Research map / Probabilistic and scientific modeling / Molecules and proteins
ReflexDiffusion: Reflection-Enhanced Trajectory Planning for High-lateral-acceleration Scenarios in Autonomous Driving
Xuemei Yao, Xiao Yang, Jianbin Sun, Liuwei Xie, Xuebin Shao, Xiyu Fang, Hang Su, Kewei Yang
This paper tried a common change: Add dropout
add conditional dropout in training · Test14-hard (reactive) · driving score: 23.86 -> 65.53 Improved
What 795 accepted papers report for this change
- 549 improved
- 87 no clear effect
- 12 mixed
- 147 hurt
- Don't Drop Dropout: Optimizing Layer Sparsity for Efficient LLM Training and Inference ICML 2026 · apply layer dropout to whole transformer blocks instead of attention and FFN sub-blocks separately · 20 TPP val loss, 271M · val loss: 2.377 -> 2.367Improved
- Quantifying the Uncertainty of Foundation Models with Singular Value Ensembles ICML 2026 · use MC Dropout · vision and SST-2 · accuracy: 86.3 -> 76.6 on Flowers102Hurt
- Motion Planning in Compressed Representation Spaces ICML 2026 · add nested dropout during training · WOMD guided generation · goal FDE: 4.145 -> 1.104Improved
See this change in detailMatched by the wording of each paper's own result, read by AI from the full text. Papers report gains more often than failures, so this shows what gets published, not the chance a change works.
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 conditional dropout in training (ablation) | ReflexDiffusion · Test14-hard (reactive) | driving score: 23.86 -> 65.53 | Improved |
| add classifier-free guidance denoising (ablation) | ReflexDiffusion · Test14-hard (reactive) | driving score: 59.85 -> 65.53 | Improved |
| add reflection module (ablation) | ReflexDiffusion · Test14-hard (reactive) | driving score: 53.21 -> 65.53 | Improved |
| add ReflexDiffusion at inference to Diffusion Planner | Diffusion Planner · Test14-hard (reactive) | driving score: 57.41 -> 65.53 | Improved |
| add ReflexDiffusion to Diffusion-es | Diffusion-es · Test14-hard | driving score: 31.88 -> 39.04 | Improved |
| add ReflexDiffusion to Diffusion Planner | Diffusion Planner · Test14-random (reactive) | driving score: 82.88 -> 71.57 | Hurt |
| add ReflexDiffusion to Diffusion Planner | Diffusion Planner · Test14-hard (non-reactive) | driving score: 58.47 -> 59.94 | Improved |
| add conditional dropout alone to Diffusion Planner | Diffusion Planner · Test14-hard (reactive) | driving score: 57.41 -> 14.04 | Hurt |
| enable reflection (triggered rarely) | ReflexDiffusion · runtime | end-to-end latency (ms): 35.7 -> 36.1 | No clear effect |
Closest papers by meaning
- Diffusion-ES: Gradient-free Planning with Diffusion for Autonomous and Instruction-guided Driving (CVPR 2024 · similarity 0.88)
- C2F-TP: A Coarse-to-Fine Denoising Framework for Uncertainty-Aware Trajectory Prediction (AAAI 2025 · similarity 0.88)
- DiRecT: Safe Diffusion-Based Planning via Receding-Horizon Denoising (NeurIPS 2026 · similarity 0.87)
- AlignDrive: Aligned Lateral-Longitudinal Planning for End-to-End Autonomous Driving (NeurIPS 2026 · similarity 0.87)
- LaKD: Length-agnostic Knowledge Distillation for Trajectory Prediction with Any Length Observations (NeurIPS 2024 · similarity 0.87)
- Don't Shake the Wheel: Momentum-Aware Planning in End-to-End Autonomous Driving (CVPR 2025 · similarity 0.87)
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.
