Talk to Your Library: AI-Powered Multimedia Discovery

Beyond Text Search: Multimodal Retrieval for Modern Libraries How can users find images using text descriptions, discover books from photos, or search audio collections with written queries? This presentation explores multimodal retrieval technologies that enable cross-modal search across text, images, and audio. We will cover how computers represent meaning through vectors, align different media types in shared mathematical spaces, and practical applications including automatic keyword extraction and intelligent search systems. Learn about breakthrough technologies like CLIP and vector databases that make large-scale multimodal search practical. The session includes live demonstrations, allowing experimentation with searching across different media types using text and image content to find similar items.

People from GLAM institutions who will use AI to bring digital media closer to users.