Research map / Probabilistic and scientific modeling / Variational and generative sampling
Einsum Networks: Fast and Scalable Learning of Tractable Probabilistic Circuits
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 |
|---|---|---|---|
| optimize EiNet with stochastic EM (sEM) instead of Adam (ablation) | EiNet on 20 binary datasets · 20 binary datasets (Table 1) | test log-likelihood: e.g. ad: -19.05 vs -48.47; sEM often better | Helped |
| optimize EiNet with stochastic EM instead of SGD (ablation) | EiNet on 20 binary datasets · 20 binary datasets (Table 1) | test log-likelihood: e.g. ad: -19.05 vs -63.81; nltcs -6.01 vs -6.17 | Helped |
| EiNet with sEM vs RAT-SPN baseline | EiNet on 20 binary datasets · 20 binary datasets (Table 1) | test log-likelihood: largely same as RAT-SPNs except ad | No clear effect |
| EiNet einsum implementation vs LibSPN and SPFlow | RAT-SPN style PCs training time · synthetic Gaussian data N=2000, D=512 | training time per epoch: one or two orders of magnitude faster | Helped |
| einsum implementation avoids explicit product nodes | EiNet memory consumption · synthetic Gaussian data N=2000, D=512 | GPU memory: for large K, memory an order of magnitude lower | Helped |
| EM implemented via automatic differentiation | EM training of PCs · algorithmic | implementation complexity | No clear effect |
Closest papers by meaning
- Scaling Probabilistic Circuits via Monarch Matrices (ICML 2025 · similarity 0.88)
- Probabilistic Neural Circuits (AAAI 2024 · similarity 0.87)
- Scaling Tractable Probabilistic Circuits: A Systems Perspective (ICML 2024 · similarity 0.87)
- The Expressivity Boundary of Probabilistic Circuits: A Comparison with Large Language Models (NeurIPS 2026 · similarity 0.87)
- Sum of Squares Circuits (AAAI 2025 · similarity 0.86)
- Neural Network Approximators for Marginal MAP in Probabilistic Circuits (AAAI 2024 · similarity 0.86)
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.