MuonRec: Shifting the Optimizer Paradigm Beyond Adam in Scalable Generative Recommendation CCS: Information systems Recommender systems
Rong Shan, Aofan Yu, Bo Chen, Kuo Cai, Qiang Luo, Ruiming Tang, Han Li, Weiwen Liu, Weinan Zhang, Jianghao Lin · 2026
Read MuonRec: Shifting the Optimizer Paradigm Beyond Adam in Scalable Generative Recommendation CCS: Information systems Recommender systems by Rong Shan, Aofan Yu, Bo Chen, Kuo Cai, Qiang Luo, Ruiming Tang, Han Li, Weiwen Liu, Weinan Zhang, Jianghao Lin (2026) 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-2603-00416.txt.
Description
Shan and nine colleagues question the default choice of Adam style optimizers in large recommender training and bring the Muon optimizer into that setting for the first time. Muon updates two dimensional weight matrices through orthogonalized momentum steps, spreading update directions more widely. Across classic sequential and modern generative recommenders, their MuonRec recipe cuts converged training steps by about a third while lifting ranking quality. Gains hold against strong Adam baselines, most of all in generative recommenders.
Catalog record
| CREATOR | Rong Shan, Aofan Yu, Bo Chen, Kuo Cai, Qiang Luo, Ruiming Tang, Han Li, Weiwen Liu, Weinan Zhang, Jianghao Lin |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 25465 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.00416