Research map / Probabilistic and scientific modeling / Causal discovery
CausalVAE: Disentangled Representation Learning via Neural Structural Causal Models
Check 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 causal structure to the supervised VAE | CausalVAE vs ConditionVAE · Pendulum | MIC: 93.8 ±3.3 -> 95.1 ±2.4 | Helped |
| Add causal structure to the supervised VAE | CausalVAE vs ConditionVAE · Pendulum | TIC: 80.5 ±1.4 -> 81.6 ±1.9 | Helped |
| Add causal structure to the supervised VAE | CausalVAE vs ConditionVAE · Flow | MIC: 75.5 ±2.3 -> 72.1 ±1.3 | Hurt |
| Add causal structure to the supervised VAE | CausalVAE vs ConditionVAE · Flow | TIC: 56.5 ±1.8 -> 56.4 ±1.6 | No clear effect |
| Add causal structure to the supervised VAE | CausalVAE vs ConditionVAE · CelebA(SMILE) | MIC: 78.8 ±10.9 -> 83.7 ±6.2 | Helped |
| Add causal structure to the supervised VAE | CausalVAE vs ConditionVAE · CelebA(SMILE) | TIC: 66.1 ±12.1 -> 71.6 ±7.2 | Helped |
| Add causal structure to the supervised VAE | CausalVAE vs ConditionVAE · CelebA(BEARD) | MIC: 89.8 ±6.2 -> 92.3 ±5.6 | Helped |
| Add causal structure to the supervised VAE | CausalVAE vs ConditionVAE · CelebA(BEARD) | TIC: 78.7 ±7.7 -> 83.3 ±8.6 | Helped |
| Add the Mask Layer and conditional prior p(z|u) (ablation) | CausalVAE vs CausalVAE-unsup · Pendulum | MIC: 21.2 ±1.4 -> 95.1 ±2.4 | Helped |
| Add the Mask Layer and conditional prior p(z|u) (ablation) | CausalVAE vs CausalVAE-unsup · Flow | MIC: 22.8 ±2.7 -> 72.1 ±1.3 | Helped |
| Add the Mask Layer and conditional prior p(z|u) (ablation) | CausalVAE vs CausalVAE-unsup · CelebA(SMILE) | MIC: 27.2 ±5.3 -> 83.7 ±6.2 | Helped |
| Add the Mask Layer and conditional prior p(z|u) (ablation) | CausalVAE vs CausalVAE-unsup · CelebA(BEARD) | MIC: 11.4 ±1.5 -> 92.3 ±5.6 | Helped |
Closest papers by meaning
- Disentangled Representation Learning in Non-Markovian Causal Systems (NeurIPS 2024 · similarity 0.91)
- Causal Representation Learning via Counterfactual Intervention (AAAI 2024 · similarity 0.90)
- $\alpha$TC-VAE: On the relationship between Disentanglement and Diversity (ICLR 2024 · similarity 0.90)
- Learning Linear Causal Representations from General Environments: Identifiability and Intrinsic Ambiguity (NeurIPS 2024 · similarity 0.89)
- Identifiability Guarantees for Causal Disentanglement from Purely Observational Data (NeurIPS 2024 · similarity 0.88)
- Graph-based Unsupervised Disentangled Representation Learning via Multimodal Large Language Models (NeurIPS 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.