Research map / Probabilistic and scientific modeling / Variational and generative sampling
Unsupervised Learning with Truncated Gaussian Graphical Models
Qinliang Su, Xuejun Liao, Chunyuan Li, Zhe Gan, Lawrence Carin
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 |
|---|---|---|---|
| Add truncated hidden variables to the bipartite GGM (ablation) | RTGGM · MNIST test | test log-prob.: -90.2 -> -83.2 | Improved |
| Add truncated hidden variables to the bipartite GGM (ablation) | RTGGM · Caltech 101 Silhouettes test | test log-prob.: -350.9 -> -105.1 | Improved |
| Use RTGGM instead of RBM | 500-hidden-node models · MNIST test | test log-prob.: -86.3 -> -83.2 | Improved |
| Use RTGGM instead of RBM | 500-hidden-node models · Caltech 101 Silhouettes test | test log-prob.: -107.7 -> -105.1 | Improved |
| Use RTGGM instead of RSM | 50-topic count models · 20news test | perplexity per word: 953 -> 915 | Improved |
| Use RTGGM instead of RSM | 50-topic count models · Reuters test | perplexity per word: 988 -> 934 | Improved |
| Pretrain the ReLU network with deep RTGGM instead of rectified-RBM | 1000-1000-1000 ReLU neural network · MNIST | classification error: 1.33% -> 1.17% | Improved |
| Pretrain the ReLU network with deep RTGGM instead of rectified-RBM | 1000-1000-1000 ReLU neural network · NORB | classification error: 16.43% -> 16.12% | Improved |
| Pretrain the ReLU network with deep RTGGM instead of no pre-training | 1000-1000-1000 ReLU neural network · MNIST | classification error: 1.43% -> 1.17% | Improved |
| Pretrain the ReLU network with deep RTGGM instead of no pre-training | 1000-1000-1000 ReLU neural network · NORB | classification error: 16.88% -> 16.12% | Improved |
Closest papers by meaning
- Nonlinear Statistical Learning with Truncated Gaussian Graphical Models (ICML 2016 · similarity 0.89)
- Asymptotic Analysis of Two-Layer Neural Networks after One Gradient Step under Gaussian Mixtures Data with Structure (ICLR 2025 · similarity 0.85)
- Approximation and Generalization Abilities of Score-based Neural Network Generative Models for Sub-Gaussian Distributions (NeurIPS 2025 · similarity 0.84)
- Feature Contamination: Neural Networks Learn Uncorrelated Features and Fail to Generalize (ICML 2024 · similarity 0.84)
- Exact, Tractable Gauss-Newton Optimization in Deep Reversible Architectures Reveal Poor Generalization (NeurIPS 2024 · similarity 0.83)
- Neural Gaussian Similarity Modeling for Differential Graph Structure Learning (AAAI 2024 · similarity 0.83)
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.
