# Neural Program Meta-Induction

arXiv: https://arxiv.org/abs/1710.04157 · HTML: https://mlsanity.com/paper/1710.04157/

## Key facts

- Cited by 23 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 11 Oct 2017.

## Built on by (most cited first, 23 of 23)

- [Large Language Models Are Human-Level Prompt Engineers](https://mlsanity.com/paper/2211.01910/) (2022)
- [Learning Performance-Improving Code Edits](https://mlsanity.com/paper/2302.07867/) (2023)
- [Learning to Synthesize Programs as Interpretable and Generalizable Policies](https://mlsanity.com/paper/2108.13643/) (2021)
- [Communicating Natural Programs to Humans and Machines](https://mlsanity.com/paper/2106.07824/) (2021)
- [Latent Execution for Neural Program Synthesis](https://mlsanity.com/paper/2107.00101/) (2021)
- [A Review of Emerging Research Directions in Abstract Visual Reasoning](https://mlsanity.com/paper/2202.10284/) (2022)
- [Emergent Representations of Program Semantics in Language Models Trained on Programs](https://mlsanity.com/paper/2305.11169/) (2023)
- [ProTo: Program-Guided Transformer for Program-Guided Tasks](https://mlsanity.com/paper/2110.00804/) (2021)
- [Learning to Combine Per-Example Solutions for Neural Program Synthesis](https://mlsanity.com/paper/2106.07175/) (2021)
- [ProtoTransformer: A Meta-Learning Approach to Providing Student Feedback](https://mlsanity.com/paper/2107.14035/) (2021)
- [From {Solution Synthesis} to {Student Attempt Synthesis} for Block-Based Visual Programming Tasks](https://mlsanity.com/paper/2205.01265/) (2022)
- [Synthesis of Mathematical programs from Natural Language Specifications](https://mlsanity.com/paper/2304.03287/) (2023)
- [Hierarchical Neural Program Synthesis](https://mlsanity.com/paper/2303.06018/) (2023)
- [Gradient-Based Program Synthesis with Neurally Interpreted Languages](https://mlsanity.com/paper/2604.18907/) (2026)
- [Learning to Improve Code Efficiency](https://mlsanity.com/paper/2208.05297/) (2022)
- [Program Synthesis Benchmark for Visual Programming in XLogoOnline Environment](https://mlsanity.com/paper/2406.11334/) (2024)
- [Learning Program Behavioral Models from Synthesized Input-Output Pairs](https://mlsanity.com/paper/2407.08597/) (2024)
- [Action abstractions for amortized sampling](https://mlsanity.com/paper/2410.15184/) (2024)
- [Neurogenetic Programming Framework for Explainable Reinforcement Learning](https://mlsanity.com/paper/2102.04231/) (2021)
- [Reevaluating Meta-Learning Optimization Algorithms Through Contextual Self-Modulation](https://mlsanity.com/paper/2410.01655/) (2024)
- [Improved Tree Search for Automatic Program Synthesis](https://mlsanity.com/paper/2303.07166/) (2023)
- [Guided Sketch-Based Program Induction by Search Gradients](https://mlsanity.com/paper/2402.06990/) (2024)
- [Optimization Problem Solving Can Transition to Evolutionary Agentic Workflows](https://mlsanity.com/paper/2505.04354/) (2025)

## Abstract (arXiv)

Most recently proposed methods for Neural Program Induction work under the assumption of having a large set of input/output (I/O) examples for learning any underlying input-output mapping. This paper aims to address the problem of data and computation efficiency of program induction by leveraging information from related tasks. Specifically, we propose two approaches for cross-task knowledge transfer to improve program induction in limited-data scenarios. In our first proposal, portfolio adaptation, a set of induction models is pretrained on a set of related tasks, and the best model is adapted towards the new task using transfer learning. In our second approach, meta program induction, a k-shot learning approach is used to make a model generalize to new tasks without additional training. To test the efficacy of our methods, we constructed a new benchmark of programs written in the Karel programming language. Using an extensive experimental evaluation on the Karel benchmark, we demonstrate that our proposals dramatically outperform the baseline induction method that does not use knowledge transfer. We also analyze the relative performance of the two approaches and study conditions in which they perform best. In particular, meta induction outperforms all existing approaches under extreme data sparsity (when a very small number of examples are available), i.e., fewer than ten. As the number of available I/O examples increase (i.e. a thousand or more), portfolio adapted program induction becomes the best approach. For intermediate data sizes, we demonstrate that the combined method of adapted meta program induction has the strongest performance.