CAREER: Democratizing Algorithmic Program Synthesis

职业：算法程序综合民主化

• 批准号: 2046071
• 负责人: Xiaokang Qiu
• 金额: $ 49.25万
• 依托单位: Purdue University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-04-01 至 2026-03-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2046071&HistoricalAwards=false
• 关键词: CAREER; Democratizing; Algorithmic; Program; Synthesis

Program synthesis is the process of automatically generating programs that meet the user's intent. As one of the most central research problems in computer science and AI, program synthesis is extremely challenging. Recent years have seen success stories of algorithmic program synthesis in real-life code generation. Despite the promising progress, modern-day synthesizers can still only be mastered by experts. One vital problem is that program synthesis has not been integrated to the traditional software development process that most regular programmers are familiar with. This project takes a major step toward lowering the barrier of algorithmic program synthesis for regular programmers. The key insight is that the challenging steps in the current synthesis process can be perceived as part of the synthesis task and incorporated into the synthesis algorithms. More specifically, this project develops new synthesis frameworks and algorithms that enable scalable, quantitative, library-based program synthesis without hand-crafted objective functions, library models or decomposition strategies. This project aims to make software development easier for regular programmers by transforming program synthesizers into more natural and accessible tools. The project also has the potential to significantly broaden the base of people who can overcome the programming-related challenges in their own fields. The project will provide valuable training for graduate students, enable novel pedagogical approaches and material suitable for compiler-related, synthesis-related and networking-related courses, and provide new computational thinking assessments and tutorials for broadening participation of K-12 and undergraduate students for broader community.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

程序合成是自动生成满足用户意图的程序的过程。作为计算机科学和人工智能领域最核心的研究问题之一，程序综合是一个极具挑战性的问题。近年来，在现实生活中的代码生成算法程序合成的成功故事。尽管取得了令人鼓舞的进展，但现代合成器仍然只能由专家掌握。一个至关重要的问题是，程序合成还没有集成到大多数普通程序员熟悉的传统软件开发过程中。这个项目朝着降低普通程序员的算法程序合成障碍迈出了重要的一步。关键的见解是，在当前的合成过程中的挑战性的步骤可以被视为合成任务的一部分，并纳入合成算法。更具体地说，该项目开发了新的合成框架和算法，使可扩展的，定量的，基于库的程序合成，而无需手工制作的目标函数，库模型或分解策略。该项目旨在通过将程序合成器转换为更自然和更易于访问的工具，使普通程序员的软件开发更容易。该项目还有可能大大扩大能够克服各自领域与方案编制有关的挑战的人员基础。该项目将为研究生提供宝贵的培训，使新的教学方法和材料适合于与汇编有关、与综合有关和与网络有关的课程，并提供新的计算思维评估和教程，以扩大K-该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准。

项目成果

Comparative Synthesis: Learning Near-Optimal Network Designs by Query

• DOI: 10.1145/3571197
• 发表时间: 2023-01-01
• 期刊: PROCEEDINGS OF THE ACM ON PROGRAMMING LANGUAGES-PACMPL
• 影响因子: 1.8
• 作者: Wang，Yanjun;Li，Zixuan;Rao，Sanjay
• 通讯作者: Rao，Sanjay

Enhanced Enumeration Techniques for Syntax-Guided Synthesis of Bit-Vector Manipulations

用于位向量操作的语法引导合成的增强枚举技术

• DOI: 10.1145/3632913
• 发表时间: 2024
• 期刊: Proceedings of the ACM on Programming Languages
• 作者: Ding, Yuantian;Qiu, Xiaokang
• 通讯作者: Qiu, Xiaokang

Bootstrapping Library-Based Synthesis

基于引导库的合成

• 发表时间: 2022
• 期刊: Lecture notes in computer science
• 作者: Huang, Kangjing;Qiu, Xiaokang
• 通讯作者: Qiu, Xiaokang