2026-09-03 04:00 UTCOriginal source2 min readUpdated: 2026-09-03 09:21 UTC

DiDrive: A Risk-Aware Hierarchical Diffusion Framework for Safe Offline Reinforcement Learning in Autonomous Driving

DiDrive is a distribution-guided offline diffusion framework combining the Risk-Aware Hierarchical Diffusion (RHDif) architecture and the 3DICE policy optimization paradigm to address distribution shift, heavy-tailed risk, out-of-distribution actions, and state redundancy. In CARLA benchmark tests, DiDrive outperforms IQL, CQL, and Diffusion-QL, reaching an 85% success rate and a 4295.68 average reward in dense 60-vehicle traffic scenarios.

SourcearXiv Machine LearningAuthor: Qisong Guo, Jingtang Chen, Zhilin Chen, Pei Xu, Mingjian Fu, Wenxi Liu, Yuanlong Yu

Article intelligence

EngineersAdvanced

Key points

Offline RL policies built on diffusion models for autonomous driving are still vulnerable to distribution shift, heavy-tailed risk, OOD action generation, and high-dimensional state redundancy.
DiDrive uses RHDif's risk-gated encoder and contextual modulator to filter environmental noise and highlight safety-critical states, while 3DICE suppresses OOD overestimation and gradient oscillation.
In CARLA's dense 60-vehicle traffic scenarios, DiDrive achieves 85% success and 4295.68 average reward, surpassing IQL, CQL, and Diffusion-QL baselines.
The 16-page paper by Qisong Guo and five co-authors was submitted to arXiv on June 4, 2026, as arXiv:2609.01609v1.

Why it matters

This matters because offline RL policies built on diffusion models for autonomous driving are still vulnerable to distribution shift, heavy-tailed risk, OOD action generation, and high-dimensional state redundancy.

Technical impact

May affect model selection, inference cost, product capability, and evaluation benchmarks.

This panel is AI-generated and reviewed for accuracy.

-->

[Submitted on 4 Jun 2026]

Title:DiDrive: A Risk-Aware Hierarchical Diffusion Framework for Safe Offline Reinforcement Learning in Autonomous Driving

View a PDF of the paper titled DiDrive: A Risk-Aware Hierarchical Diffusion Framework for Safe Offline Reinforcement Learning in Autonomous Driving, by Qisong Guo and 5 other authors

View PDF HTML (experimental)

Abstract:While diffusion models effectively capture multimodal behavioral priors for autonomous driving, offline reinforcement learning (RL) policies remain susceptible to distribution shift, heavy-tailed risk signals, out-of-distribution (OOD) action generation, and high-dimensional state redundancy. To address these challenges, we propose DiDrive, a distribution-guided offline diffusion framework featuring two synergistic components: the Risk-Aware Hierarchical Diffusion (RHDif) architecture and the 3DICE policy optimization paradigm. In the state space, RHDif utilizes a low-level risk-gated encoder and a high-level contextual modulator to filter environmental redundancy and focus on safety-critical threats. In the action space, 3DICE mitigates OOD overestimation and gradient oscillation through in-sample calibrated guidance, spatiotemporal optimization, and ensemble-based candidate ranking. Evaluations on the CARLA benchmark demonstrate DiDrive's superiority over baselines like IQL, CQL, and Diffusion-QL, particularly in complex, high-density traffic scenarios with 60 vehicles, where it achieves an 85% success rate and a 4295.68 average reward, providing a robust pathway for safe autonomous driving decision-making.

Comments: 16 pages

Subjects:

Machine Learning (cs.LG); Robotics (cs.RO)

Cite as: arXiv:2609.01609 [cs.LG]

(or arXiv:2609.01609v1 [cs.LG] for this version)

https://doi.org/10.48550/arXiv.2609.01609

arXiv-issued DOI via DataCite

Submission history

From: Mingjian Fu [view email] [v1] Thu, 4 Jun 2026 04:04:39 UTC (4,148 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled DiDrive: A Risk-Aware Hierarchical Diffusion Framework for Safe Offline Reinforcement Learning in Autonomous Driving, by Qisong Guo and 5 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.LG

new | recent | 2026-09

Change to browse by:

cs cs.RO

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)