AI Paper Atlas

Research map / Large language models / Agents, code and math

Learning Compositional Rules via Neural Program Synthesis

Maxwell Nye, Armando Solar-Lezama, Josh Tenenbaum, Brenden M Lake

NeurIPS 2020 · Large language models › Agents, code and math · approach: LLM-guided program synthesis · 29 citations (OpenAlex)

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 changedWhereMeasuredEffect
Add guess-and-check search to neural grammar synthesis (ablation)Synth · SCAN length splitaccuracy: 0.0 -> 100Improved
Add guess-and-check search to neural grammar synthesis (ablation)Synth · SCAN simple splitaccuracy: 13.3 (3.3) -> 100Improved
Add guess-and-check search to neural grammar synthesis (ablation)Synth · SCAN add jump splitaccuracy: 3.5 (0.7) -> 100Improved
Add guess-and-check search to neural grammar synthesis (ablation)Synth · SCAN add around right splitaccuracy: 0.0 -> 100Improved
Use program synthesis with explicit checking rather than direct neural input-output predictionSynth · SCAN length splitaccuracy: 0.04 (0.02) -> 100Improved
Use program synthesis with explicit checking rather than direct neural input-output predictionSynth · Spanish number-word learningaccuracy: 64.4 (3.2) -> 80.0 (17.9)Improved
Add guess-and-check search to neural grammar synthesis (ablation)Synth · Spanish number-word learningaccuracy: 0.0 -> 80.0 (17.9)Improved
Add guess-and-check search to neural grammar synthesis (ablation)Synth · French number-word learningaccuracy: 0.0 -> 75.5 (2.4)Improved
Add guess-and-check search to neural grammar synthesis (ablation)Synth · Greek number-word learningaccuracy: 70.0 (10.2) -> 94.5 (4.9)Improved
Add guess-and-check search to neural grammar synthesis (ablation)Synth · Vietnamese number-word learningaccuracy: 69.5 (2.3) -> 69.5 (2.3)No clear effect
Use search to find a support-consistent grammar (ablation)Synth · MiniSCAN human instruction-learning query setaccuracy: 37.5% -> 98.75%Improved
Use neural attention across support examples in place of separate example encodersNeural program synthesis · MiniSCANaccuracy: 3%, 4%, 3%, 3.5% vs. synthesis resultsImproved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy