Learning Higher-Order Structure from Incomplete Spatiotemporal Data: Multi-Scale Hypergraph Laplacians with Neural Refinement
Keshu Wu, Sixu Li, Zihao Li, Zhiwen Fan, Xiaopeng Li, Yang Zhou · 2026
Read Learning Higher-Order Structure from Incomplete Spatiotemporal Data: Multi-Scale Hypergraph Laplacians with Neural Refinement by Keshu Wu, Sixu Li, Zihao Li, Zhiwen Fan, Xiaopeng Li, Yang Zhou (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-2605-17316.txt.
Introduction
Wu and colleagues introduce MoHL, a mix of hypergraph Laplacians learning group relations from incomplete traffic-sensor data with long gaps. It finds overlapping sensor groups from observed co-variation, builds one Laplacian per interaction order, and mixes them through holdouts mimicking the missing pattern. Sharing weights across group couplings pools thin evidence, and tests on three traffic networks give the lowest error under day-long gaps with far fewer parameters than deep imputers.
Catalog record
| CREATOR | Keshu Wu, Sixu Li, Zihao Li, Zhiwen Fan, Xiaopeng Li, Yang Zhou |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 130116 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.17316