Dynamic Population Distribution Aware Human Trajectory Generation with Diffusion Model Journal: TIST CCS: Information systems Spatial-temporal systems CCS: Computing methodologies Modeling methodologies CCS: Computing methodologies Neural networks

Qingyue Long, Can Rong, Tong Li, Yong Li · 2025

Read Dynamic Population Distribution Aware Human Trajectory Generation with Diffusion Model Journal: TIST CCS: Information systems Spatial-temporal systems CCS: Computing methodologies Modeling methodologies CCS: Computing methodologies Neural networks by Qingyue Long, Can Rong, Tong Li, Yong Li (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-2511-01929.txt.

Description

Qingyue Long, Can Rong, Tong Li, and Yong Li tackle a data problem in city science: how to create fake human travel records that still behave like real ones. Real trajectory data, the paths people take through a city, serve planning, traffic, and health work. Yet real records raise privacy risks, cost much to gather, and often have quality flaws.

Earlier generators mostly copy single person movement patterns. They miss how the spread of a whole population, shifting across districts over time, shapes each person's trip. The team's diffusion model adds that missing piece. A spatial graph strengthens links between places. A denoising network built to sense population spread learns the patterns of movement in space and time. In tests, the fake trips matched real ones on key statistical measures, beating the best prior methods by more than half.

Catalog record

CREATORQingyue Long, Can Rong, Tong Li, Yong Li
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE63 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-01929

Browse the full library · About the Tower of Babel