Execution-Guided Neural Program Synthesis

By Xinyun Chen, Chang Liu, Dawn Song. 2018

174 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Natural Language Processing

Citations in the corpus

Reasoning Like Program Executors (2022, background): “Other lines of research leverage program execution in inference as a reliable sanity guarantee for generated programs by pruning non-executable candidates (Wang et al., 2018b; Chen et al., 2019b, 2021; Odena et al., 2020; Ellis et al., 2019; Chen et al., 2019b; Sun et al., 2018; Zohar and Wolf, 2018”
Program Synthesis with Large Language Models (2021, methodology): “Many learning methods have been proposed for program repair [Allamanis et al., 2018b, Tarlow et al., 2019, Hellendoorn et al., 2019, Dinella et al., 2019, Yasunaga and Liang, 2020, Chen et al., 2019b, Pradel and Sen, 2018].”
CodeT: Code Generation with Generated Tests (2022, methodology): “This execution-guided approach has been widely adopted in various code-related tasks, such as code generation (Chen et al., 2021; Li et al., 2022b; Shi et al., 2022), code translation (Roziere et al., 2021), and program synthesis (Chen et al., 2018; Ellis et al., 2019).”
BUSTLE: Bottom-up program-Synthesis Through Learning-guided Exploration (2020, background): “, 2018), and learned search based on partial executions (Ellis et al., 2019; Zohar & Wolf, 2018; Chen et al., 2019).”
Synthetic Data for Deep Learning (2019, background): “Learning-based program synthesis (earlier attempts were based on logical inference [392]) began with learning string manipulation programs [218] and soon branched into deep learning, using recurrent architectures to guide search-based methods [614] or to generate programs in encoder-decoder architec”
Large Language Models as Analogical Reasoners (2023, background): “…et al., 2019; Feng et al., 2020; Zhang et al., 2022a; Yasunaga et al., 2021; 2022a; Xie et al., 2022), and incorporating external modules for reasoning such as logic and program execution (Chen et al., 2018; 2019; Yasunaga & Liang, 2020; Gupta et al., 2020; Ren et al., 2021; Zhang et al., 2023).”
Improving Molecular Design by Stochastic Iterative Target Augmentation (2020, methodology): “Our task is based on the educational Karel programming language (Pattis, 1981) used for evaluation in Bunel et al. (2018) and Chen et al. (2019).”
Demystifying GPT Self-Repair for Code Generation (2023, background): “Our work is also different in that we assume access to the full collection of input-output examples, as is typically done in inductive synthesis [6, 10, 12, 16, 30].”
Voyager: An Open-Ended Embodied Agent with Large Language Models (2023, background): “[83] Xinyun Chen, Chang Liu, and Dawn Song.”
LEVER: Learning to Verify Language-to-Code Generation with Execution (2023, methodology): “Intermediate execution results were used to guide program search at both training (Chen et al., 2019; 2021b) and inference time (Wang et al., 2018).”
Planning with Large Language Models for Code Generation (2023, methodology): “The code generation problem has been formulated as a sequential decision-making problem (Bunel et al., 2018; Chen et al., 2018), which enables designing and applying reinforcement learning (RL) and planning algorithms for code generation.”
Learning to Synthesize Programs as Interpretable and Generalizable Policies (2021, background): “[18] Xinyun Chen, Chang Liu, and Dawn Song.”
LEGO: Latent Execution-Guided Reasoning for Multi-Hop Question Answering on Knowledge Graphs (2021, methodology): “To overcome the requirement of the rules, recent KGQA methods leverage neural semantic parsing with a deep network (Berant et al., 2013; Lan & Jiang, 2020; Liang et al., 2017; Yih et al., 2015; 2016; Luo et al., 2018; Chen et al., 2019b; Hu et al., 2018; Qiu et al., 2020b).”
In-IDE Code Generation from Natural Language: Promise and Challenges (2021, methodology): “to automatically generate functions, with some recent advances that use machine learning models [9, 20, 26, 103].”
Write, Execute, Assess: Program Synthesis with a REPL (2019, methodology): “The policy-only beam search method is inspired by [12] and [13], which performed execution-guided synthesis in Karel and list processing domains, respectively.”
MetaGPT: Meta Programming for Multi-Agent Collaborative Framework (2023, methodology): “There are also state-of-the-art methods that utilize majority voting to select candidate programs [28] and use execution results to improve program synthesis [29, 30].”
Learning Execution through Neural Code Fusion (2019, background): “2 Using Program State to Guide Program Synthesis There are several works that learn from program state to aid program synthesis [3, 32, 12, 47, 7, 42, 30].”
Learning Compositional Rules via Neural Program Synthesis (2020, methodology): “We are inspired by work such as RobustFill [11], enumerative approaches [16], execution guided work [17, 18, 10, 19], and hybrid models [20, 21].”
Teaching Large Language Models to Self-Debug (2023, methodology): “, 2023), and performing selection based on the consensus on unit test execution outputs among samples (Chen et al., 2019; 2021b; Roziere et al., 2022; Shi et al., 2022; Li et al., 2022; Chen et al., 2023b).”
IReEn: Iterative Reverse-Engineering of Black-Box Functions via Neural Program Synthesis (2020, background): “We base our notation on (Bunel et al., 2018; Shin et al., 2018; Chen et al., 2019).”