Research map / Label-efficient and robust learning / Recognition and long-tail
Predict and Resist: Long-Term Accident Anticipation Under Sensor Noise
Xingcheng Liu, Bin Rao, Yanchen Guan, Chengyue Wang, Haicheng Liao, Jiaxun Zhang, Chengyu Lin, Meixin Zhu, Zhenning Li
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 anticipation loss (ablation) | Ours · CCD | AP (%): 33.3 -> 99.8 | Improved |
| add value loss (ablation) | Ours · CCD | AP (%): 92.8 -> 99.8 | Improved |
| add value loss (ablation) | Ours · CCD | mTTA (s): 3.03 -> 4.29 | Improved |
| add self-adaptive object-aware module (ablation) | Ours · CCD | AP (%): 99.3 -> 99.8 AP; 4.61 -> 4.29 mTTA | Mixed |
| add time-weight layer (ablation) | Ours · CCD | AP (%): 99.5 -> 99.8 AP; 4.47 -> 4.29 mTTA | Mixed |
| add policy gradient loss (ablation) | Ours · CCD | AP (%): 99.6 -> 99.8 | No clear effect |
| add dual diffusion denoising (ablation) | Ours · CCD, Gaussian sigma=5 | AP (%): 99.0 -> 99.6 | Improved |
| add image diffusion under severe noise (ablation) | Ours · CCD, Gaussian sigma=20 | AP (%): 92.8 -> 91.6 | Hurt |
| increase early-reward weight from x1 to x50 (ablation) | Ours · A3D | AP (%): 95.7 -> 92.7 | Hurt |
| raise misprediction penalty x10 (ablation) | Ours · A3D | mTTA (s): 4.60 -> 4.92 mTTA; 95.7 -> 91.2 AP | Mixed |
| ours vs LATTE on CCD timeliness | Ours · CCD | mTTA (s): 4.53 -> 4.29 | Hurt |
| add Gaussian noise sigma=20 | Ours · CCD | AP (%): 99.8 -> 91.6 | Hurt |
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.
- LATTE — LATTE: Latent Type Modeling for Biomedical Entity Linking (AAAI 2020) · better on 3 of 4 shared results
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
- Progressive Risk Estimation for Accident Anticipation (NeurIPS 2026 · similarity 0.90)
- Accident Anticipation via Temporal Occurrence Prediction (NeurIPS 2025 · similarity 0.88)
- ReflexDiffusion: Reflection-Enhanced Trajectory Planning for High-lateral-acceleration Scenarios in Autonomous Driving (AAAI 2026 · similarity 0.88)
- RiskProp: Collision-Anchored Self-Supervised Risk Propagation For Early Accident Anticipation (CVPR 2026 · similarity 0.87)
- DiffScene: Diffusion-Based Safety-Critical Scenario Generation for Autonomous Vehicles (AAAI 2025 · similarity 0.85)
- Future-Aware End-to-End Driving: Bidirectional Modeling of Trajectory Planning and Scene Evolution (NeurIPS 2025 · similarity 0.85)
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.
