Cyberlearning: Sensei: High-Fidelity, Non-Invasive Classroom Sensing for Professional Development

多年来，研究表明，放弃大型讲座，增加学生在课堂上的参与度和参与度，可以显著提高学习效果。不幸的是，教授缺乏高质量的专业发展机会来提高他们的教学水平，而且通常没有接受过如何教学的培训。这个项目是通过一种名为“老师”的网络学习创新来解决大学教授的专业发展问题。Sensei使用传感器捕捉、隔离和分析语音和视频，提供课堂互动的近实时数据，如学生与教授交谈的时间百分比，学生与学生交谈的时间百分比，通过面部表情分析的学生参与度，学生与教授之间的轮流等等，所有这些都涉及语音和视频的多模态分析。本研究的第二个组成部分是发展建议的行动，以提高教授作为教师的表现。更准确地说，Sensei借鉴了传感阵列、计算机视觉、智能环境和个人信息学方面的技术和社会技术进步，以及高等教育专业发展框架。在这个项目中，研究人员将1)开发自动感知和显示反馈给教师所需的技术，2)将该系统部署到大学教师体内，在一系列基于设计的研究中使用该系统，并对结果进行解释，3)迭代我们的框架，以便将课堂数据常规地纳入专业发展。这项研究是通过网络创新实现的，其中计算通过最先进的多模态传感方法的能力进行扩展，以实现课堂规模的非侵入式传感。这种网络创新推动了一种学习创新，通过在可管理的教学单元中提供快速、频繁的反馈和指导，支持以学生为中心的信念，在综合反思和培训系统中提供近乎实时的教学实践数据。反过来，学习创新通过探索专业发展框架中的机制，促进了对教师如何在技术丰富的学习环境中学习的理解，而没有这种新的网络学习类型，这些机制是不可能实现的。特别是，通过与教授STEM大学课程的教师进行一系列基于设计的研究，研究人员探索了Sensei a)可以触发批判性自我反思的方式，b)这种自我反思如何根据所查看数据的特征发生变化，c)数据驱动的目标设定如何促进教学中的自我效能感，以及d)这些影响如何随时间变化。所有的代码都将作为开源开发，如果成功，Sensei可以推广到包括K-12教师。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。

For years, research has shown that moving away from large lectures and increasing student engagement and participation in classrooms significantly improves learning. Unfortunately, professors lack quality professional development opportunities to improve their instruction, and typically receive no training on how to teach. This project is addressing the issue of college professors' professional development through a cyberlearning innovation called Sensei. Sensei has novel capabilities using sensors to capture, isolate, and analyze voice and video that will provide near real time data on classroom interactions such as the percent time students talk vs professors, the percent time students talk to students, student engagement through facial expression analysis, turn taking between students and professors, etc all of which involve multimodal analysis of voice and video. The second component of this research is the development of suggested actions to improve the professor's performance as a teacher.More precisely, Sensei draws on technical and socio-technical advances in sensing arrays, computer vision, intelligent environments, and personal informatics, as well as frameworks of professional development in higher education. In this project the researchers will 1) develop the technologies needed to automatically sense and display feedback to instructors, 2) deploy this system in-vivo to college instructors over semesters of use in a series of design-based research studies, and interpret the results to 3) iterate on our framework for the routine incorporation of classroom data into professional development. This research is enabled by a cyber innovation in which computing is expanded by the capabilities of state of the art multimodal sensing approaches to achieve non-invasive sensing at classroom-scale. This cyber innovation drives a learning innovation of delivering near-real-time data on teaching practices in a combined reflection and training system by delivering rapid and frequent feedback and instruction on good strategies in manageable instructional units, that support a focus on student-centered beliefs. In turn, the learning innovation advances understanding of how instructors learn in technology-rich learning environments by exploring mechanisms in a framework of professional development that would not be possible without this new cyberlearning genre. In particular, through a series of design-based research studies with instructors teaching STEM college courses, the researchers explore ways in which Sensei a) can trigger critical self reflection, b) how this self-reflection changes based on the features of the data viewed, c) how datadriven goal-setting can foster self-efficacy in teaching, and d) how these effects vary over time. All of the code will be developed as open source and, if successful, Sensei could be generalized to include K-12 teachers.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.