GeomHerd: A Forward-looking Herding Quantification via Ricci Flow Geometry on Agent Interactive Simulations
Lake Yang, Junwei Su, Jingfeng Zeng, Wenhao Lu, Xingzhi Qian, Weitong Zhang, Chuan Wu, Dunhong Jin · 2026
Read GeomHerd: A Forward-looking Herding Quantification via Ricci Flow Geometry on Agent Interactive Simulations by Lake Yang, Junwei Su, Jingfeng Zeng, Wenhao Lu, Xingzhi Qian, Weitong Zhang, Chuan Wu, Dunhong Jin (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-2605-11645.txt.
Description
Yang and seven colleagues present GeomHerd, forward-looking herding quantification via Ricci flow geometry. Herding, where agents align and act collectively, drives market fragility, but price statistics lag because they detect coordination only after it moves returns. GeomHerd quantifies coordination directly on upstream agent-interaction graphs from an LLM-driven multi-agent simulator, tracking discrete Ollivier-Ricci curvature. A mean-field bridge links the graph metric to CSAD, the classical herding statistic. On a continuous-spin testbed the main detector fires a median 272 steps before order-parameter onset.
Catalog record
| CREATOR | Lake Yang, Junwei Su, Jingfeng Zeng, Wenhao Lu, Xingzhi Qian, Weitong Zhang, Chuan Wu, Dunhong Jin |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 60348 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.11645