FlowR2A: Learning Reward-to-Action Distribution for Multimodal Driving Planning

NeurIPS 2026

News

• [2026-10] We released the reward simulation pipeline (guide) and interactive visualization tools (guide).
• [2026-10] FlowR2A has been accepted to NeurIPS 2026.

Overview

FlowR2A is a multimodal driving planner that learns the reward-conditioned action distribution p(a | r) with flow matching. Instead of treating simulation rewards as discriminative targets, FlowR2A treats them as a condition, unifying the dense supervision of scoring-based methods with the generative proposal modeling of anchor-based methods. At inference, generation is steered toward high-reward trajectories via classifier-free guidance.

Training pipeline of FlowR2A.

Table of Contents

• News
• Overview
• Table of Contents
• Qualitative Results on NAVSIM
• Getting Started
• To-Do
• Acknowledgement
• Citation
• License

Qualitative Results on NAVSIM

Each row is one planner; trajectory proposals are colored by PDM score, from red (0) to green (1). FlowR2A (bottom row) produces proposals that are both diverse and consistently high-scoring.

More qualitative comparisons are available on the project page.

Getting Started

• Data and Environment Preparation — install dependencies and download the data/models needed for evaluation.
• Evaluation — download the pre-trained checkpoints and run evaluation on the test set.
• Training — download the training data and prepare the cache, then train the model in two stages.
• Reward Simulation — (optional) generate the reward cache from scratch instead of downloading it.
• Visualization — interactively inspect model outputs on navtest and the navtrain reward cache in the browser.

To-Do

• Inference and evaluation code
• Training code
• Reward simulation / caching pipeline

Acknowledgement

FlowR2A is built upon the following outstanding open-source contributions: NAVSIM, DiffusionDrive.

Citation

If you find FlowR2A useful in your research or applications, please consider giving us a star 🌟 and citing it with the following BibTeX entry.

@inproceedings{flowr2a2026,
  title     = {FlowR2A: Learning Reward-to-Action Distribution for Multimodal Driving Planning},
  author    = {Li, Xirui and Liu, Zhe and Ye, Xiaoqing and Han, Wenhua and Pan, Yifeng and Han, Junyu and Zhao, Hengshuang},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year      = {2026}
}

License

MIT — see LICENSE.