YouTube is one of the most popular online platforms used for educational purposes as it has a vast repository of educational content available openly. It makes learning easy by explaining concepts with the help of visuals and animations, unlike textbooks and reference books that use static text and images. However, YouTube poses challenges such as a lack of real-time interaction and doubt-clarification, passive learning, and distraction from the learning path as YouTube is unrestricted to only educational content. To overcome these challenges, we propose the Sensei application, which is a browser extension that redirects users to a Streamlined web application and aims to improve their learning experience. It uses Llama2, a Large Language Model that powers the chatbot for real-time doubt clarification and also provides prerequisites and personalized quizzes. Whisper generates a transcript of the YouTube video which is used to fine-tune the chatbot. This allows Sensei to offer relevant answers based on the video being viewed. Sensei also provides similar video and course recommendations using YouTube Data API and Coursera Course API respectively. The results show high user satisfaction, generation of accurate transcript, and context-related quiz with unique and relevant options for each question. The performance of Sensei was tested across various browsers and was found to be good. Future enhancements can increase the effectiveness of this application. Overall, Sensei is a personalized, and interactive educational platform that allows users to resolve their queries independently.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Sensei: Generative AI Based YouTube Study Assistant

  • Deepak S. Khachane,
  • Rushikesh R. Nikam,
  • Chetan Kumbhar,
  • Vaishnavi Kothari,
  • Anwesha Pani,
  • Aastha Patil

摘要

YouTube is one of the most popular online platforms used for educational purposes as it has a vast repository of educational content available openly. It makes learning easy by explaining concepts with the help of visuals and animations, unlike textbooks and reference books that use static text and images. However, YouTube poses challenges such as a lack of real-time interaction and doubt-clarification, passive learning, and distraction from the learning path as YouTube is unrestricted to only educational content. To overcome these challenges, we propose the Sensei application, which is a browser extension that redirects users to a Streamlined web application and aims to improve their learning experience. It uses Llama2, a Large Language Model that powers the chatbot for real-time doubt clarification and also provides prerequisites and personalized quizzes. Whisper generates a transcript of the YouTube video which is used to fine-tune the chatbot. This allows Sensei to offer relevant answers based on the video being viewed. Sensei also provides similar video and course recommendations using YouTube Data API and Coursera Course API respectively. The results show high user satisfaction, generation of accurate transcript, and context-related quiz with unique and relevant options for each question. The performance of Sensei was tested across various browsers and was found to be good. Future enhancements can increase the effectiveness of this application. Overall, Sensei is a personalized, and interactive educational platform that allows users to resolve their queries independently.