mlsanityData to 7 Oct 2026

LoRD: Adapting Differentiable Driving Policies to Distribution Shifts

arXiv 2410.09681 ↗First posted 13 Oct 2024cs.RO cs.AI cs.LGcited by 4

On the map: End-to-end autonomous driving → End-to-end system architecture

Abstract (arXiv)

Distribution shifts between operational domains can severely affect the performance of learned models in self-driving vehicles (SDVs). While this is a well-established problem, prior work has mostly explored naive solutions such as fine-tuning, focusing on the motion prediction task. In this work, we explore novel adaptation strategies for differentiable autonomy stacks consisting of prediction, planning, and control, perform evaluation in closed-loop, and investigate the often-overlooked issue of catastrophic forgetting. Specifically, we introduce two simple yet effective techniques: a low-rank residual decoder (LoRD) and multi-task fine-tuning. Through experiments across three models conducted on two real-world autonomous driving datasets (nuPlan, exiD), we demonstrate the effectiveness of our methods and highlight a significant performance gap between open-loop and closed-loop evaluation in prior approaches. Our approach improves forgetting by up to 23.33% and the closed-loop OOD driving score by 9.93% in comparison to standard fine-tuning.

Key facts. Cited by 4 papers in the map (data through 7 Oct 2026). Placed in End-to-end system architecture, one of 9 directions in End-to-end autonomous driving. Builds on 49 of its 60 references that are in the map. First posted on arXiv 13 Oct 2024.

Builds on

49 of 60 references.

Closest ancestors · 5

Same direction, newest first

  1. 2024QuAD: Query-based Interpretable Neural Motion Planning for Autonomous Driving1 Apr
  2. 2024Towards learning-based planning:The nuPlan benchmark for real-world autonomous driving7 Mar
  3. 2023Rethinking the Open-Loop Evaluation of End-to-End Autonomous Driving in nuScenes17 May
  4. 2022DiffStack: A Differentiable and Modular Control Stack for Autonomous Vehicles13 Dec
  5. 2021End-to-end Interpretable Neural Motion Planner17 Jan

