USDB: Scaling Storage for Next-Generation Recommender Systems at ByteDance

arxiv.org

49 points·1equalsequals1··8 comments

Setting up a classical DLRM at work, show of hands: who has managed to bridge the gap to generative recommenders or LRMs?

Our survey identified quite a substantial list of challenges: product cardinality (Semantic IDs?), low performance vs discriminative recommenders (our user intents are heavily biases towards repetitive products), cold starts, entity features, missing explainability suite, training cost, prediction cost, etc.

Has anyone done it easily? Where is the tutorial for generative recommenders ?

GR expertise is essentially only in-house at major social media companies or people hired from these companies. Closest thing to a tutorial for LRM is the Kuaishou onerec papers.

Yeah OneRec is on our radar. I'd love to adapt their pipeline, but it's a very long shot in the small company I'm in.

So bytedance has the luxury of building custom storage for this, meanwhile we're over here trying to shoehorn user sequences into redis sorted sets

The future of recommenders is just an agent paging through your data and trying picking the best next options

That's how they've been working for decades, ever since Facebook changed timeline to feed.

What does this actually mean? The term "agent" is designed to evoke an image of a person riffling through their files on you but can't the way you just described it describe any feed algorithm, including chronological?

an ai agent, not a human, its an llm

Reply on news.ycombinator.com