AI Coders Are Among Us: Rethinking Programming Language Grammar Towards Efficient Code Generation
作者: Zhensu Sun, Xiaoning Du, Zhou Yang, Li Li, David Lo
分类: cs.SE, cs.AI, cs.PL
发布日期: 2024-04-25 (更新: 2024-08-14)
备注: Accepted by ISSTA'24
💡 一句话要点
提出AI导向语法以提高代码生成效率
🎯 匹配领域: 支柱九：具身大模型 (Embodied Foundation Models)
关键词: AI导向语法 代码生成 大型语言模型 Python 计算效率 抽象语法树 人机协作
📋 核心要点
- 现有编程语言语法主要为人类开发者设计，导致AI模型在处理代码时面临不必要的计算负担。
- 本文提出AI导向语法，通过最小化语法标记和格式，优化代码表示以适应AI模型的工作机制。
- 实验表明，使用SimPy的CodeLlama和GPT-4在完成相同任务时，标记使用量分别减少13.5%和10.4%，且性能保持或提升。
📝 摘要（中文）
随着大型语言模型（LLMs）的发展，人工智能（AI）模型成为编程语言的重要用户。然而，现有编程语言的语法设计主要为人类开发者服务，导致LLMs在处理代码时面临不必要的计算负担。为此，本文提出AI导向语法的概念，旨在通过最小化语法和格式标记来提高推理效率。我们实现了首个Python的AI导向语法SimPy，经过一系列启发式规则修订原有Python语法，保持相同的抽象语法树结构。实验结果表明，SimPy在完成相同代码任务时，相较于Python，CodeLlama和GPT-4的标记使用量分别减少了13.5%和10.4%，且模型性能得以维持或提升。
🔬 方法详解
问题定义：本文旨在解决现有编程语言语法对AI模型的低效支持问题。现有语法设计过于复杂，增加了LLMs的计算负担。
核心思路：提出AI导向语法，通过简化语法结构和减少标记数量，使代码更适合AI模型的处理方式，从而提高推理效率。
技术框架：整体架构包括对原有Python语法的修订，形成新的SimPy语法。该框架允许通过修改的AST解析器执行代码，并实现Python与SimPy之间的无缝转换。
关键创新：SimPy作为首个AI导向语法，显著减少了代码中的标记数量，同时保持与标准Python相同的AST结构，这是与现有方法的本质区别。
关键设计：在设计SimPy时，采用了一系列启发式规则来修订Python语法，确保代码语义的有效传达，同时优化了标记的使用。
🖼️ 关键图片
📊 实验亮点
实验结果显示，使用SimPy的CodeLlama和GPT-4在完成相同代码任务时，标记使用量分别减少了13.5%和10.4%。此外，这些模型在使用SimPy时能够维持或提升其性能，表明AI导向语法的有效性和实用性。
🎯 应用场景
该研究的潜在应用领域包括AI编程助手、自动代码生成工具以及人机协作编程环境。通过优化代码语法，AI模型可以更高效地生成和理解代码，从而提升开发效率，降低计算成本，推动软件开发的智能化进程。
📄 摘要（原文）
Artificial Intelligence (AI) models have emerged as another important audience for programming languages alongside humans and machines, as we enter the era of large language models (LLMs). LLMs can now perform well in coding competitions and even write programs like developers to solve various tasks, including mathematical problems. However, the grammar and layout of current programs are designed to cater the needs of human developers -- with many grammar tokens and formatting tokens being used to make the code easier for humans to read. While this is helpful, such a design adds unnecessary computational work for LLMs, as each token they either use or produce consumes computational resources. To improve inference efficiency and reduce computational costs, we propose the concept of AI-oriented grammar. This aims to represent code in a way that better suits the working mechanism of AI models. Code written with AI-oriented grammar discards formats and uses a minimum number of tokens to convey code semantics effectively. To demonstrate the feasibility of this concept, we explore and implement the first AI-oriented grammar for Python, named SimPy. SimPy is crafted by revising the original Python grammar through a series of heuristic rules. Programs written in SimPy maintain identical AST structures to those in standard Python. This allows for not only execution via a modified AST parser, but also seamless transformation between programs written in Python and SimPy, enabling human developers and LLMs to use Python and SimPy, respectively, when they need to collaborate. In the experiments, compared with Python, SimPy enables a reduction in token usage by 13.5% and 10.4% for CodeLlama and GPT-4, respectively, when completing the same set of code-related tasks. Additionally, these models can maintain or even improve their performance when using SimPy instead of Python for these tasks.