SBIR Phase I: Empowering Distance Learning for CS/CE Students with Real-time, Personalized Feedback through Anti-Pattern Detection and Correction (COVID-19)

这项小型企业创新研究（SBIR）第一阶段项目的更广泛影响/商业潜力是通过自动反馈机制解决对计算机科学（CS）和计算机工程（CE）教育工作者日益增长的需求。CS和CE的课程正在迅速增长。该项目将开发一个利用机器学习提供自动反馈的平台。这将使课程在不影响质量的情况下扩大规模。特别是，这个项目将导致一个系统的发展，为学生提供实时的，个性化的反馈。这一制度也可能有助于吸引和留住在行政长官和行政长官内代表性不足的人口。该产品将大大提高所有学生对计算机的科学和技术理解。这个小企业创新研究（SBIR）第一阶段项目将创建一个新颖的智能辅导系统，为CS和CE学生提供实时、个性化的反馈。特别是，该项目将开发一种深度学习架构，可以检测学生代码中的错误，并根据语义和行为相似性将其聚类到反模式中。提出的研究将设计和调整深度学习架构，以达到足够的准确性和精度。要实现这一目标，必须克服几个技术障碍，包括抽象语法树分析和下游本地化。最终的任务是通过编辑路径从不正确的代码到正确的代码找到补丁，并以人类可读的方式报告这些修复。该项目将推进一个全面的学习管理系统，强调学生的学习和表现。特别是，该平台将允许学生提交课程材料并迅速获得个性化反馈。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。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project is to address the increasing demand placed on educators in computer science (CS) and computer engineering (CE) through an automated feedback mechanism. Courses in CS and CE are growing rapidly. This project will develop a platform providing automated feedback by leveraging machine learning. This will allow courses to scale without compromising quality. In particular, this project would lead to the development of a system that provides real-time, personalized feedback to students. This system may also help attract and retain under-represented populations within CS and CE. This product will greatly enhance the scientific and technical understanding of computing for all students.This Small Business Innovation Research (SBIR) Phase I project will create a novel intelligent tutoring system to provide real-time, personalized feedback to CS and CE students. In particular, this project will develop a deep learning architecture that detects errors in student code and clusters them into anti-patterns according to semantic and behavioral similarity. The proposed research will design and tune the deep learning architecture to reach sufficient accuracy and precision. To realize this goal, several technical hurdles must be overcome, including abstract syntax tree analysis and downstream localization. The ultimate task is to find patches through edit-paths from incorrect code to correct code and report these repairs in a human-readable way. This project will advance a comprehensive learning management system that emphasizes student learning and performance. In particular, the platform will allow for students to submit course material and rapidly receive personalized feedback.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.