An Empirical Survey and Benchmark of Learned Distance Indexes for Road Networks
Gautam Choudhary, Libin Zhou, Yeasir Rayhan, Walid G. Aref · 2026
Read An Empirical Survey and Benchmark of Learned Distance Indexes for Road Networks by Gautam Choudhary, Libin Zhou, Yeasir Rayhan, Walid G. Aref (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-2602-04068.txt.
Description
Learned indexes promise fast distance answers on road networks. Gautam Choudhary and 3 co-authors ran the first broad empirical comparison in a 2026 survey paper. Exact algorithms like Dijkstra's are too slow for real-time, planet-scale queries, so indexes precompute or learn shortcuts.
Ten representative machine-learning techniques were benchmarked on thirteen real road networks. The study measured training time, query latency, storage, and accuracy, using both trajectory-derived and synthetic workloads. No prior survey had compared these methods head to head under one protocol. The results give system builders practical guidance on which learned index fits which deployment constraints.
Catalog record
| CREATOR | Gautam Choudhary, Libin Zhou, Yeasir Rayhan, Walid G. Aref |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 73 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-04068