mlsanityData to 7 Oct 2026

Adaptive Prediction Ensemble: Improving Out-of-Distribution Generalization of Motion Forecasting

arXiv 2407.09475 ↗First posted 12 Jul 2024cs.RO cs.AI cs.CVcited by 7

Not on the map. Why?

Abstract (arXiv)

Deep learning-based trajectory prediction models for autonomous driving often struggle with generalization to out-of-distribution (OOD) scenarios, sometimes performing worse than simple rule-based models. To address this limitation, we propose a novel framework, Adaptive Prediction Ensemble (APE), which integrates deep learning and rule-based prediction experts. A learned routing function, trained concurrently with the deep learning model, dynamically selects the most reliable prediction based on the input scenario. Our experiments on large-scale datasets, including Waymo Open Motion Dataset (WOMD) and Argoverse, demonstrate improvement in zero-shot generalization across datasets. We show that our method outperforms individual prediction models and other variants, particularly in long-horizon prediction and scenarios with a high proportion of OOD data. This work highlights the potential of hybrid approaches for robust and generalizable motion prediction in autonomous driving. More details can be found on the project page: https://sites.google.com/view/ape-generalization.

Key facts. Cited by 7 papers in the map (data through 7 Oct 2026). Builds on 40 of its 51 references that are in the map. First posted on arXiv 12 Jul 2024.

Builds on

40 of 51 references.

References · 40

  1. 2022Training language models to follow instructions with human feedbackcited by 15,712
  2. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  3. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
  4. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,634
  5. 2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 3,413
  6. 2017Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real Worldcited by 1,401
  7. 2022Planning-oriented Autonomous Drivingcited by 1,068
  8. 2019Argoverse: 3D Tracking and Forecasting with Rich Mapscited by 728
  9. 2022Fine-Tuning can Distort Pretrained Features and Underperform Out-of-Distributioncited by 722
  10. 2021Large Scale Interactive Motion Forecasting for Autonomous Driving : The Waymo Open Motion Datasetcited by 532
  11. 2020VectorNet: Encoding HD Maps and Agent Dynamics from Vectorized Representationcited by 528
  12. 2022Wayformer: Motion Forecasting via Simple & Efficient Attention Networkscited by 279
28 more
  1. 2022Motion Transformer with Global Intention Localization and Local Movement Refinementcited by 267
  2. 2018Selective Experience Replay for Lifelong Learningcited by 239
  3. 2022MegaBlocks: Efficient Sparse Training with Mixture-of-Expertscited by 204
  4. 2022Exact Feature Distribution Matching for Arbitrary Style Transfer and Domain Generalizationcited by 122
  5. 2021The Devil is in the Detail: Simple Tricks Improve Systematic Generalization of Transformerscited by 117
  6. 2022Agreement-on-the-Line: Predicting the Performance of Neural Networks under Distribution Shiftcited by 94
  7. 2024Transformers Can Achieve Length Generalization But Not Robustlycited by 82
  8. 2022Differentiable Integrated Motion Prediction and Planning with Learnable Cost Function for Autonomous Drivingcited by 71
  9. 2021Adaptive Methods for Real-World Domain Generalizationcited by 66
  10. 2023Mixture-of-Experts Meets Instruction Tuning:A Winning Combination for Large Language Modelscited by 65
  11. 2024UniTraj: A Unified Framework for Scalable Vehicle Trajectory Predictioncited by 51
  12. 2022Hierarchical Planning Through Goal-Conditioned Offline Reinforcement Learningcited by 48
  13. 2022Compound Domain Generalization via Meta-Knowledge Encodingcited by 40
  14. 2021Vehicle trajectory prediction works, but not everywherecited by 39
  15. 2021Spatio-Temporal Graph Dual-Attention Network for Multi-Agent Prediction and Trackingcited by 35
  16. 2024CMP: Cooperative Motion Prediction with Multi-Agent Communicationcited by 25
  17. 2021Dealing with the Unknown: Pessimistic Offline Reinforcement Learningcited by 18
  18. 2022Dynamics-Aware Spatiotemporal Occupancy Prediction in Urban Environmentscited by 17
  19. 2023Guided Online Distillation: Promoting Safe Reinforcement Learning by Offline Demonstrationcited by 14
  20. 2023Scene Informer: Anchor-based Occlusion Inference and Trajectory Prediction in Partially Observable Environmentscited by 14
  21. 2023Pedestrian Crossing Action Recognition and Trajectory Prediction with 3D Human Keypointscited by 12
  22. 2023Interactive Autonomous Navigation with Internal State Inference and Interactivity Estimationcited by 10
  23. 2023DE-TGN: Uncertainty-Aware Human Motion Forecasting using Deep Ensemblescited by 6
  24. 2024Quantifying Uncertainty in Motion Prediction with Variational Bayesian Mixturecited by 6
  25. 2021Interaction-Aware Behavior Planning for Autonomous Vehicles Validated with Real Traffic Datacited by 4
  26. 2023Failure Detection for Motion Prediction of Autonomous Driving: An Uncertainty Perspectivecited by 4
  27. 2023Multi-Predictor Fusion: Combining Learning-based and Rule-based Trajectory Predictorscited by 3
  28. 2024RuleFuser: An Evidential Bayes Approach for Rule Injection in Imitation Learned Planners and Predictors for Robustness under Distribution Shiftscited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 2; 2025: 3; 2026: 2.

7 papers.

  1. 2025UniOcc: A Unified Benchmark for Occupancy Forecasting and Prediction in Autonomous Drivingcited by 27
  2. 2024Generating Out-Of-Distribution Scenarios Using Language Modelscited by 10
  3. 2025TrajEvo: Trajectory Prediction Heuristics Design via LLM-driven Evolutioncited by 5
  4. 2025Towards Generalizable Safety in Crowd Navigation via Conformal Uncertainty Handlingcited by 5
  5. 2026Drive My Way: Preference Alignment of Vision-Language-Action Model for Personalized Drivingcited by 4
  6. 2024Importance Sampling-Guided Meta-Training for Intelligent Agents in Highly Interactive Environmentscited by 3
  7. 2026A Generalizable Physics-guided Causal Model for Trajectory Prediction in Autonomous Drivingcited by 1