@inproceedings{gorinski-etal-2023-automatic,
title = "Automatic Unit Test Data Generation and Actor-Critic Reinforcement Learning for Code Synthesis",
author = "Gorinski, Philip and
Zimmer, Matthieu and
Lampouras, Gerasimos and
Deik, Derrick Goh Xin and
Iacobacci, Ignacio",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.28",
doi = "10.18653/v1/2023.findings-emnlp.28",
pages = "370--384",
abstract = "The advent of large pre-trained language models in the domain of Code Synthesis has shown remarkable performance on various benchmarks, treating the problem of Code Generation in a fashion similar to Natural Language Generation, trained with a Language Modelling (LM) objective. In addition, the property of programming language code being precisely evaluable with respect to its semantics {--} through the use of Unit Tests to check its functional correctness {--} lends itself to using Reinforcement Learning (RL) as a further training paradigm. Previous work has shown that RL can be applied as such to improve models{'} coding capabilities; however, such RL-based methods rely on a reward signal based on defined Unit Tests, which are much harder to obtain compared to the huge crawled code datasets used in LM objectives. In this work, we present a novel approach to automatically obtain data consisting of function signatures and associated Unit Tests, suitable for RL training of Code Synthesis models. We also introduce a straightforward, simple yet effective Actor-Critic RL training scheme and show that it, in conjunction with automatically generated training data, leads to improvement of a pre-trained code language model{'}s performance by up to 9.9{\%} improvement over the original underlying code synthesis LM, and up to 4.3{\%} over RL-based models trained with standard PPO or CodeRL.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="gorinski-etal-2023-automatic">
<titleInfo>
<title>Automatic Unit Test Data Generation and Actor-Critic Reinforcement Learning for Code Synthesis</title>
</titleInfo>
<name type="personal">
<namePart type="given">Philip</namePart>
<namePart type="family">Gorinski</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Matthieu</namePart>
<namePart type="family">Zimmer</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gerasimos</namePart>
<namePart type="family">Lampouras</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Derrick</namePart>
<namePart type="given">Goh</namePart>
<namePart type="given">Xin</namePart>
<namePart type="family">Deik</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ignacio</namePart>
<namePart type="family">Iacobacci</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-12</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2023</title>
</titleInfo>
<name type="personal">
<namePart type="given">Houda</namePart>
<namePart type="family">Bouamor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Juan</namePart>
<namePart type="family">Pino</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kalika</namePart>
<namePart type="family">Bali</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Singapore</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>The advent of large pre-trained language models in the domain of Code Synthesis has shown remarkable performance on various benchmarks, treating the problem of Code Generation in a fashion similar to Natural Language Generation, trained with a Language Modelling (LM) objective. In addition, the property of programming language code being precisely evaluable with respect to its semantics – through the use of Unit Tests to check its functional correctness – lends itself to using Reinforcement Learning (RL) as a further training paradigm. Previous work has shown that RL can be applied as such to improve models’ coding capabilities; however, such RL-based methods rely on a reward signal based on defined Unit Tests, which are much harder to obtain compared to the huge crawled code datasets used in LM objectives. In this work, we present a novel approach to automatically obtain data consisting of function signatures and associated Unit Tests, suitable for RL training of Code Synthesis models. We also introduce a straightforward, simple yet effective Actor-Critic RL training scheme and show that it, in conjunction with automatically generated training data, leads to improvement of a pre-trained code language model’s performance by up to 9.9% improvement over the original underlying code synthesis LM, and up to 4.3% over RL-based models trained with standard PPO or CodeRL.</abstract>
<identifier type="citekey">gorinski-etal-2023-automatic</identifier>
<identifier type="doi">10.18653/v1/2023.findings-emnlp.28</identifier>
<location>
<url>https://aclanthology.org/2023.findings-emnlp.28</url>
</location>
<part>
<date>2023-12</date>
<extent unit="page">
<start>370</start>
<end>384</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Automatic Unit Test Data Generation and Actor-Critic Reinforcement Learning for Code Synthesis
%A Gorinski, Philip
%A Zimmer, Matthieu
%A Lampouras, Gerasimos
%A Deik, Derrick Goh Xin
%A Iacobacci, Ignacio
%Y Bouamor, Houda
%Y Pino, Juan
%Y Bali, Kalika
%S Findings of the Association for Computational Linguistics: EMNLP 2023
%D 2023
%8 December
%I Association for Computational Linguistics
%C Singapore
%F gorinski-etal-2023-automatic
%X The advent of large pre-trained language models in the domain of Code Synthesis has shown remarkable performance on various benchmarks, treating the problem of Code Generation in a fashion similar to Natural Language Generation, trained with a Language Modelling (LM) objective. In addition, the property of programming language code being precisely evaluable with respect to its semantics – through the use of Unit Tests to check its functional correctness – lends itself to using Reinforcement Learning (RL) as a further training paradigm. Previous work has shown that RL can be applied as such to improve models’ coding capabilities; however, such RL-based methods rely on a reward signal based on defined Unit Tests, which are much harder to obtain compared to the huge crawled code datasets used in LM objectives. In this work, we present a novel approach to automatically obtain data consisting of function signatures and associated Unit Tests, suitable for RL training of Code Synthesis models. We also introduce a straightforward, simple yet effective Actor-Critic RL training scheme and show that it, in conjunction with automatically generated training data, leads to improvement of a pre-trained code language model’s performance by up to 9.9% improvement over the original underlying code synthesis LM, and up to 4.3% over RL-based models trained with standard PPO or CodeRL.
%R 10.18653/v1/2023.findings-emnlp.28
%U https://aclanthology.org/2023.findings-emnlp.28
%U https://doi.org/10.18653/v1/2023.findings-emnlp.28
%P 370-384
Markdown (Informal)
[Automatic Unit Test Data Generation and Actor-Critic Reinforcement Learning for Code Synthesis](https://aclanthology.org/2023.findings-emnlp.28) (Gorinski et al., Findings 2023)
ACL