Research map / Probabilistic and scientific modeling / Molecules and proteins
Learning Libraries of Subroutines for Neurally–Guided Bayesian Program Induction
Kevin Ellis, Lucas Morales, Mathias Sablé-Meyer, Armando Solar-Lezama, Josh Tenenbaum
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 neural recognition model (ablation) | EC2 · List Processing | % solved: 79% -> 94% | Improved |
| learn DSL with fragments instead of only subexpressions (ablation) | EC2 · List Processing | % solved: 71% -> 94% | Improved |
| add neural recognition model (ablation) | EC2 · Text Editing | % solved: 43% -> 74% | Improved |
| learn DSL with fragments instead of only subexpressions (ablation) | EC2 · Text Editing | % solved: 30% -> 74% | Improved |
| add neural recognition model (ablation) | EC2 · Symbolic Regression | % solved: 75% -> 84% | Improved |
| learn DSL with fragments instead of only subexpressions (ablation) | EC2 · Symbolic Regression | % solved: 62% -> 84% | Improved |
| learn DSL structure (vs recognition model only on fixed DSL) (ablation) | EC2 · List Processing | % solved: 35% -> 94% | Improved |
| learn DSL structure (vs recognition model only on fixed DSL) (ablation) | EC2 · Text Editing | % solved: 33% -> 74% | Improved |
| learn DSL structure (vs recognition model only on fixed DSL) (ablation) | EC2 · Symbolic Regression | % solved: 38% -> 84% | Improved |
| learn DSL structure (vs PCFG weights only, no NN) (ablation) | No NN · List Processing | % solved: 62% -> 79% | Improved |
| learn DSL on SyGuS text editing (ablation) | EC2 · SyGuS text editing | % solved: 3.7% -> 74.1% | Improved |
| add recognition model (ablation) | EC2 · List Processing training tasks | % solved | No clear effect |
Closest papers by meaning
- Gradient-Based Program Synthesis with Neurally Interpreted Languages (ICLR 2026 · similarity 0.88)
- HYSYNTH: Context-Free LLM Approximation for Guiding Program Synthesis (NeurIPS 2024 · similarity 0.87)
- Learning Differentiable Programs with Admissible Neural Heuristics (NeurIPS 2020 · similarity 0.87)
- Bayesian Program Learning by Decompiling Amortized Knowledge (ICML 2024 · similarity 0.87)
- Program Synthesis via Test-Time Transduction (NeurIPS 2025 · similarity 0.86)
- Learning Structure-Aware Representations of Dependent Types (NeurIPS 2024 · similarity 0.85)
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.
