Tensor arithmetics in search and ranking for Ecommerce.
Session Abstract
Small, domain-specific vision models can dramatically enhance the buyer search experience by delivering more relevant visual understanding. But the real opportunity comes from controllable image embeddings: by fusing base search embeddings with additional control vectors, representing features such as color, shape, and style,
Session Description
In this talk, we’ll show how Vespa’s ranking engine, advanced fashion visual model, supported by tensor operations enables dense vector–based search tuning for fashion apparel search, making product discovery faster, more intuitive, and more aligned with user preferences. We’ll walk through the configuration examples of how these ranking pipelines are built in Vespa: from schema design and tensor field definitions, to ranking expressions using tensor arithmetic, to practical patterns for production deployment.
—
This talk is sponsored by Vespa.ai
