Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
CURVE: Learning Causality-Inspired Invariant Representations for Robust Scene Understanding via Uncertainty-Guided Regularization
Yue Liang, Jiatong Du, Ziyi Yang, Yanjun Huang, Hong Chen
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 prototype-conditioned causal intervention (ablation) | CURVE · CARLA-SR to DeepAccident (zero-shot OOD) | MCC: 25.77 -> 35.07 | Improved |
| add prototype-conditioned causal intervention (ablation) | CURVE · CARLA-SR test (ID) | MCC: 49.93 -> 68.85 | Improved |
| add prototype-conditioned causal intervention (ablation) | CURVE · CARLA-SR test (ID) | accuracy: 89.94 -> 92.92 | Improved |
| add prototype-conditioned causal intervention (ablation) | CURVE · CARLA-SR to DeepAccident (zero-shot OOD) | accuracy: 63.74 -> 68.13 | Improved |
| add uncertainty-guided pruning (ablation) | CURVE · CARLA-SR test (ID) | average node degree: 76.61 -> 5.78 | Improved |
| add uncertainty-guided pruning (ablation) | CURVE · CARLA-SR to DeepAccident (zero-shot OOD) | MCC: 30.46 -> 35.07 | Improved |
| add uncertainty-guided pruning (ablation) | CURVE · CARLA-SR test (ID) | MCC: 49.92 -> 68.85 | Improved |
| raise RS2G edge confidence threshold to 0.6 (sparser graph) (ablation) | RS2G · CARLA-SR to DeepAccident (zero-shot OOD) | accuracy: 64.84 -> 57.14 | Hurt |
| raise RS2G edge confidence threshold to 0.6 (sparser graph) (ablation) | RS2G · CARLA-SR test (ID) | MCC: 67.61 -> 82.01 | Improved |
| replace RS2G topology and relation learning with CURVE | RS2G · CARLA-SR to DeepAccident (zero-shot OOD) | accuracy: 64.84 -> 68.13 | Improved |
| replace RS2G topology and relation learning with CURVE | RS2G · CARLA-SR to DeepAccident (zero-shot OOD) | MCC: 36.44 -> 35.07 | No clear effect |
| replace RS2G topology and relation learning with CURVE | RS2G · CARLA-SR test (ID) | average node degree: 75.92 -> 5.78 | Improved |
Closest papers by meaning
- SPARTAN: A Sparse Transformer World Model Attending to What Matters (NeurIPS 2025 · similarity 0.86)
- Conformal Prediction and MLLM aided Uncertainty Quantification in Scene Graph Generation (CVPR 2025 · similarity 0.85)
- DiffusionSfM: Predicting Structure and Motion via Ray Origin and Endpoint Diffusion (CVPR 2025 · similarity 0.84)
- Robust Inverse Graphics via Probabilistic Inference (ICML 2024 · similarity 0.84)
- Generative Perception of Shape and Material from Differential Motion (NeurIPS 2025 · similarity 0.84)
- Variational Multi-scale Representation for Estimating Uncertainty in 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.84)
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.
