Research map / Probabilistic and scientific modeling / Molecules and proteins
Adaptive Latent Feature Sharing for Piecewise Linear Dimensionality Reduction
Adam Farooq, Yordan P. Raykov, Petar Raykov, Max A. Little
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 |
|---|---|---|---|
| increase assumed number of features K from 2 to 3K | aFA · Synthetic Sparse (sigma2=4) | mean squared out-of-sample reconstruction error: 4.20 -> 0.25 | Improved |
| increase assumed number of features K from 2 to 3K | aFA · Synthetic Sparse (sigma2=0.01) | mean squared out-of-sample reconstruction error: 0.15 -> 0.00 | Improved |
| increase assumed number of factors from 2 to 3K | FA · Synthetic Sparse (sigma2=4) | mean squared out-of-sample reconstruction error: 4.06 -> 1.07 | Improved |
| use aFA instead of fsFA at K=3K | aFA vs fsFA · Synthetic Sparse (sigma2=4) | mean squared out-of-sample reconstruction error: 4.01 -> 0.25 | Improved |
| use aFA instead of conventional FA at K=3K | aFA vs FA · Synthetic Sparse (sigma2=4) | mean squared out-of-sample reconstruction error: 1.07 -> 0.25 | Improved |
| use aFA instead of isFA at K=3K | aFA vs isFA · Synthetic Dense (sigma2=4) | mean squared out-of-sample reconstruction error: 5.10 -> 0.24 | Improved |
| increase K from 2K to 3K | aFA · Synthetic Sparse (sigma2=4) | mean squared out-of-sample reconstruction error: 2.07 -> 0.25 | Improved |
| increase K from 2 to 3K at low noise | FA · Synthetic Sparse (sigma2=0.01) | mean squared out-of-sample reconstruction error: 0.96 -> 0.01 | No clear effect |
| increase K from 2 to 3K | aFA · Synthetic data sparsity (K*=10) | sparsity (portion of zeros): 0.60 -> 0.87 | Improved |
| use sparse latent feature aPPCA instead of global PCA for 2D projection | aPPCA · MNIST 10,000 digits (VAE 10-D) | qualitative cluster separation | Improved |
Closest papers by meaning
- Additive Component Analysis (CVPR 2017 · similarity 0.88)
- Diverse Influence Component Analysis: A Geometric Approach to Nonlinear Mixture Identifiability (NeurIPS 2025 · similarity 0.85)
- Efficient Sparse PCA via Block-Diagonalization (ICLR 2025 · similarity 0.85)
- Probabilistic Geometric Principal Component Analysis with application to neural data (ICLR 2025 · similarity 0.85)
- Unconstrained Stochastic CCA: Unifying Multiview and Self-Supervised Learning (ICLR 2024 · similarity 0.85)
- Capturing the denoising effect of PCA via compression ratio (NeurIPS 2024 · similarity 0.85)
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.
