Research map / Probabilistic and scientific modeling / Variational and generative sampling
pomegranate: Fast and Flexible Probabilistic Modeling in Python
Jacob Schreiber
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 |
|---|---|---|---|
| use 8 threads instead of 1 (ablation) | pomegranate Gaussian naive Bayes · 3M samples, 1K dims | training time (s): ~65 -> ~17 | Improved |
| use 4 threads instead of 1 (ablation) | pomegranate full-covariance GMM · 2M samples, 150 dims | training time (s): ~470 -> ~135 | Improved |
| use 16 threads instead of 1 (ablation) | pomegranate full-covariance GMM · 2M samples, 150 dims | training time (s): ~470 -> ~57 | Improved |
| use GPU instead of 1 CPU thread (ablation) | pomegranate full-covariance GMM · 2M samples, 150 dims | training time (s): ~470 -> ~200 | Improved |
| use GPU instead of 16 CPU threads (ablation) | pomegranate full-covariance GMM · 2M samples, 150 dims | training time (s): ~57 -> ~200 | Hurt |
| use 16 threads instead of 1 (ablation) | pomegranate HMM Baum-Welch · 100 sequences x 1000 obs, 10 states | training time (s): ~13 -> ~2 | Improved |
| use pomegranate instead of hmmlearn | 10-state Gaussian HMM training · 100 sequences x 1000 obs | training time (s): ~25 -> ~13 | Improved |
| use pomegranate (1 thread) instead of scikit-learn | Gaussian naive Bayes training · 3M samples, 1K dims | training time (s): ~53 -> ~65 | Hurt |
| use pomegranate EM-based semi-supervised learning instead of scikit-learn label propagation (RBF kernel) | semi-supervised Gaussian models · 100k samples, 10-D | training time (s): ~220 -> ~0.04 | Improved |
| use pomegranate EM-based semi-supervised learning instead of scikit-learn label propagation | semi-supervised Gaussian models · 100k samples, 10-D | validation accuracy: over 0.75 vs chance | Improved |
Closest papers by meaning
- SimPro: A Simple Probabilistic Framework Towards Realistic Long-Tailed Semi-Supervised Learning (ICML 2024 · similarity 0.84)
- Recursive PAC-Bayes: A Frequentist Approach to Sequential Prior Updates with No Information Loss (NeurIPS 2024 · similarity 0.84)
- Discrete Markov Probabilistic Models: An Improved Discrete Score-Based Framework with sharp convergence bounds under minimal assumptions (ICML 2025 · similarity 0.84)
- Automated Statistical Model Discovery with Language Models (ICML 2024 · similarity 0.83)
- Scalable Bayesian Learning with posteriors (ICLR 2025 · similarity 0.83)
- Scaling Tractable Probabilistic Circuits: A Systems Perspective (ICML 2024 · similarity 0.82)
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.
