Research map / Probabilistic and scientific modeling / Variational and generative sampling
Learning to Reason: Leveraging Neural Networks for Approximate DNF Counting
Ralph Abboud, Ismail Ceylan, Thomas Lukasiewicz
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 message-passing iterations rather than 2 (ablation) | NeuralDNF · Structure test set | accuracy at threshold 0.10: 91.36 -> 99.95 | Improved |
| Use 8 message-passing iterations rather than 2 (ablation) | NeuralDNF · 10K-variable size generalization | accuracy at threshold 0.10: 84.20 -> 97.13 | Improved |
| Use 8 message-passing iterations rather than 2 (ablation) | NeuralDNF · 15K-variable size generalization | accuracy at threshold 0.10: 79.31 -> 94.83 | Improved |
| Use 32 rather than 8 message-passing iterations (ablation) | NeuralDNF · Structure test set | accuracy at threshold 0.10: 99.95 -> 99.83 | No clear effect |
| Use 32 rather than 8 message-passing iterations (ablation) | NeuralDNF · 10K-variable size generalization | accuracy at threshold 0.10: 97.13 -> 93.67 | Hurt |
| Use 32 rather than 8 message-passing iterations (ablation) | NeuralDNF · 15K-variable size generalization | accuracy at threshold 0.10: 94.83 -> 87.93 | Hurt |
| Use NeuralDNF instead of KLM | Weighted DNF counting · 15K-variable formulas, width 3 | runtime (s): 2375.56 -> 0.104 | Improved |
Closest papers by meaning
- Where to Approximate in Neurosymbolic Inference? (NeurIPS 2026 · similarity 0.89)
- On the Hardness of Probabilistic Neurosymbolic Learning (ICML 2024 · similarity 0.85)
- The Gradient of Algebraic Model Counting (AAAI 2025 · similarity 0.83)
- Language Models Need Inductive Biases to Count Inductively (ICLR 2025 · similarity 0.83)
- Model Counting and Sampling via Semiring Extensions (AAAI 2024 · similarity 0.82)
- On Neural Networks as Infinite Tree-Structured Probabilistic Graphical Models (NeurIPS 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.
