Fully Autonomous Programming with Large Language Models

被引:8
|
作者
Liventsev, Vadim [1 ,2 ]
Grishina, Anastasiia [3 ,4 ]
Harma, Aki [2 ]
Moonen, Leon [3 ,5 ]
机构
[1] TU Eindhoven, Eindhoven, Netherlands
[2] Philips Res, Eindhoven, Netherlands
[3] Simula, Oslo, Norway
[4] Univ Oslo, Oslo, Norway
[5] BI Norwegian Business Sch, Oslo, Norway
关键词
automatic programming; large language models; program repair;
D O I
10.1145/3583131.3590481
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Current approaches to program synthesis with Large Language Models (LLMs) exhibit a "near miss syndrome": they tend to generate programs that semantically resemble the correct answer (as measured by text similarity metrics or human evaluation), but achieve a low or even zero accuracy as measured by unit tests due to small imperfections, such as the wrong input or output format. This calls for an approach known as Synthesize, Execute, Debug (SED), whereby a draft of the solution is generated first, followed by a program repair phase addressing the failed tests. To effectively apply this approach to instruction-driven LLMs, one needs to determine which prompts perform best as instructions for LLMs, as well as strike a balance between repairing unsuccessful programs and replacing them with newly generated ones. We explore these trade-offs empirically, comparing replace-focused, repair-focused, and hybrid debug strategies, as well as different template-based and model-based prompt-generation techniques. We use OpenAI Codex as the LLM and Program Synthesis Benchmark 2 as a database of problem descriptions and tests for evaluation. The resulting framework outperforms both conventional usage of Codex without the repair phase and traditional genetic programming approaches.
引用
收藏
页码:1146 / 1155
页数:10
相关论文
共 50 条
  • [1] Prompting Is Programming: A Query Language for Large Language Models
    Beurer-Kellner, Luca
    Fischer, Marc
    Vechev, Martin
    PROCEEDINGS OF THE ACM ON PROGRAMMING LANGUAGES-PACMPL, 2023, 7 (PLDI):
  • [2] Autonomous chemical research with large language models
    Boiko, Daniil A.
    Macknight, Robert
    Kline, Ben
    Gomes, Gabe
    NATURE, 2023, 624 (7992) : 570 - +
  • [3] Autonomous chemical research with large language models
    Daniil A. Boiko
    Robert MacKnight
    Ben Kline
    Gabe Gomes
    Nature, 2023, 624 : 570 - 578
  • [4] AskIt: Unified Programming Interface for Programming with Large Language Models
    Okuda, Katsumi
    Amarasinghe, Saman
    2024 IEEE/ACM INTERNATIONAL SYMPOSIUM ON CODE GENERATION AND OPTIMIZATION, CGO, 2024, : 41 - 54
  • [5] Propagating Large Language Models Programming Feedback
    Koutcheme, Charles
    Hellas, Arto
    PROCEEDINGS OF THE ELEVENTH ACM CONFERENCE ON LEARNING@SCALE, L@S 2024, 2024, : 366 - 370
  • [6] A Survey on Multimodal Large Language Models for Autonomous Driving
    Cui, Can
    Ma, Yunsheng
    Cao, Xu
    Ye, Wenqian
    Zhou, Yang
    Liang, Kaizhao
    Chen, Jintai
    Lu, Juanwu
    Yang, Zichong
    Liao, Kuei-Da
    Gao, Tianren
    Li, Erlong
    Tang, Kun
    Cao, Zhipeng
    Zhou, Tong
    Liu, Ao
    Yan, Xinrui
    Mei, Shuqi
    Cao, Jianguo
    Wang, Ziran
    Zheng, Chao
    2024 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION WORKSHOPS, WACVW 2024, 2024, : 958 - 979
  • [7] A review of large language models and autonomous agents in chemistry
    Ramos, Mayk Caldas
    Collison, Christopher J.
    White, Andrew D.
    CHEMICAL SCIENCE, 2025, 16 (06) : 2514 - 2572
  • [8] A Superalignment Framework in Autonomous Driving with Large Language Models
    Kong, Xiangrui
    Braunl, Thomas
    Fahmi, Marco
    Wang, Yue
    2024 35TH IEEE INTELLIGENT VEHICLES SYMPOSIUM, IEEE IV 2024, 2024, : 1715 - 1720
  • [9] Facilitating Autonomous Driving Tasks With Large Language Models
    Wu, Mengyao
    Yu, F. Richard
    Liu, Peter Xiaoping
    He, Ying
    IEEE INTELLIGENT SYSTEMS, 2025, 40 (01) : 45 - 52
  • [10] Large Language Models in Robot Programming Potential in the programming of industrial robots
    Syniawa, Daniel
    Ates, Baris
    Boshoff, Marius
    Kuhlenkoetter, Bernd
    ATP MAGAZINE, 2024, (6-7):