Research map / Learning theory and algorithms / Combinatorial and search optimization
Write, Execute, Assess: Program Synthesis with a REPL
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 |
|---|---|---|---|
| use value-guided SMC instead of the compared inference strategies | REPL policy for inverse CAD · 2D and 3D inverse CAD | test accuracy | Helped |
| add a learned value function to guide search (ablation) | policy-based beam search and sampling · 2D and 3D inverse CAD | program quality (IoU) | Helped |
| remove the REPL and decode from the specification in one shot (ablation) | REPL policy for inverse CAD · 2D and 3D inverse CAD | program quality (IoU) | Mixed |
| use value-guided SMC rather than the compared inference techniques | REPL policy for string editing · String Editing, 87-problem dataset | number of correct programs | Helped |
| use value-guided SMC rather than the compared inference techniques | REPL policy for string editing · String Editing, 87-problem dataset | node expansions: 10x fewer node expansions | Helped |
| add a learned value function to beam search (ablation) | policy-only beam search · String Editing, 87-problem dataset | number of correct programs | Helped |
| remove value guidance from beam search (ablation) | REPL policy for string editing · String Editing, 87-problem dataset | number of correct programs | Hurt |
Closest papers by meaning
- Latent Execution for Neural Program Synthesis Beyond Domain-Specific Languages (NeurIPS 2021 · similarity 0.90)
- Representing Partial Programs with Blended Abstract Semantics (ICLR 2021 · similarity 0.89)
- HYSYNTH: Context-Free LLM Approximation for Guiding Program Synthesis (NeurIPS 2024 · similarity 0.88)
- Program Synthesis via Test-Time Transduction (NeurIPS 2025 · similarity 0.87)
- Guess & Sketch: Language Model Guided Transpilation (ICLR 2024 · similarity 0.87)
- From I/O to Code with Discovery Agent (NeurIPS 2026 · 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.