digital representations that allow for the comparison of texts and images based on their semantic proximity. Through projects and case studies, we will discover how this approach can enrich search engines and open new exploration pathways within collections.

At the heart of this session:

• Understanding the principle of embeddings and similarity search.

• Searching from a natural language query, a document, or an image.

• Understanding the limitations of the proposed matches 

• Articulating embedding search with traditional search (text and metadata)

The agenda

• 15 minutes for an overview of new developments (technology watch and public policies)
• 15 min pour un focus sur les moteurs de recherche par embeddings
• 15 minutes of questions and answers

Who is it for? Professionals from archives, museums, and heritage institutions involved in the description, documentation, research, or promotion of collections. 

Online — Duration: 45 min (in French)

TEKLIA AI Meetings

The TEKLIA AI Meetings offer a series of five webinars dedicated to the uses of artificial intelligence in archives and museums. Each session combines a time for monitoring technological developments and public policies, a thematic focus, and an exchange with participants.