VLDB 2025 | 时空数据（Spatial-Temporal）论文总结(时空预测，轨迹相似度，轨迹表示等)

时空探索之旅

发布于 2026-03-10 14:58:52

530

VLDB 2025于2025年9月1号-5号在英国伦敦(London, United Kingdom)举行。

本文总结了VLDB 2025有关时空数据（Spatial Temporal）的相关论文，主要包含如有疏漏，欢迎大家补充。

时空数据Topic：时空预测，交通预测，轨迹相似度，轨迹表示等。

1. TEAM: Topological Evolution-aware Framework for Traffic Forecasting2. Quantifying Point Contributions: A Lightweight Framework for Efficient and Effective Query-Driven Trajectory Simplification3. SIMformer: Single-Layer Vanilla Transformer Can Learn Free-Space Trajectory Similarity4. RED: Effective Trajectory Representation Learning with Comprehensive Information5. T-Assess: An Efficient Data Quality Assessment System Tailored for Trajectory Data6. Revisiting CNNs for Trajectory Similarity Learning7. Mining Platoon Patterns from Traffic Videos8. SimRN: Trajectory Similarity Learning in Road Networks based on Distributed Deep Reinforcement Learning9. GraphSparseNet: a Novel Method for Large Scale Traffic Flow Prediction10. Scalable Pre-Training of Compact Urban Spatio-Temporal Predictive Models on Large-Scale Multi-Domain Data11. TMLKD: Few-shot Trajectory Metric Learning via Knowledge Distillation12. BiST: A Lightweight and Efficient Bi-directional Model for Spatiotemporal Prediction

1 TEAM: Topological Evolution-aware Framework for Traffic Forecasting

链接：https://www.vldb.org/pvldb/vol18/p265-kieu.pdf

代码：https://github.com/kvmduc/TEAM-topo-evo-traffic-forecasting

作者：Duc Kieu, Tung Kieu, Peng Han, Bin Yang, Christian S. Jensen, Bac Le

关键词：交通预测，拓扑演化感知

2 Quantifying Point Contributions: A Lightweight Framework for Efficient and Effective Query-Driven Trajectory Simplification

链接：https://www.vldb.org/pvldb/vol18/p453-gu.pdf

代码：https://github.com/yumengs-exp/MLSimp

作者：Yumeng Song, Yu Gu, Tianyi Li, Yushuai Li, Christian S. Jensen, Ge Yu

关键词：轨迹查询，轨迹简化

3 SIMformer: Single-Layer Vanilla Transformer Can Learn Free-Space Trajectory Similarity

链接：https://www.vldb.org/pvldb/vol18/p390-jiang.pdf

代码：https://github.com/SUSTC-ChuangYANG/SIMformer/

作者：Chuang Yang, Renhe Jiang, Xiaohang Xu, Chuan Xiao, Kaoru Sezaki

关键词：轨迹相似度，Transformer

4 RED: Effective Trajectory Representation Learning with Comprehensive Information

链接：https://www.vldb.org/pvldb/vol18/p80-zhou.pdf

代码：https://github.com/slzhou-xy/RED

作者：Silin Zhou, Shuo Shang, Lisi Chen, Christian S. Jensen, Panos Kalnis

关键词：轨迹表示学习

5 T-Assess: An Efficient Data Quality Assessment System Tailored for Trajectory Data

链接：https://www.vldb.org/pvldb/vol18/p666-chen.pdf

代码：https://github.com/ZJU-DAILY/T-Assess

作者：Junhao Zhu, Tao Wang, Danlei Hu, Ziquan Fang, Lu Chen, Yunjun Gao, Tianyi Li, Christian S. Jensen

关键词：自动化，轨迹数据质量评估系统

6 Revisiting CNNs for Trajectory Similarity Learning

链接：https://www.vldb.org/pvldb/vol18/p1013-chang.pdf

代码：https://github.com/Proudc/ConvTraj

作者：Zhihao Chang, Linzhu Yu, Huan Li, Sai Wu, Gang Chen, Dongxiang Zhang

关键词：轨迹相似度，CNN

7 Mining Platoon Patterns from Traffic Videos

链接：https://www.vldb.org/pvldb/vol18/p1839-bei.pdf

代码：https://github.com/Mateng0228/Vplatoon

作者：Yijun Bei, Teng Ma, Dongxiang Zhang, Sai Wu, Kian-Lee Tan, Gang Chen

关键词：交通视频，车队模式

8 SimRN: Trajectory Similarity Learning in Road Networks based on Distributed Deep Reinforcement Learning

链接：https://www.vldb.org/pvldb/vol18/p2057-chen.pdf

代码：https://github.com/ZJU-DAILY/SimRN

作者：Danlei Hu, Yilin Li, Lu Chen, Ziquan Fang, Yushuai Li, Yunjun Gao, Tianyi Li

关键词：轨迹相似度，路网，分布式强化学习

9 GraphSparseNet: a Novel Method for Large Scale Traffic Flow Prediction

链接：https://www.vldb.org/pvldb/vol18/p2295-liu.pdf

代码：https://github.com/PolynomeK/GSNet

作者：Weiyang Kong, Wu Kaiqi, Sen Zhang, Yubao Liu

关键词：大规模交通预测

10 Scalable Pre-Training of Compact Urban Spatio-Temporal Predictive Models on Large-Scale Multi-Domain Data

链接：https://www.vldb.org/pvldb/vol18/p2149-han.pdf

代码：https://github.com/usail-hkust/CompactST

作者：Jindong Han, Hao Wang, Hui Xiong, Hao Liu

关键词：时空预测，可扩展，多域（模态）数据

11 TMLKD: Few-shot Trajectory Metric Learning via Knowledge Distillation

链接：https://www.vldb.org/pvldb/vol18/p2308-lai.pdf

代码：https://github.com/LynnTakesAHint/TMLKD

作者：Danling Lai, Jiajie Xu, Jianfeng Qu, Pingfu Chao, Junhua Fang, Chengfei Liu

关键词：轨迹度量学习，知识蒸馏

12 BiST: A Lightweight and Efficient Bi-directional Model for Spatiotemporal Prediction

链接：https://www.vldb.org/pvldb/vol18/p1663-wang.pdf

代码：https://github.com/PoorOtterBob/BiST

作者：Jiaming Ma, Binwu Wang, Pengkun Wang, Zhengyang Zhou, Xu Wang, Yang Wang

关键词：时空预测，轻量化