EAGER: Cultivating Scientific Mindsets in the Machine Learning Era

人工智能（AI）已经融入到日常生活中。目前，人工智能在未来的劳动力中存在巨大的技能缺口。K-12学生有限的AI学习机会和教师的专业发展机会可能导致劳动力和教育中的AI不平等。该项目通过引入机器学习（ML）作为K-12 STEM教室中数据驱动的科学探究的发现工具来解决这些挑战。该项目的重点是创建和研究一个新的无编程和基于视觉的ML驱动的学习环境。它旨在使数学、编程和数据技能有限的高中学生和教师能够发现复杂的科学现象，并从隐藏在现实世界数据中的发人深省的模式中提出重大问题。研究人员将在整个研究活动中包括来自STEM边缘化背景的高中生，并与大卫T.罗切斯特大学卡恩斯多样性和领导力中心。该项目将有助于NSF促进下一代STEM教育包容性的使命，并推动K-12人工智能素养成为国家繁荣的推动力。 研究人员将开展三项协同研究活动：（1）创建一个ML驱动的视觉学习环境，利用基于字形的新型数据可视化和类比学习过程的组合，以缓解ML和多维模式发现的陡峭学习曲线。（2）采用共同设计的方法，让K-12 STEM教师和数据科学专家参与创建ML驱动的科学探究活动;以及（3）迭代地评估新学习环境在支持高中生的三个关键学习目标方面的有效性：多维模式发现，ML概念和聚类和分类方法，以及通过提问、假设生成、解释和论证进行模式启发的科学探究。该项目的研究结果将推进我们对ML驱动的视觉学习环境的设计和教学指南的了解，这些环境可以最大限度地减少K-12 AI初学者的认知负荷。该奖项反映了NSF的法定使命，并被认为值得通过使用基金会的智力价值和更广泛的影响审查标准进行评估来支持。

Artificial Intelligence (AI) is woven into the fabric of everyday life. There is currently an enormous skills gap in AI for the future workforce. Limited AI learning opportunities among K-12 students and professional development opportunities for teachers may lead to AI inequity in the workforce and education. This project addresses these challenges by introducing Machine Learning (ML) as a discovery tool for data-driven scientific inquiry in K-12 STEM classroom. The project focuses on creating and studying a novel programming-free and visual-based ML-powered learning environment. It aims to enable high school students and teachers with limited mathematical, programming and data skills to discover complex scientific phenomena and ask big questions from thought-provoking patterns hidden in real-world data. Researchers will include high school students from marginalized backgrounds in STEM throughout the research activities and engage in outreach in collaboration with the David T. Kearns Center for Diversity and Leadership at the University of Rochester. This project will contribute to NSF’s missions on promoting inclusion in next-generation STEM education, and advance K-12 AI literacy as a driving force of national prosperity. Researchers will carry out three synergistic research activities: (1) creating a ML-powered visual learning environment that utilizes a combination of novel glyph-based data visualization and analogical learning process to mitigate the steep learning curve of ML and multi-dimensional pattern discovery for high school learners; (2) adopting a co-design approach to include K-12 STEM teachers and data science experts in creating ML-powered scientific inquiry activities; and (3) iteratively evaluating the effectiveness of the new learning environment in supporting three key learning goals for high school students: multi-dimensional pattern discovery, ML concepts and methods around clustering and classification, and pattern-inspired scientific inquiry through question asking, hypothesis generation, explanation and argument. Findings of this project will advance our knowledge on the design and pedagogical guidelines of ML-powered visual learning environments that minimize cognitive load for novice K-12 AI learners. The resulting novel learning environment, and ML-powered scientific inquiry activities will be made publicly available.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.