Research map / Probabilistic and scientific modeling / Variational and generative sampling
A Probabilistic Framework for Deep Learning
Read the paperCheck 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 |
|---|---|---|---|
| Train the 1-layer RFM with EG instead of DCN back-propagation | 1-layer RFM vs. Convnet · MNIST supervised classification | test error: 1.30% -> 1.21% | Helped |
| Train the 1-layer RFM with EG instead of DCN back-propagation | 1-layer RFM vs. DCN · MNIST supervised classification | epochs to reach 1.5% test error: 18 vs. 40 epochs | Helped |
| Use a 5-layer DRMM instead of the corresponding 5-layer DCN | 5-layer DRMM · MNIST supervised classification | test error: 0.81% vs 0.89% | Hurt |
| Add the KL regularizer (ablation) | DRMM 5-layer semi-sup NN+KL · MNIST semi-supervised test, 100 labeled | test error: 3.50 -> 0.57 | Helped |
| Increase depth from 2 layers to 5 layers (ablation) | DRMM semi-sup · MNIST semi-supervised test, 600 labeled | test error: 3.73 -> 1.56 | Helped |
| Increase depth from 2 layers to 5 layers (ablation) | DRMM semi-sup · MNIST semi-supervised test, 1K labeled | test error: 2.88 -> 1.67 | Helped |
| Increase depth from 2 layers to 5 layers (ablation) | DRMM semi-sup · MNIST semi-supervised test, 3K labeled | test error: 1.72 -> 0.91 | Helped |
| Use DRMM 5-layer unsup-pretr instead of SWWAE unsup-pretr | Unsupervised pretraining models · MNIST unsup-pretr test, 600 labeled | test error: 9.80 -> 3.61 | Helped |
| Use DRMM 5-layer unsup-pretr instead of SWWAE unsup-pretr | Unsupervised pretraining models · MNIST unsup-pretr test, 1K labeled | test error: 6.135 -> 2.73 | Helped |
| Use DRMM 5-layer unsup-pretr instead of SWWAE unsup-pretr | Unsupervised pretraining models · MNIST unsup-pretr test, 3K labeled | test error: 4.41 -> 1.68 | Helped |
Closest papers by meaning
- Deep Probabilistic Supervision for Image Classification (NeurIPS 2026 · similarity 0.85)
- Deep Homogeneous Mixture Models: Representation, Separation, and Approximation (NeurIPS 2018 · similarity 0.85)
- Direct Discriminative Optimization: Your Likelihood-Based Visual Generative Model is Secretly a GAN Discriminator (ICML 2025 · similarity 0.85)
- Analytical Probability Distributions and Exact Expectation-Maximization for Deep Generative Networks (NeurIPS 2020 · similarity 0.85)
- M3D: Dataset Condensation by Minimizing Maximum Mean Discrepancy (AAAI 2024 · similarity 0.84)
- On Neural Networks as Infinite Tree-Structured Probabilistic Graphical Models (NeurIPS 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.