Research map / Large language models / Agents, code and math
Neural Programming by Example
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 NPBE instead of an LSTM encoder-decoder | String manipulation program induction · All 45 Tasks | program accuracy (Top1): 20.4% -> 74.1% | Helped |
| Use NPBE instead of an LSTM with attention encoder-decoder | String manipulation program induction · All 45 Tasks | program accuracy (Top1): 22.0% -> 74.1% | Helped |
| Allow three candidate predictions for Select integer arguments | NPBE · All 45 Tasks | program accuracy: 74.1% -> 85.8% | Helped |
| Allow five candidate predictions for Select integer arguments | NPBE · All 45 Tasks | program accuracy: 74.1% -> 91.0% | Helped |
| Use previously unseen argument settings rather than seen settings | NPBE · 19 selected Tasks | program accuracy (Top5): 87.6% -> 87.4% | No clear effect |
| Train on known task structures; attempt to generalize to completely unseen tasks | NPBE · Unseen task structures | program generalization | Hurt |
Closest papers by meaning
- Is Programming by Example Solved by LLMs? (NeurIPS 2024 · similarity 0.86)
- Neuro-Symbolic Program Synthesis (ICLR 2017 · similarity 0.86)
- DeepCoder: Learning to Write Programs (ICLR 2017 · similarity 0.84)
- Grammar Filtering for Syntax-Guided Synthesis (AAAI 2020 · similarity 0.84)
- Generating Pragmatic Examples to Train Neural Program Synthesizers (ICLR 2024 · similarity 0.81)
- Data-Efficient Learning with Neural Programs (NeurIPS 2024 · similarity 0.81)
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.