Other references · 44

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
  3. 2022Planning-oriented Autonomous Drivingcited by 1,068
  4. 2016End to End Learning for Self-Driving Carscited by 1,000
  5. 2023Argoverse 2: Next Generation Datasets for Self-Driving Perception and Forecastingcited by 701
  6. 2017OptNet: Differentiable Optimization as a Layer in Neural Networkscited by 588
  7. 2021Implicit Behavioral Cloningcited by 442
  8. 2019MultiPath: Multiple Probabilistic Anchor Trajectory Hypotheses for Behavior Predictioncited by 417
  9. 2019Causal Confusion in Imitation Learningcited by 286
  10. 2021Parameter-efficient Multi-task Fine-tuning for Transformers via Shared Hypernetworkscited by 278
  11. 2018Learning to Drive in a Daycited by 266
  12. 2023Parting with Misconceptions about Learning-based Vehicle Motion Planningcited by 238
32 more
  1. 2021MP3: A Unified Model to Map, Perceive, Predict and Plancited by 211
  2. 2018Differentiable MPC for End-to-end Planning and Controlcited by 178
  3. 2018Efficient parametrization of multi-domain deep neural networkscited by 168
  4. 2016A Connection between Generative Adversarial Networks, Inverse Reinforcement Learning, and Energy-Based Modelscited by 167
  5. 2018Residual Policy Learningcited by 146
  6. 2022Imitation Is Not Enough: Robustifying Imitation with Reinforcement Learning for Challenging Driving Scenarioscited by 124
  7. 2019Urban Driving with Conditional Imitation Learningcited by 84
  8. 2021Unsupervised Learning of Compositional Energy Conceptscited by 71
  9. 2022Differentiable Integrated Motion Prediction and Planning with Learnable Cost Function for Autonomous Drivingcited by 71
  10. 2022GoRela: Go Relative for Viewpoint-Invariant Motion Forecastingcited by 68
  11. 2022Symphony: Learning Realistic and Diverse Agents for Autonomous Driving Simulationcited by 61
  12. 2023LLM-Assist: Enhancing Closed-Loop Planning with Language-Based Reasoningcited by 59
  13. 2024UniTraj: A Unified Framework for Scalable Vehicle Trajectory Predictioncited by 51
  14. 2023DTPP: Differentiable Joint Conditional Prediction and Cost Evaluation for Tree Policy Planning in Autonomous Drivingcited by 46
  15. 2021Towards Robust and Adaptive Motion Forecasting: A Causal Representation Perspectivecited by 40
  16. 2024Can Vehicle Motion Planning Generalize to Realistic Long-tail Scenarios?cited by 38
  17. 2021Differentiable SLAM-net: Learning Particle SLAM for Visual Navigationcited by 34
  18. 2023Rethinking Imitation-based Planner for Autonomous Drivingcited by 25
  19. 2019Differentiable Algorithm Networks for Composable Robot Learningcited by 24
  20. 2021You Mostly Walk Alone: Analyzing Feature Attribution in Trajectory Predictioncited by 23
  21. 2021UMBRELLA: Uncertainty-Aware Model-Based Offline Reinforcement Learning Leveraging Planningcited by 23
  22. 2022Expanding the Deployment Envelope of Behavior Prediction via Adaptive Meta-Learningcited by 23
  23. 2022Interpretable Self-Aware Neural Networks for Robust Trajectory Predictioncited by 22
  24. 2021Injecting Knowledge in Data-driven Vehicle Trajectory Predictorscited by 18
  25. 2021Hyperparameter Selection for Imitation Learningcited by 17
  26. 2022Embedding Synthetic Off-Policy Experience for Autonomous Driving via Zero-Shot Curriculacited by 17
  27. 2022Motion Style Transfer: Modular Low-Rank Adaptation for Deep Motion Forecastingcited by 12
  28. 2023Energy-based Potential Games for Joint Motion Forecasting and Controlcited by 9
  29. 2022Differentiable Constrained Imitation Learning for Robot Motion Planning and Controlcited by 7
  30. 2023Inferring Relational Potentials in Interacting Systemscited by 5
  31. 2024Transfer Learning Study of Motion Transformer-based Trajectory Predictionscited by 5
  32. 2023On a Connection between Differential Games, Optimal Control, and Energy-based Models for Multi-Agent Interactionscited by 4

Built on by

Cited by year

Citing papers per year: 2025: 4.

4 papers.

  1. 2025Reinforced Refinement with Self-Aware Expansion for End-to-End Autonomous Drivingcited by 19
  2. 2025Trends in Motion Prediction Toward Deployable and Generalizable Autonomy: A Revisit and Perspectivescited by 5
  3. 2025A Concept for Efficient Scalability of Automated Driving Allowing for Technical, Legal, Cultural, and Ethical Differencescited by 0
  4. 2025Vehicle Dynamics Embedded World Models for Autonomous Drivingcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026The Blind Spot of Adaptation: Quantifying and Mitigating Forgetting in Fine-tuned Driving Models
  2. 2025RoaD: Rollouts as Demonstrations for Closed-Loop Supervised Fine-Tuning of Autonomous Driving Policies
  3. 2026BridgeSim: Unveiling the OL-CL Gap in End-to-End Autonomous Driving
  4. 2023Combating the effects of speed and delays in end-to-end self-driving
  5. 2025Model-Based Policy Adaptation for Closed-Loop End-to-End Autonomous Driving
  6. 2026Post-Training in End-to-End Autonomous Driving
  7. 2025Closing the Loop: Motion Prediction Models beyond Open-Loop Benchmarks
  8. 2025AdaDrive: Self-Adaptive Slow-Fast System for Language-Grounded Autonomous Driving