mlsanityData to 7 Oct 2026

Exploiting map information for self-supervised learning in motion forecasting

arXiv 2210.04672 ↗First posted 10 Oct 2022cs.CVcited by 8

Not on the map. Why?

Abstract (arXiv)

Inspired by recent developments regarding the application of self-supervised learning (SSL), we devise an auxiliary task for trajectory prediction that takes advantage of map-only information such as graph connectivity with the intent of improving map comprehension and generalization. We apply this auxiliary task through two frameworks - multitasking and pretraining. In either framework we observe significant improvement of our baseline in metrics such as minFDE6 (as much as 20.3%) and MissRate6 (as much as 33.3%), as well as a richer comprehension of map features demonstrated by different training configurations. The results obtained were consistent in all three data sets used for experiments: Argoverse, Interaction and NuScenes. We also submit our new pretrained model's results to the Interaction challenge and achieve textit1st place with respect to minFDE6 and minADE6.

Key facts. Cited by 8 papers in the map (data through 7 Oct 2026). Builds on 33 of its 43 references that are in the map. First posted on arXiv 10 Oct 2022.

Builds on

33 of 43 references.

References · 33

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  3. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,339
  4. 2019nuScenes: A multimodal dataset for autonomous drivingcited by 4,485
  5. 2015Fast R-CNNcited by 2,930
  6. 2019Argoverse: 3D Tracking and Forecasting with Rich Mapscited by 728
  7. 2018Social GAN: Socially Acceptable Trajectories with Generative Adversarial Networkscited by 687
  8. 2020VectorNet: Encoding HD Maps and Agent Dynamics from Vectorized Representationcited by 528
  9. 2019MultiPath: Multiple Probabilistic Anchor Trajectory Hypotheses for Behavior Predictioncited by 417
  10. 2020Learning Lane Graph Representations for Motion Forecastingcited by 380
  11. 2021AgentFormer: Agent-Aware Transformers for Socio-Temporal Multi-Agent Forecastingcited by 327
  12. 2020Social-STGCNN: A Social Spatio-Temporal Graph Convolutional Neural Network for Human Trajectory Predictioncited by 324
21 more
  1. 2021DenseTNT: End-to-end Trajectory Prediction from Dense Goal Setscited by 292
  2. 2022Wayformer: Motion Forecasting via Simple & Efficient Attention Networkscited by 279
  3. 2019CoverNet: Multimodal Behavior Prediction using Trajectory Setscited by 259
  4. 2021MultiPath++: Efficient Information Fusion and Trajectory Aggregation for Behavior Predictioncited by 246
  5. 2021Multimodal Motion Prediction with Stacked Transformerscited by 207
  6. 2021GOHOME: Graph-Oriented Heatmap Output for future Motion Estimationcited by 146
  7. 2021HOME: Heatmap Output for future Motion Estimationcited by 122
  8. 2021Multimodal Trajectory Prediction Conditioned on Lane-Graph Traversalscited by 112
  9. 2022HDGT: Heterogeneous Driving Graph Transformer for Multi-Agent Trajectory Prediction via Scene Encodingcited by 109
  10. 2020Self-supervised Learning on Graphs: Deep Insights and New Directioncited by 108
  11. 2021TPCN: Temporal Point Cloud Networks for Motion Forecastingcited by 94
  12. 2019Multi-Head Attention for Multi-Modal Joint Vehicle Motion Forecastingcited by 87
  13. 2018An LSTM Network for Highway Trajectory Predictioncited by 81
  14. 2020What-If Motion Prediction for Autonomous Drivingcited by 80
  15. 2021Multi-modal Motion Prediction with Transformer-based Neural Network for Autonomous Drivingcited by 73
  16. 2021Learning to Predict Vehicle Trajectories with Model-based Planningcited by 63
  17. 2021GRI: General Reinforced Imitation and its Application to Vision-Based Autonomous Drivingcited by 59
  18. 2022SSL-Lanes: Self-Supervised Learning for Motion Forecasting in Autonomous Drivingcited by 31
  19. 2020Map-Adaptive Goal-Based Trajectory Predictioncited by 29
  20. 2022PreTraM: Self-Supervised Pre-training via Connecting Trajectory and Mapcited by 25
  21. 2020Trajectory Prediction for Autonomous Driving based on Multi-Head Attention with Joint Agent-Map Representationcited by 12

Built on by

Cited by year

Citing papers per year: 2023: 4; 2024: 2; 2025: 2.

8 papers.

  1. 2023Traj-MAE: Masked Autoencoders for Trajectory Predictioncited by 42
  2. 2023Pre-training on Synthetic Driving Data for Trajectory Predictioncited by 18
  3. 2025Motion Forecasting for Autonomous Vehicles: A Surveycited by 8
  4. 2024SmartPretrain: Model-Agnostic and Dataset-Agnostic Representation Learning for Motion Predictioncited by 7
  5. 2023RedMotion: Motion Prediction via Redundancy Reductioncited by 6
  6. 2023RMP: A Random Mask Pretrain Framework for Motion Predictioncited by 6
  7. 2025Trends in Motion Prediction Toward Deployable and Generalizable Autonomy: A Revisit and Perspectivescited by 5
  8. 2024Forecast-PEFT: Parameter-Efficient Fine-Tuning for Pre-trained Motion Forecasting Modelscited by 3