Research map / Large language models / Agents, code and math
Inferring the Invisible: Neuro-Symbolic Rule Discovery for Missing Value Imputation
Wendi Ren, Ke Wan, Junyu Leng, Shuang Li
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 |
|---|---|---|---|
| add joint fine-tuning with soft-OR after sequential covering (ablation) | NS-FCN · Synthetic disjunctive rule X5 | imputation accuracy: 0.8729 -> 1.0 | Improved |
| joint optimization of rule discovery and target inference vs impute-then-classify | NS-FCN · Heart Disease downstream | accuracy: 0.89 vs best baseline 0.85 | Improved |
| model continuous features directly with soft predicates instead of discretizing | NS-FCN · Heart Disease imputation | imputation accuracy: trestbps 0.87 vs 0.88; oldpeak 0.78 vs 0.89 (GAIN) | Mixed |
| use NS-FCN instead of MLP on simple Birds logic | NS-FCN · Birds | imputation accuracy: 0.96 -> 0.95 (abnormal bird) | Hurt |
| use NS-FCN rule learner instead of rule-based baselines (LEN, RRL, BRCG, DR-NET) | NS-FCN · Birds | imputation accuracy: 0.57 -> 1.00 | Improved |
| change missingness from MCAR to MNAR (ablation) | NS-FCN · Synthetic chain X5 | imputation accuracy: 0.95 -> 0.93 | No clear effect |
| change rule optimization order (ablation) | NS-FCN · Synthetic chain (a) | rule learning success: X4 imputation 0.87 vs 1.00 across orders | Mixed |
Closest papers by meaning
- Right for the Right Reasons: Avoiding Reasoning Shortcuts via Prototypical Neurosymbolic AI (NeurIPS 2025 · similarity 0.85)
- Mastering Symbolic Operations: Augmenting Language Models with Compiled Neural Networks (ICLR 2024 · similarity 0.85)
- Learning What Not to Impute: An Uncertainty-Aware Diffusion Framework for Meaningful Missingness (NeurIPS 2026 · similarity 0.85)
- Neuro-Symbolic Temporal Point Processes (ICML 2024 · similarity 0.85)
- Deciphering Raw Data in Neuro-Symbolic Learning with Provable Guarantees (AAAI 2024 · similarity 0.85)
- Concept-RuleNet: Grounded Multi-Agent Neurosymbolic Reasoning in Vision Language Models (AAAI 2026 · similarity 0.84)
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.
