Research map / Probabilistic and scientific modeling / Causal discovery
Identification of Nonlinear Latent Hierarchical Models
Lingjing Kong, Biwei Huang, Feng Xie, Eric Xing, Yuejie Chi, Kun Zhang
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 |
|---|---|---|---|
| assume joint invertibility of g1,g2 instead of individual invertibility | basis model estimator · d_z=d_s1=d_s2=2 | R^2 for z: 0.67 -> 0.93 | Improved |
| assume joint invertibility instead of individual invertibility | basis model estimator · d_z=d_s1=d_s2=4 | R^2 for z: 0.67 -> 0.89 | Improved |
| allow different dimensionalities of s1 and s2 (drop g1=g2 assumption) | basis model estimator · d_s1=2, d_s2=3 | R^2 for z: NA -> 0.90 | Improved |
| increase latent dimensionality from 2 to 4 with different dims for s2 (d_s2=6) | joint invertibility estimator · basis model | R^2 for z: 0.93 -> 0.83 | Mixed |
| treat v1 = x3 (shared parent of z2 and z3) in V-structure model | hierarchical identification algorithm · synthetic V-structure | pairwise R^2: 0.53 -> 0.90 | Improved |
| predict within sibling pairs vs non-sibling pairs | hierarchical identification algorithm · synthetic V-structure | pairwise R^2: 0.57 -> 0.86 | Improved |
| asymmetric prediction direction (x3 predicts others vs others predict x3) | hierarchical identification algorithm · synthetic V-structure | pairwise R^2: 0.52 -> 0.88 | Improved |
| run algorithm on nonlinear personality data | hierarchical identification algorithm · personality dataset | structure recovery | Improved |
| run algorithm on high-dimensional multi-view digit data | hierarchical identification algorithm · multi-view MNIST | structure recovery | Improved |
Closest papers by meaning
- Identification of Latent Confounders via Investigating the Tensor Ranks of the Nonlinear Observations (ICML 2025 · similarity 0.93)
- Differentiable Causal Discovery for Latent Hierarchical Causal Models (ICLR 2025 · similarity 0.93)
- Identifiable Latent Polynomial Causal Models through the Lens of Change (ICLR 2024 · similarity 0.92)
- Local Identifying Causal Relations in the Presence of Latent Variables (ICML 2025 · similarity 0.91)
- Local Causal Structure Learning in the Presence of Latent Variables (ICML 2024 · similarity 0.91)
- Identification of Causal Structure with Latent Variables Based on Higher Order Cumulants (AAAI 2024 · similarity 0.90)
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.
