Research map / Probabilistic and scientific modeling / Causal discovery
Deep Counterfactual Estimation with Categorical Background Variables
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 |
|---|---|---|---|
| model hidden categorical response groups with CFQP | counterfactual estimation · Harmonic Oscillator, additive noise | test counterfactual MSE: 0.082 \pm 0.023 -> 0.013 \pm 0.001 | Helped |
| model hidden categorical response groups with CFQP | counterfactual estimation · colored-MNIST, additive noise | test counterfactual MSE: 0.008 \pm 0.004 -> 0.001 \pm 0.001 | Helped |
| model hidden categorical response groups with CFQP | counterfactual estimation · Cardiovascular, additive noise | test counterfactual MSE: 0.206\pm0.036 -> 0.077 \pm 0.050 | Helped |
| model hidden categorical response groups with CFQP | counterfactual estimation · Harmonic Oscillator, non-additive noise | test counterfactual MSE: 0.106\pm0.038 -> 0.009\pm0.001 | Helped |
| model hidden categorical response groups with CFQP | counterfactual estimation · colored-MNIST, non-additive noise | test counterfactual MSE: 0.009\pm0.002 -> 0.002\pm0.001 | Helped |
| model hidden categorical response groups with CFQP | counterfactual estimation · Cardiovascular, non-additive noise | test counterfactual MSE: 0.311\pm0.073 -> 0.188\pm0.114 | Helped |
| use CFQP rather than Deep-SCM for image counterfactuals | colored-MNIST · colored-MNIST, additive noise | SSIM: 0.857 \pm 0.003 -> 0.964 \pm 0.003 | Helped |
| use CFQP rather than Deep-SCM for image counterfactuals | colored-MNIST · colored-MNIST, non-additive noise | SSIM: 0.856 \pm 0.004 -> 0.959 \pm 0.005 | Helped |
| use CFQP rather than SC | Cardiovascular fluids · Cardiovascular | PEHE: 0.258\pm0.016 -> 0.143\pm0.108 | Helped |
| set the cluster count to the true K=3 (ablation) | CFQP · Harmonic Oscillator, additive noise | test counterfactual MSE: 0.08058\pm0.01735 -> 0.01332 \pm 0.00123 | Helped |
| replace K-means initialization with GMM (ablation) | CFQP · Cardiovascular, additive noise | test counterfactual MSE: 0.077 \pm 0.050 -> 0.067\pm0.042 | Mixed |
| replace K-means initialization with GMM (ablation) | CFQP · Cardiovascular, non-additive noise | test counterfactual MSE: 0.188\pm0.114 -> 0.192\pm0.0834 | Mixed |
Closest papers by meaning
- Towards Characterizing Domain Counterfactuals for Invertible Latent Causal Models (ICLR 2024 · similarity 0.92)
- Learning Counterfactual Outcomes Under Rank Preservation (NeurIPS 2025 · similarity 0.90)
- Counterfactual Generative Modeling with Variational Causal Inference (ICLR 2025 · similarity 0.88)
- An Empirical Examination of Balancing Strategy for Counterfactual Estimation on Time Series (ICML 2024 · similarity 0.86)
- Counterfactual Contrastive Learning with Normalizing Flows for Robust Treatment Effect Estimation (ICML 2025 · similarity 0.86)
- Exogenous Matching: Learning Good Proposals for Tractable Counterfactual Estimation (NeurIPS 2024 · similarity 0.86)
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.