Research map / Probabilistic and scientific modeling / Variational and generative sampling
Logical Credal Networks
Radu Marinescu, Haifeng Qian, Alexander G. Gray, Debarun Bhattacharjya, Francisco Barahona, Tian Gao, Ryan Riegel, Pravinda Sahu
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 LCN maximin inference instead of Credal maximin (adds logic formulas with bounds) | credal network · Mastermind | accuracy: 64.4 -> 71.1 | Improved |
| use LCN maximax instead of Credal maximax | credal network · Mastermind | accuracy: 60.3 -> 69.2 | Improved |
| use maximum-entropy joint distribution (LCN_maxent) instead of maximin (ablation) | LCN · Mastermind | accuracy: 71.1 -> 72.5 | Improved |
| use LCN maximin instead of Bayesian midpoint | Bayesian_midpoint · Mastermind | accuracy: 65.7 -> 71.1 | Improved |
| add logic formulas as soft constraints (cProbLog/MLN) over ProbLog | ProbLog_midpoint · Mastermind | accuracy: 65.7 -> 65.3 | No clear effect |
| use Nilsson maximin without independence assumptions | Nilsson · Mastermind | accuracy: 0% | Hurt |
| use LCN instead of ProbLog_midpoint | ProbLog_midpoint · Credit card fraud | F1 score: 0.599 -> 0.630 | Improved |
| use LCN instead of Naive_Bayes (adding expert rules) | Naive_Bayes · Credit card fraud | F1 score: 0.408 -> 0.630 | Improved |
| use LCN instead of MLN_midpoint | MLN_midpoint · Credit card fraud | F1 score: 0.472 -> 0.630 | Improved |
| expand Naive Bayes with rule antecedents as parents (Bayesian_midpoint) via noisy-OR | Naive_Bayes · Credit card fraud | F1 score: 0.408 -> 0.090 | Hurt |
| use credal network intervals instead of midpoints | Bayesian_midpoint · Credit card fraud | F1 score: 0.090 -> 0.089 | No clear effect |
Closest papers by meaning
- Abductive Reasoning in Logical Credal Networks (NeurIPS 2024 · similarity 0.91)
- Efficient Probabilistic Logic Reasoning with Graph Neural Networks (ICLR 2020 · similarity 0.85)
- Decomposing Constraint Networks for Calculating c-Representations (AAAI 2024 · similarity 0.85)
- From Probability to Counterfactuals: the Increasing Complexity of Satisfiability in Pearl's Causal Hierarchy (ICLR 2025 · similarity 0.83)
- A Logical Analysis of Hanabi (AAAI 2025 · similarity 0.83)
- A Compositional Atlas for Algebraic Circuits (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.
