Research map / Probabilistic and scientific modeling / Causal discovery
Weakly supervised causal representation learning
Johann Brehmer, Pim De Haan, Phillip Lippe, Taco Cohen
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 |
|---|---|---|---|
| model causal structure (ILCM) instead of assuming independent factors (dVAE) | weakly supervised VAE · 2D toy data | DCI disentanglement: 0.35 -> 0.99 | Improved |
| model causal structure (ILCM) instead of independent factors (dVAE) | weakly supervised VAE · Causal3DIdent | DCI disentanglement: 0.82 -> 0.99 | Improved |
| model causal structure (ILCM) instead of independent factors (dVAE) | weakly supervised VAE · CausalCircuit | DCI disentanglement: 0.34 -> 0.97 | Improved |
| ILCM representations versus dVAE representations for causal graph recovery | ILCM-E · CausalCircuit | SHD: 5.00 -> 0.00 | Improved |
| use ENCO versus heuristic algorithm for graph extraction (ablation) | ILCM · Causal3DIdent | SHD: 0.00 vs 0.17 | No clear effect |
| use implicit instead of explicit graph parameterization (ILCM vs ELCM) (ablation) | latent causal model · simple datasets | optimization robustness | Improved |
| make causal variables discrete instead of continuous (ablation) | ILCM · CausalCircuit variant | disentanglement | Hurt |
| increase number of causal variables beyond about 10 (ablation) | ILCM · synthetic linear SCMs | disentanglement and graph accuracy | Hurt |
Closest papers by meaning
- A Sparsity Principle for Partially Observable Causal Representation Learning (ICML 2024 · similarity 0.90)
- Learning Linear Causal Representations from General Environments: Identifiability and Intrinsic Ambiguity (NeurIPS 2024 · similarity 0.90)
- Causal Representation Learning Made Identifiable by Grouping of Observational Variables (ICML 2024 · similarity 0.90)
- Disentangled Representation Learning in Non-Markovian Causal Systems (NeurIPS 2024 · similarity 0.89)
- Object centric architectures enable efficient causal representation learning (ICLR 2024 · similarity 0.89)
- Causal Representation Learning from Multiple Distributions: A General Setting (ICML 2024 · similarity 0.88)
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.
