Research map / Probabilistic and scientific modeling / Dynamical systems
Filtered-CoPhy: Unsupervised Learning of Counterfactual Physics in Pixel Space
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 appearance coefficients to the N-keypoint representation (ablation) | Filtered-CoPhy image reconstruction · BlocktowerCF | PSNR: 32.53 -> 34.40 | Helped |
| Add appearance coefficients to the N-keypoint representation (ablation) | Filtered-CoPhy image reconstruction · BallsCF | PSNR: 34.62 -> 37.76 | Helped |
| Add appearance coefficients to the N-keypoint representation (ablation) | Filtered-CoPhy image reconstruction · CollisionCF | PSNR: 30.65 -> 32.00 | Helped |
| Increase the keypoint count from N to 2N (ablation) | Filtered-CoPhy image reconstruction · BlocktowerCF | PSNR: 34.40 -> 35.41 | Helped |
| Increase the keypoint count from N to 2N (ablation) | Filtered-CoPhy image reconstruction · BallsCF | PSNR: 37.76 -> 37.06 | Hurt |
| Increase the keypoint count from N to 2N (ablation) | Filtered-CoPhy image reconstruction · CollisionCF | PSNR: 32.00 -> 35.41 | Helped |
| Apply the identifiability constraint when generating the benchmark (ablation) | BlocktowerCF mass estimation · BlocktowerCF | Accuracy: 56% -> 84% | Helped |
| Apply the identifiability constraint when generating the benchmark (ablation) | BlocktowerCF corrected mass estimation · BlocktowerCF | Corrected Acc.: 58% -> 91% | Helped |
| Increase trajectory sampling frequency from 5 FPS to 25 FPS (ablation) | BlocktowerCF trajectory forecaster · BlocktowerCF | MSE: 4.58 -> 3.74 | Helped |
| Increase trajectory sampling frequency from 25 FPS to 45 FPS (ablation) | BlocktowerCF trajectory forecaster · BlocktowerCF | MSE: 3.74 -> 3.93 | Hurt |
| Add appearance coefficients to the N-keypoint representation (ablation) | Filtered-CoPhy image reconstruction · BlocktowerCF | MSE Grad: 41.86 -> 27.24 | Helped |
| Add appearance coefficients to the N-keypoint representation (ablation) | Filtered-CoPhy image reconstruction · BallsCF | MSE Grad: 6.22 -> 3.47 | Helped |
Closest papers by meaning
- Latent Intuitive Physics: Learning to Transfer Hidden Physics from A 3D Video (ICLR 2024 · similarity 0.85)
- VideoREPA: Learning Physics for Video Generation through Relational Alignment with Foundation Models (NeurIPS 2025 · similarity 0.85)
- Rethinking Visual Counterfactual Explanations Through Region Constraint (ICLR 2025 · similarity 0.84)
- Causal Representation Learning via Counterfactual Intervention (AAAI 2024 · similarity 0.84)
- Learning Physics From Video: Unsupervised Physical Parameter Estimation for Continuous Dynamical Systems (CVPR 2025 · similarity 0.84)
- Learning What's Real: Disentangling Signals and Measurement Artifacts in Multi-Sensor Data, with Applications to Astrophysics (NeurIPS 2026 · 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.