top of page
From Signals to Trends: Building Multimodal AI for Fashion
What if fashion trends could be understood as evolving signals over time?
This workshop will explore how multimodal AI connects fashion images and language through a shared taxonomy. We will present the technical journey behind a system that understands garments, colors, patterns, materials and visual noise, enabling natural-language discovery across visual and textual data.
The session will focus on taxonomy design, image annotation, model alignment, search relevance and LLMs at scale, with time-series analysis providing additional insight into how trends evolve.
Speaker list
bottom of page


