Optimal Sequential Recommendations: Exploiting User and Item Structure
Mina Karzand, Guy Bresler · 2025
Read Optimal Sequential Recommendations: Exploiting User and Item Structure by Mina Karzand, Guy Bresler (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-19476.txt.
Description
Mina Karzand and Guy Bresler design recommendation methods that learn from both users and items in this 2025 paper. Their model serves one item per user per round and hears only like or dislike. Users and items each fall into hidden types, and tastes run by type pairs.
Most real systems lean on just one side's structure. The pair's algorithm uses both at once. They prove it nearly optimal by matching it against information-theoretic limits no method can beat. The gap left by one-sided methods is shown plainly. The result guides how streaming and shopping services should pool their data.
Catalog record
| CREATOR | Mina Karzand, Guy Bresler |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 251 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-19476