Predictive Synthesis of API-Centric Code

TL;DR

This paper introduces a deep learning model that predicts API function sequences to accelerate enumerative program synthesis, particularly for data processing tasks, by learning from input-output examples without explicit intermediate steps.

Contribution

It presents a novel deep learning approach to predict API sequences for program synthesis, improving speed over previous ML-based methods like DeepCoder.

Findings

Model achieves significant speedups in synthesis tasks

Outperforms previous ML-based synthesis methods

Effective in predicting API sequences from input-output data

Abstract

Today's programmers, especially data science practitioners, make heavy use of data-processing libraries (APIs) such as PyTorch, Tensorflow, NumPy, Pandas, and the like. Program synthesizers can provide significant coding assistance to this community of users; however program synthesis also can be slow due to enormous search spaces. In this work, we examine ways in which machine learning can be used to accelerate enumerative program synthesis. We present a deep-learning-based model to predict the sequence of API functions that would be needed to go from a given input to a desired output, both being numeric vectors. Our work is based on two insights. First, it is possible to learn, based on a large number of input-output examples, to predict the likely API function needed in a given situation. Second, and crucially, it is also possible to learn to compose API functions into a sequence, given an input and the desired final output, without explicitly knowing the intermediate values. We show that we can speed up an enumerative program synthesizer by using predictions from our model variants. These speedups significantly outperform previous ways (e.g. DeepCoder) in which researchers have used ML models in enumerative synthesis.