CAREER: Designing an Interactive Partner to Support Pair Programming

语音辅助设备展示了具有类似人类互动的会话虚拟助手的潜力。这项研究旨在开发一个类似alexa的编程伙伴，以改变编程的学习方式和编程的完成方式。该研究借鉴并扩展了人机交互、人工智能和软件工程的基本思想，以解决设计支持结对编程的会话代理的问题，在结对编程中，人类和代理参与丰富的协作交互。结对编程提高了代码质量、效率、知识转移和自我效能，并且通过开发结对编程伙伴，这项工作将影响软件开发以及编程的教学和学习，同时帮助保留计算机科学程序中代表性不足的群体。该项目将通过创建支持结对编程的以人为中心的交互伙伴来推进程序员与计算机的交互。基础研究将进行：(i)了解配对编程伙伴的可行性，（ii）调查拟人化对话代理的设计空间，（iii）确定人类程序员话语的意图并设计适当的代理响应，以及（iv）实现稳健和支持性的会话代理。通过整合各种交互机制（可视化、对话风格和代理行为）、最先进的机器学习算法、创造力和解决问题的策略，会话代理的拟人化属性将会出现。关键架构组件的紧密数据收集、设计、评估和细化周期将逐步发展功能，以实现健壮的编程会话代理。研究和教育活动的整合将有助于创建一个可靠而实用的结对编程代理，从而提高学生的参与度、学习能力和编程技能。本项目由以人为中心的计算（HCC）和促进竞争研究的既定计划（EPSCoR）共同资助。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Voice-assisted devices showcase the potential of conversational virtual assistants that have human-like interactions. This research seeks to develop an Alexa-like programming partner to transform how programming is learned and how programming is done. The research draws on and expands underlying ideas from human-computer interaction, artificial intelligence, and software engineering to address the problem of designing a conversational agent that supports pair programming, where a human and agent engage in rich collaborative interactions on programming activities. Pair programming increases code quality, efficiency, knowledge transfer and self-efficacy, and through developing pair-programming partners, the work will impact software development as well as teaching and learning programming, while helping retain underrepresented groups in computer science programs. The project will advance programmer-computer interactions by creating a human-centered interactive partner that supports pair programming. Foundational research will be performed to: (i) understand the feasibility of a pair programming partner, (ii) investigate the design space of anthropomorphic conversational agents, (iii) determine the intent of human programmer utterances and design appropriate agent responses and (iv) implement a robust and supportive conversational agent. Anthropomorphic properties of the conversational agent will arise by integrating diverse interaction mechanisms (visualizations, dialogue styles, and agent actions), state-of-the-art machine learning algorithms, and creativity and problem-solving strategies. Tight data collection, design, evaluation and refinement cycles for key architectural components will incrementally evolve functionality to realize the robust conversational agent for programming. Integration of research and educational activities will contribute to the creation of a sound and practical pair programming agent that enhances student engagement, learning, and programming skills. This project is jointly funded by Human-Centered Computing (HCC) and the Established Program to Stimulate Competitive Research (EPSCoR).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.