Research map / Learning theory and algorithms / Combinatorial and search optimization
Representing Partial Programs with Blended Abstract Semantics
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 blended abstract semantics instead of neural abstract semantics (ablation) | Program synthesis in tower construction · Tower construction | percentage of test problems solved | Helped |
| Add a learned value function as the search heuristic | Blended abstract semantics · Tower construction | percentage of test problems solved | Helped |
| Use blended abstract semantics instead of RNN encoding (ablation) | Program synthesis · Tower construction | percentage of test problems solved | Helped |
| Use blended abstract semantics instead of RNN encoding (ablation) | Program synthesis · List processing | accuracy: 5-10% | Helped |
| Use blended abstract semantics instead of neural abstract semantics (ablation) | Program synthesis · List processing | accuracy | Helped |
| Use blended semantics instead of the unmodified RobustFill baseline | Program synthesis · String editing | accuracy | Helped |
| Use the domain-specific execution-guided REPL instead of blended semantics | Program synthesis · String editing | accuracy | Hurt |
| Use blended semantics instead of neural abstract semantics (ablation) | Program synthesis · String editing | accuracy | Helped |
| Use blended semantics instead of neural abstract semantics (ablation) | Program synthesis · Tower construction, list processing, and string editing | programs solved or accuracy: at least 5% | Helped |
Closest papers by meaning
- Write, Execute, Assess: Program Synthesis with a REPL (NeurIPS 2019 · similarity 0.89)
- Latent Execution for Neural Program Synthesis Beyond Domain-Specific Languages (NeurIPS 2021 · similarity 0.88)
- Execution-Guided Neural Program Synthesis (ICLR 2019 · similarity 0.88)
- HYSYNTH: Context-Free LLM Approximation for Guiding Program Synthesis (NeurIPS 2024 · similarity 0.88)
- ExeDec: Execution Decomposition for Compositional Generalization in Neural Program Synthesis (ICLR 2024 · similarity 0.87)
- Emergent Representations of Program Semantics in Language Models Trained on Programs (ICML 2024 · similarity 0.87)
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.