Gen-C: Populating Virtual Worlds with Generative Crowds
Andreas Panayiotou, Panayiotis Charalambous, Ioannis Karamouzas · 2025
Read Gen-C: Populating Virtual Worlds with Generative Crowds by Andreas Panayiotou, Panayiotis Charalambous, Ioannis Karamouzas (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-2504-01924.txt.
Description
Andreas Panayiotou, Panayiotis Charalambous, and Ioannis Karamouzas present Gen-C, a system that fills virtual worlds with crowds that act out rich plans, in a research paper from 2025. Older crowd simulators handle low-level tasks well, like avoiding bumps and following paths. They struggle with the higher-level behavior that grows from long contact among people and places.
Gen-C stands for Generative Crowds and builds crowd scenes as time-expanded graphs of actions, contacts, and places. To train it, the team used large language models to draft synthetic crowd data, avoiding costly video labeling. Two linked learning parts then learn the graph links and the node traits from text and structure. Tests on a campus and a train station gave mixed crowds with coherent, context-fitting plans.
Catalog record
| CREATOR | Andreas Panayiotou, Panayiotis Charalambous, Ioannis Karamouzas |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 41 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-01924