Distinguished Quantized Guidance for Diffusion-based Sequence Recommendation
Wenyu Mao*, Shuchang Liu*, Haoyang Liu, Haozhe Liu, Xiang Li, Lantao Hu · 2025
Read Distinguished Quantized Guidance for Diffusion-based Sequence Recommendation by Wenyu Mao*, Shuchang Liu*, Haoyang Liu, Haozhe Liu, Xiang Li, Lantao Hu (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-2501-17670.txt.
Description
Wenyu Mao and five co-authors improve recommendation with diffusion models in this two thousand twenty-five paper. Sequential recommenders guess the next item a user will want from their history. Diffusion models, which build items by slowly removing noise, have joined this field. The history guiding them, though, is messy: users act at random, and lengths vary widely. Noisy guidance can steer the model wrong. The models also lean toward popular items, slighting personal taste.
Their method adds a distinguished quantized guidance step. It cleans and compresses the user history into a steadier guide. The diffusion process then aims at items that fit the person, not just the crowd. The paper targets researchers who build next-item recommenders for shops and media apps.
Catalog record
| CREATOR | Wenyu Mao*, Shuchang Liu*, Haoyang Liu, Haozhe Liu, Xiang Li, Lantao Hu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-17670