REU Site: Interdisciplinary AI Research for Undergraduates

REU网站的总体目标是为本科生提供研究经验，从而在跨学科人工智能领域发表可发表的工作，并教育学生人工智能（AI）的伦理。人工智能正在对我们的日常生活产生越来越大的影响，无论是通过与智能扬声器和数字助理的互动、防止碰撞、发现新药、智能辅导、谷歌翻译，还是其他更多的途径。人工智能也是最具争议的新技术之一，因为许多人对其作为生存威胁的潜在风险提出了批评，它被专制政府用作压迫工具，被法官用于量刑决定，以及由于基于人类判断的数据集而产生的偏见。学生和公众必须意识到对我们未来的承诺和潜在的影响，无论是好是坏。此外，2016年世界经济论坛的一份报告《2018年工作的未来》估计，到2022年，人工智能将创造5800万个新工作岗位。因此，必须招募更多的学生，更多样化的学生，并对他们进行人工智能方法和道德方面的培训。该项目将在三年内每年培训11名学生，其中一半来自当地社区大学，为他们在负责任的人工智能领域的职业生涯做好准备。该项目每年将培训11名跨学科人工智能学生，以及人工智能研究的伦理，包括其潜在的好处和危险。此外，这种方法是合作的，而不是竞争的，因为现代人工智能的发现将由愿意与其他学科的科学家合作的研究人员做出，以便能够用最好的工具来解决问题，无论这些问题是由什么学科创造的。向学生介绍这种人工智能的协作方法是使他们为21世纪的科学实践做出贡献的必要步骤。该项目的一个关键方面是，它是一个为期30周的学年计划，给受训者必要的时间来吸收材料，思考项目，开发方法，并测试它们。该项目将从人工智能训练营开始，该训练营将在本学年的前八个周六举行，为学生们提供梯度下降及其在深度学习中的应用的速成课程。学生们将从大量搭建的初始项目开始，这将使他们初步尝到成功的滋味。从那里，他们将把获得的技能应用到加州大学圣地亚哥分校活跃的研究实验室中进行的跨学科项目中。项目导师包括计算机科学与工程系、电气与计算机工程系、认知科学系、语言学系、心理学系和上下文机器人研究所的教师，他们在研究中使用人工智能。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 overall aim of the REU Site is to provide undergraduate students with a research experience that leads to publishable work in interdisciplinary Artificial Intelligence, as well as educate students in the ethics of Artificial Intelligence (AI). AI is having increasing impact in our everyday lives, whether through interacting with smart speakers and digital assistants, preventing collisions, discovering new medicines, intelligent tutoring, Google Translate, and many more avenues. AI is also one of the most controversial new technologies, as many critiques have been made concerning its potential risk as an existential threat, its use as a tool of oppression by authoritarian governments, its use in sentencing decisions by judges, and its biases due to datasets based on human judgments. It is imperative that students and the public become aware of the promise and the potential impacts, good or bad, on our future. In addition, a 2016 World Economic Forum Report “The Future of Jobs 2018” estimates that there will be 58 million new jobs in AI by 2022. Hence, more students, and more diverse students, must be recruited and trained in AI methods and ethics. This project will train eleven students per year for three years, including half from local community colleges, to prepare them for careers in responsible AI.This project will train eleven students per year in interdisciplinary Artificial Intelligence, as well as the ethics of AI research, including its potential benefits and dangers. In addition, the approach is collaborative, rather than competitive, research, as modern AI discoveries will be made by researchers who are willing to work with scientists from other disciplines in order to be able to approach problems with the best tools available no matter what discipline created them. Introducing students to this collaborative approach to Artificial Intelligence is a necessary step to prepare them to contribute to science as it is practiced in the 21st century. A crucial aspect of the project is that it is a thirty week academic year program, giving the trainees the necessary time to absorb the material, think about the project, develop approaches, and test them. The program will begin with an AI Boot Camp, which will meet the first eight Saturdays of the academic year, giving the students a crash course in gradient descent and its application in deep learning. The students will begin with heavily scaffolded starter projects that will give them an initial taste of success. From there, they will apply acquired skills to interdisciplinary projects being carried out in active research labs at UC San Diego. The project mentors include faculty who are using AI in their research in the departments of Computer Science and Engineering, Electrical and Computer Engineering, Cognitive Science, Linguistics, Psychology, and the Contextual Robotics Institute.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.