TrajDLM: Topology-Aware Block Diffusion Language Model for Trajectory Generation
Wilson Wongso 1 Lihuan Li 1 Arian Prabowo 1 Xiachong Lin 1 Baiyu Chen 1 Hao Xue 1,2 Flora Salim 1 1 University of New South Wales 2 Hong Kong University of Science and Technology (Guangzhou) · 2026
Read TrajDLM: Topology-Aware Block Diffusion Language Model for Trajectory Generation by Wilson Wongso 1 Lihuan Li 1 Arian Prabowo 1 Xiachong Lin 1 Baiyu Chen 1 Hao Xue 1,2 Flora Salim 1 1 University of New South Wales 2 Hong Kong University of Science and Technology (Guangzhou) (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-2605-10020.txt.
Description
Wongso and six colleagues built TrajDLM, a topology-aware block diffusion text model for trip making. True-to-life synthetic GPS trips matter for transport, urban planning, and what-if tests, since privacy concerns limit access to real travel data. Current models trade speed against road network shape. Free-space methods run fast but ignore the road network. Topology-aware methods rely on slow look-back search.
TrajDLM models trips as chains of single road segments. It joins a block core for fast cleanup, road-aware codes from a road network encoder, and road-limited sampling for linked, true-to-life trips. Across three city-scale sets, it scored well on fine-grained local match scores while running up to 2.8 times faster than prior work. It also showed strong zero-shot transfer across domains, plus unseen transport modes.
Catalog record
| CREATOR | Wilson Wongso 1 Lihuan Li 1 Arian Prabowo 1 Xiachong Lin 1 Baiyu Chen 1 Hao Xue 1,2 Flora Salim 1 1 University of New South Wales 2 Hong Kong University of Science and Technology (Guangzhou) |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 55 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.10020