AI-Powered Learning: Increasing User Retention and Productivity Through Personalised Learning Paths

人工智能最近的创新在所有行业都掀起了颠覆性的浪潮，特别是在教育领域。世界各地的学校都禁止使用ChatGPT，担心它会影响学习，而教育科技初创公司则抓住了降低教育门槛的机会。线下和在线学习之间的鸿沟越来越大，学校和高等教育被迫提供混合学习方法。知识已经成为一种商品。为了保持相关性和经济竞争力，学习机构需要调整，重新将教学方法的重点不是知识本身，而是其交付机制，以便学生在学习中茁壮成长，在遇到困难时得到支持，并受到诱惑继续学习。该项目由音乐技术在线学习平台Music Hackspace和伦敦艺术大学创意计算研究所牵头。它的重点是研究个性化学习之旅的可行性，利用现有的由Music Hackspace在2020至2023年间创建的300多门点播课程的资源库。该项目将分析音乐Hackspace课程的成绩单，帮助学生发现新内容并决定学习什么，使用一个对话工具，根据学生的需求和背景提供定制的建议。两个团队已经在2021年合作获得了IUK的拨款，建立了一个机器学习引擎，根据课程的复杂程度对课程进行排名。这个项目是当时开始的成功研究的延续，但没有使用大型语言模型。两个团队拥有现有的代码库和原型，明确的知识产权战略，并习惯于合作。这个项目利用了三个关键组成部分：(1)丽贝卡·菲布林克教授团队的世界级机器学习和音乐技术专业知识，(2)学者和专业人员在线和面对面教学的经验，(3)一家蓬勃发展的初创公司，非常适合整合这项研究的结果，以获得吸引力和市场份额。Statista估计，全球电子学习市场将增长9.84%(2023：GB 133B，2027：GB 191b)，虽然全球音乐电子学习领域的增长速度快了一倍(2023年：GB 174M，2027：GB 341M/CAGR：18.40%，Research&Markets)，但这突显了英国在这一快速增长领域的地位的迫切需要。机器学习是区分这一领域创新的关键。该项目的成功完成预计将使MHS的收入比目前的预测增加62%，突显出人工智能在教育领域的变革潜力。

The recent innovations in Artificial Intelligence have sent disrupting waves across all industries, and across the Education sector in particular. Schools around the world have banned the use of ChatGPT, fearing its impact on learning, while edtech startups embraced the opportunity to lower barriers to education. The divide between offline and online learning grows stronger, and schools and higher education are forced to offer hybrid approaches to learning.Knowledge has become a commodity. To remain relevant and economically competitive, learning institutions need to adapt, and refocus their pedagogical approach not on the knowledge itself, but on its delivery mechanisms, such that students will thrive in their learning, be supported when encountering difficulties and be enticed to continuously learn.This project is led by Music Hackspace, an online learning platform for music technologies, and the Creative Computing Institute of University of the Arts London. It focuses on studying the feasibility of personalised learning journeys, using an existing repository of over 300 on-demand courses created by Music Hackspace between 2020 and 2023\. The project will analyse the transcripts of Music Hackspace courses to help students discover new content and decide what to study, with a conversational tool that offers bespoke advice based on student needs and contexts.Both teams have collaborated on a IUK grant in 2021, to build a Machine Learning powered engine to rank courses by complexity. This project is a continuation of the successful research started then, but which didn't use large language models. The teams have an existing codebase and prototype, a clear IP strategy, and are used to working together.This project leverages three key components: (1) the world-class Machine Learning and music technology expertise of Prof Rebecca Fiebrink's team, (2) the experience of teaching online and in-person of academics and professionals, (3) a thriving startup ideally positioned to integrate the results of this research to gain traction and market share.Statista estimates the global e-learning market to grow CAGR 9.84%(2023:£133B,2027:£191B), while the global music e-learning segment grows twice as fast (2023: £174M,2027:£341M/CAGR:18.40%, Research&Markets), highlighting an urgent need to bolster the UK position in this rapidly growing sector. Machine Learning is key to differentiating innovation in this space. Successful completion of this project is projected to increase MHS's revenue by 62% over its current forecast, underscoring the transformative potential of AI in education.