Research map / Multi-agent systems and games / Human-AI and goal-conditioned
CausalDrive: Real-time Causal World Models for Autonomous Driving
Tianyi Yan, Huan Zheng, Dubing Chen, Meizhi Qu, Yingying Shen, Lijun Zhou, Mingfei Tu, Bing Wang, Guang Chen, Hangjun Ye, Haiyang Sun, Cheng-zhong Xu and 1 more
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add Context-Forced DMD distillation to the AR teacher (CausalDrive+ to CausalDrive) | CausalDrive · nuPlan | FPS: 0.5 -> 12.4 | Improved |
| distill the AR teacher into a few-step student with Context-Forced DMD | CausalDrive · nuPlan | FVD: 113.6 -> 121.6 | Mixed |
| add sociology text prompt (Polite) to action conditioning | CausalDrive vs Vista · SocioDrive-Bench | YCR: 12.5% -> 82.0% | Improved |
| add sociology prompt conditioning, lowering ghost collisions | CausalDrive vs Vista · SocioDrive-Bench | FCR: 87.5% -> 18.0% | Improved |
| replace log-replay with CausalDrive reactive simulation | CausalDrive vs Log-Replay(GT) · SocioDrive-Bench | FCR: 100.0% -> 18.0% | Improved |
| use CausalDrive instead of DrivingSphere as closed-loop simulator | UniAD planner · Closed-loop PDM-Closed | PDMS: 0.7281 -> 0.7792 | Improved |
| use CausalDrive instead of DriveArena as closed-loop simulator | UniAD planner · Closed-loop PDM-Closed | PDMS: 0.6901 -> 0.7792 | Improved |
| RL post-train DiffusionDrive inside CausalDrive | DiffusionDrive · Navsim | PDMS: 88.1 -> 90.7 | Improved |
| RL post-train in CausalDrive versus DiffusionDrive-v2 | DiffusionDrive · Navsim | PDMS: 90.3 -> 90.7 | Improved |
| use CausalDrive over Orbis for action following | CausalDrive · nuPlan | ADE precision: 0.39 -> 0.45 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DriveArena — DriveArena: A Closed-loop Generative Simulation Platform for Autonomous Driving (ICCV 2025) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Prioritizing Perception-Guided Self-Supervision: A New Paradigm for Causal Modeling in End-to-End Autonomous Driving (NeurIPS 2025 · similarity 0.90)
- ChainFlow-VLA: Causal Flow Planning with Vision-Language Models (NeurIPS 2026 · similarity 0.90)
- ReSim: Reliable World Simulation for Autonomous Driving (NeurIPS 2025 · similarity 0.89)
- Vista: A Generalizable Driving World Model with High Fidelity and Versatile Controllability (NeurIPS 2024 · similarity 0.88)
- Generating Multimodal Driving Scenes via Next-Scene Prediction (CVPR 2025 · similarity 0.88)
- Beyond Self-Play and Scale: A Behavior Benchmark for Generalization in Autonomous Driving (NeurIPS 2026 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
