Research map / Label-efficient and robust learning / Self-supervised and contrastive
Self-Supervised Cross-City Trajectory Representation Learning Based on Meta-Learning
Yanwei Yu, Hong Xia, Shaoxuan Gu, Xingyu Zhao, Dongliang Chen, Yuan Cao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| train baselines on BJ then fine-tune on Porto (cross-city) | TrajRL · Porto destination prediction | Micro-F1: 0.563 -> 0.535 | Hurt |
| train MetaTRL cross-city (BJ->Porto) instead of on Porto only | MetaTRL · Porto destination prediction | Micro-F1: 0.594 -> 0.642 | Improved |
| add the Meta Knowledge Enhanced Road Encoder (vs plain GAT) (ablation) | MetaTRL · Porto (CLA, TTE, DP) | Micro-F1 / MAE | Improved |
| add the Transferable Time Encoder (ablation) | MetaTRL · Porto (CLA, TTE, DP) | Micro-F1 / MAE | Improved |
| add the shared time interval encoder (ablation) | MetaTRL · Porto (CLA, TTE, DP) | MAE | Improved |
| add the Knowledge Fusion Trajectory Encoder (vs standard transformer) (ablation) | MetaTRL · Porto (CLA, TTE, DP) | Micro-F1 / MAE | Improved |
| add time masking and recovery task (ablation) | MetaTRL · Porto (CLA, TTE, DP) | Micro-F1 / MAE | Improved |
| increase embedding size to very large values (ablation) | MetaTRL · Porto (TTE, DP) | MAE / Micro-F1 | Hurt |
| increase batch size to excessive values (ablation) | MetaTRL · Porto (TTE, DP) | MAE / Micro-F1 | Hurt |
| MetaTRL vs TrajRL on destination prediction in larger BJ dataset | destination prediction · BJ | Micro-F1: 0.173 -> 0.396 | Improved |
Closest papers by meaning
- UniTR: A Unified Framework for Joint Representation Learning of Trajectories and Road Networks (AAAI 2025 · similarity 0.86)
- Bridging Traffic State and Trajectory for Dynamic Road Network and Trajectory Representation Learning (AAAI 2025 · similarity 0.86)
- GTR: A General, Multi-View, and Dynamic Framework for Trajectory Representation Learning (ICML 2025 · similarity 0.86)
- TrajMamba: An Efficient and Semantic-rich Vehicle Trajectory Pre-training Model (NeurIPS 2025 · similarity 0.85)
- STraj: Self-training for Bridging the Cross-Geography Gap in Trajectory Prediction (AAAI 2025 · similarity 0.85)
- Enhancing Trajectory Prediction through Self-Supervised Waypoint Distortion Prediction (ICML 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
