SANSA: Our approach to recommendation from millions of items

Providing accurate recommendations to new users with minimal interaction history presents an untapped gold mine for commercial recommender systems. The challenge, intensified by a vast item catalog, resembles finding a needle in a haystack. The question arises: How can we identify the few relevant items from millions, based solely on a handful of past interactions? The solution involves linking feedback from numerous users into 'chains.' Traditional implementations of this strategy, however, are very resource-intensive.

We have developed an approach that is orders of magnitude more efficient and are pleased to introduce our model, SANSA. Remarkably, SANSA can train a model of equivalent complexity on a standard laptop, which would normally demand a much more powerful computer. As an open-source solution, we encourage you to explore SANSA at 👉 https://github.com/glami/sansa.

👩‍💻 Explore SANSA: As an open-source solution, SANSA is accessible to everyone. Check it out at https://github.com/glami/sansa.

📄 Read Our Paper: For a deeper understanding of SANSA, read our award-winning paper available here: https://dl.acm.org/doi/10.1145/3604915.3608827.

👉🏻 SECURE YOUR FREE TICKET NOW! 👈🏻
------------------------------------------------------------------------------------------------------
⌚️ Start: 16:00 and end at 17:00, both online https://meet.google.com/pzm-inpb-rww and offline in Miton offices, Křížíkova 34, Praha - reception will have the necessary information.
🎙️ Speakers: Martin Spišák (GLAMI)
🍻 Networking after the talk.
🎥 Recording: After the event we will publish a recording and post a link to it in the comments.
🚪Doors open at 15:45, and the event officially starts at 16:00.

Don't miss out—SECURE YOUR FREE TICKET NOW! 🎉🎟️