Short-video Propagation Influence Rating: A New Real-world Dataset and A New Large Graph Model DOI: XXXXXXX.XXXXXXX Conference: ; ;
Dizhan Xue, Jing Cui, Shengsheng Qian, Chuanrui Hu, Changsheng Xu · 2025
Read Short-video Propagation Influence Rating: A New Real-world Dataset and A New Large Graph Model DOI: XXXXXXX.XXXXXXX Conference: ; ; by Dizhan Xue, Jing Cui, Shengsheng Qian, Chuanrui Hu, Changsheng Xu (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-2503-23746.txt.
Description
Dizhan Xue and co-authors build a dataset and a model for predicting how far short videos will spread online, in a 2025 paper. Short-video apps draw billions of users, and knowing which clips will travel matters for business and for tracking public opinion. The team's dataset gathers 117,720 videos from the five biggest Chinese platforms, with counts of views, likes, shares, saves, and comments, and rates each video's spread on a ten-step scale.
Their model, NetGPT, joins a graph of how videos and users connect with the reasoning power of a large language model. It trains in three stages so the language model learns to read the graph's structure. On tests against the new dataset, NetGPT beat rival methods on both sorting videos into spread levels and estimating spread as a number. The authors offer the pair as a base for studying how short clips move across platforms.
Catalog record
| CREATOR | Dizhan Xue, Jing Cui, Shengsheng Qian, Chuanrui Hu, Changsheng Xu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 51 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-23746