Collaborative Research: Using AI-enabled Smart Objects to Understand and Support Spatial Reasoning and Learning

空间思维是理解和使用有关地点的信息来解决问题的能力。空间思维对STEM学科的成就至关重要。幸运的是，空间思维是可塑性的，它对需要人们使用空间技能的训练和生活经历做出反应。这项研究将开发和评估一个有助于幼儿空间思维的人工智能系统。该系统依靠一种久经考验的方法来提高儿童的空间技能：积木游戏。最大的进步是这些积木将是“智能的”，因为它们可以通过跟踪移动、加速度、旋转以及不同对象彼此之间的放置和处理方式来检测儿童之间的互动。通过分析儿童的建筑模式，研究人员将使用机器学习来突出积木的模式，并在儿童建造时向他们建议可能导致更高水平的空间思维的策略。这一结果可能会支持我们支持幼儿空间思维的方式发生革命性变化，最终支持STEM学习。该项目开发了一种革命性的工程概念，赋予儿童玩具和日常教室/游戏室物品以智能和互联网连接。这类物体的例子包括积木、玩具、拼图等。目标是让这些玩具和物体准确地记录作为他们游戏和学习过程的一部分的儿童如何单独和小组与他们互动。这种数据是以非常高的时间和空间分辨率无线记录和收集的。该项目还开发了一系列基于新兴人工智能的算法，以分析智能对象生成的数据，以找出儿童处理这些对象的特定模式。然后对这些模式进行分析，以评估空间思维和空间智力，以及它们如何在不同的训练方案下随时间演变。研究人员将开发嵌入式计算工具，包括硬件和软件，以及低功率无线通信机制，以实现智能对象这一转变的工程概念。从长远来看，他们设想将无处不在的智能引入许多日常物品中，以便可以无缝地跟踪和分析空间情报。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Spatial thinking is the ability to understand and use information about locations to solve problems. Spatial thinking is critically important to achievement in STEM disciplines. Fortunately, spatial thinking is malleable, it responds to training and to life experiences that require people to use their spatial skills. This research will develop and assess an AI system that will facilitate young children’s spatial thinking. The system relies on a time-tested method for improving children’s spatial skills: block play. The major advancement is that the blocks will be “smart”, in that they can detect child-interactions by tracking movement, acceleration, rotation, and how different objects are placed and handled with respect to each other. By analyzing children’s building patterns, the researchers will use machine learning to highlight patterns of block building and to suggest strategies to children as they build that may result in higher levels of spatial thinking. The results may support a transformative change in how we support spatial thinking, and ultimately STEM learning, in young children.The project develops a transformative engineering concept of endowing smartness and internet connectivity to children’s toys and everyday classroom/playroom objects. Examples of such objects include playing blocks, toys, jigsaw puzzle pieces, etc. The objective is for those toys and objects to precisely record how a child interacts with them individually and in a group as parts of their play and learning process. Such data is recorded and collected wirelessly at very high temporal and spatial resolutions. The project also develops a family of emerging Artificial Intelligence based algorithms to analyze the smart object-generated data to find out specific patterns in children’s handling of those objects. Such patterns are then analyzed for evaluating spatial thinking and spatial intelligence, and how they evolve over time under different training regimens. The investigators will develop embedded computing tools, both hardware and software, as well as low power wireless communication mechanisms for realizing this transforming engineering concept of smart objects. For the long term, they envision introducing ubiquitous smartness into many everyday objects so that tracking and analysis of spatial intelligence can be seamless and highly scalable.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.