STTR Phase I: Enabling Student Project Collaboration with Artificial Intelligence Augmented Mentorship

这个小企业技术转移第一阶段项目的更广泛/商业影响是通过相互依赖的学习经验改善学生的学习和劳动力的准备。该项目将创造基于项目的环境，促进沟通、批判性思维、解决问题、时间管理、创造力和团队合作等技能——所有这些都反映了专业的工作环境。该技术还将促进项目管理、写作、商业和数据分析、设计和演示等技能的发展。项目协作要求学生在项目完成过程中频繁互动，包括频繁的导师或老师互动。这样一个相互依存的环境创造了一个现实世界的动态，更好地为学生进入劳动力市场做好准备。该项目开发的平台在参与增长最快的电子学习领域的同时，可能会产生重大的社会影响。该提案旨在开发一个协作社区平台，使用专有项目协作模型与人工智能（AI）增强指导相结合，以提高学生的劳动力准备程度。该技术将被设计为在正确的时间为学生和导师提供正确的信息。通过分析和统一领域特定语义表示下的所有内容，系统将能够聚合和组织所有内容，并识别上下文最有用的干预部分。为了使学生和导师之间的项目协作和指导更容易，以一种值得信赖的方式，建模将利用最小的监督来完成。该建模将包括将语言模型中的上下文嵌入与基于图的神经网络相结合，以捕获跨多个方面的交互。该技术将建立在深度神经网络的可解释性之上，为决策提供适当的透明度，既可以让用户学会信任平台，也可以让平台开发人员构建有助于可靠、值得信赖和公平指导的系统。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/commercial impact of this Small Business Technology Transfer Phase I project is in improving both student learning and workforce readiness through interdependent learning experiences. The project will create project-based environments that promote skills such as communication, critical thinking, problem solving, time management, creativity, and teamwork – all mirroring professional work environments. The technology will also promote development of skills such as project management, writing, business and data analysis, design, and presentation. Project collaboration requires that students interact frequently throughout the project completion process, including frequent mentor or teacher interactions. Such an interdependent environment creates a real-world dynamic that better prepares students to enter the workforce. The platform developed by this project is likely to create significant societal impact while participating in the fastest growing e-learning sector.The proposal seeks to develop a collaborative community platform using proprietary project collaboration models integrated with Artificial Intelligence (AI) augmented mentorship to enhance student workforce readiness. The technology will be designed to provide the right piece of information to the students and mentors at the right time. By analyzing and unifying all the content under a domain-specific semantic representation, the system will be able to aggregate and organize all the content and identify the piece for intervention that is contextually most useful. To make project collaboration and mentorship easier between students and mentors in a trustworthy manner, modeling will be done utilizing minimal supervision. This modelling will include combining contextual embeddings from language models with graph-based neural networks to capture interactions across multiple facets. The technology will build upon explainability of deep neural networks to provide an appropriate level of transparency into the decision making, both for the users to learn to trust the platform, as well as for the platform developers to build systems that aid in reliable, trustworthy, and fair mentoring.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.