HGMP:Heterogeneous Graph Multi-Task Prompt Learning
Pengfei Jiao, Jialong Ni, Di Jin, Xuan Guo, Huan Liu, Hongjiang Chen, Yanxian Bi · 2025
Read HGMP:Heterogeneous Graph Multi-Task Prompt Learning by Pengfei Jiao, Jialong Ni, Di Jin, Xuan Guo, Huan Liu, Hongjiang Chen, Yanxian Bi (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-2507-07405.txt.
Description
Pengfei Jiao and co-authors present HGMP, a prompt based framework for learning on mixed type graphs, in a 2025 paper. Such graphs hold many kinds of nodes and links, as in social or product networks. Pre-trained graph models often fit new tasks poorly, since the training goal and the task goal clash.
HGMP recasts all target tasks into one graph level form. A contrast based pre-training step is shaped for mixed graphs, and prompts that refine node features help the model adapt. Public data tests show the method fits varied tasks and beats prior baselines by clear margins. The work extends prompt learning, a hit in language models, to graph data.
Catalog record
| CREATOR | Pengfei Jiao, Jialong Ni, Di Jin, Xuan Guo, Huan Liu, Hongjiang Chen, Yanxian Bi |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-07405