CHIME: A Compressive Framework for Holistic Interest Modeling
Yong Bai, Rui Xiang, Kaiyuan Li, Yongxiang Tang, Yanhua Cheng, Xialong Liu, Peng Jiang, Kun Gai · 2025
Read CHIME: A Compressive Framework for Holistic Interest Modeling by Yong Bai, Rui Xiang, Kaiyuan Li, Yongxiang Tang, Yanhua Cheng, Xialong Liu, Peng Jiang, Kun Gai (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-2504-06780.txt.
Description
Yong Bai and co-authors present CHIME in a 2025 paper on modeling what users like as a whole. Recommender systems gain from full interest models, but cost soars and mixed data types are hard to fuse. Search based methods that pick a slice of past behavior can drop key signals.
CHIME adapts large language models to encode a user's full behavior with mixed inputs. A learning aim at two grain sizes catches both lasting tastes and brief whims. A coding step then packs those codes into compact vectors for fast use. Tests across diverse data sets show top ranking scores for the method. The authors pitch CHIME as a robust, scalable fix for interest modeling at platform scale.
Catalog record
| CREATOR | Yong Bai, Rui Xiang, Kaiyuan Li, Yongxiang Tang, Yanhua Cheng, Xialong Liu, Peng Jiang, Kun Gai |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 46 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-06780