# Latent Attention For If-Then Program Synthesis

arXiv: https://arxiv.org/abs/1611.01867 · HTML: https://mlsanity.com/paper/1611.01867/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 7 Nov 2016.

## Built on by (most cited first, 10 of 10)

- [Synchromesh: Reliable code generation from pre-trained language models](https://mlsanity.com/paper/2201.11227/) (2022)
- [Reducing hallucination in structured outputs via Retrieval-Augmented Generation](https://mlsanity.com/paper/2404.08189/) (2024)
- [SeaD: End-to-end Text-to-SQL Generation with Schema-aware Denoising](https://mlsanity.com/paper/2105.07911/) (2021)
- [A Survey on Machine Learning Techniques for Source Code Analysis](https://mlsanity.com/paper/2110.09610/) (2021)
- [Deep Learning for Code Intelligence: Survey, Benchmark and Toolkit](https://mlsanity.com/paper/2401.00288/) (2023)
- [ProTo: Program-Guided Transformer for Program-Guided Tasks](https://mlsanity.com/paper/2110.00804/) (2021)
- [Neural Attention Models in Deep Learning: Survey and Taxonomy](https://mlsanity.com/paper/2112.05909/) (2021)
- [A Survey of Deep Learning Models for Structural Code Understanding](https://mlsanity.com/paper/2205.01293/) (2022)
- [A Survey on Natural Language Processing for Programming](https://mlsanity.com/paper/2212.05773/) (2022)
- [Proactive Detection of Physical Inter-rule Vulnerabilities in IoT Services Using a Deep Learning Approach](https://mlsanity.com/paper/2406.03836/) (2024)

## Abstract (arXiv)

Automatic translation from natural language descriptions into programs is a longstanding challenging problem. In this work, we consider a simple yet important sub-problem: translation from textual descriptions to If-Then programs. We devise a novel neural network architecture for this task which we train end-to-end. Specifically, we introduce Latent Attention, which computes multiplicative weights for the words in the description in a two-stage process with the goal of better leveraging the natural language structures that indicate the relevant parts for predicting program elements. Our architecture reduces the error rate by 28.57% compared to prior art. We also propose a one-shot learning scenario of If-Then program synthesis and simulate it with our existing dataset. We demonstrate a variation on the training procedure for this scenario that outperforms the original procedure, significantly closing the gap to the model trained with all data.