跳到正文
arXiv cs.AI· Hanyu Guo, Jiedong Yang, Chao Chen, Longfei Xu, Kaikui Liu, Xiangxiang Chu·· 4 小时前AI 评分42

TransitLM:面向无地图公交路线生成的大规模数据集与基准

TransitLM: A Large-Scale Dataset and Benchmark for Map-Free Transit Route Generation

AI 导读

TransitLM 数据集与基准被 NeurIPS 2026 评估与数据集赛道接收,包含来自中国四个城市的近 1300 万条公交路线规划记录,覆盖 120,845 个站点和 13,666 条线路,作为持续预训练语料及三项评估任务发布。

正文 · 原文

View PDF HTML (experimental)

Abstract:Public transit route planning traditionally depends on structured map infrastructure and complex routing engines, and no existing dataset supports training models to bypass this dependency. We present TransitLM, a large-scale dataset of nearly 13 million transit route planning records from four Chinese cities covering 120,845 stations and 13,666 lines, released as a continual pre-training corpus and benchmark data for three evaluation tasks with complementary metrics. Experiments show that an LLM trained on TransitLM produces structurally valid routes at high accuracy and implicitly grounds arbitrary GPS coordinates to appropriate stations without any explicit mapping. These results demonstrate that transit route planning can be learned entirely from data, enabling end-to-end, map-free route generation directly from origin-destination information. The dataset and benchmark are available at this https URL, with evaluation code at this https URL.
Comments: Accepted at NeurIPS 2026 (Evaluations & Datasets Track)
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Cite as: arXiv:2605.22355 [cs.CL]
  (or arXiv:2605.22355v2 [cs.CL] for this version)
  https://doi.org/10.48550/arXiv.2605.22355

arXiv-issued DOI via DataCite

Submission history

From: Hanyu Guo [view email]
[v1] Thu, 21 May 2026 11:42:33 UTC (3,126 KB)
[v2] Fri, 13 Nov 2026 09:37:08 UTC (3,132 KB)

来源:arXiv cs.AI · arxiv.org