Open Source AI Hackathon 2024

GenAI makers and creators contest and showcase

Tickets

Loading…

Gagan Gupta

Gagan Gupta

@gagangupta

Harshita Kamani

@harshitakamani

EduPrepAI

Submitted Jan 28, 2024

Check for latest updates:
https://gitlab.com/hkxgg/eduprepai/-/blob/main/README.md

Presentation:
http://surl.li/rqmik

Demo:
https://youtu.be/ikYq7MhHAW8?si=30Fqnghp9Om54ggK

Problem Research 📚

We conducted extensive research and found that students often struggle to find relevant practice questions, leading to a lack of practice, while teachers face challenges in creating personalized content. Students tend to rely on summary videos for exam preparation, which may not sufficiently cover the material. To address these issues, we propose a solution that offers quick summarization, interactive quizzes, and personalized content generation based on educational videos.

Problem Statement 🎯

  • Watching lengthy lecture videos is time-consuming and tedious.
  • Students struggle to find relevant questions for practice.
  • Teachers spend considerable time creating quizzes and notes manually.
  • Limited access to personalized content affects students’ exam preparation.

Solution Overview 💡

Our innovative EdTech platform aims to enhance the learning experience by converting educational video lectures into summary, lecture notes and engaging MCQ-based quizzes and providing automatic assessment. This project not only streamlines the quiz creation process for teachers but also offers students a dynamic and interactive learning environment. With features such as instant grading and performance analytics, our platform facilitates efficient and effective learning.


Roadmap 🛣️

  1. Scraping Data [Successful]

  2. Data Collection and Selection [Successful]

    • Utilize educational videos from open access university websites and YouTube
  3. Frame Extraction [Successful]

    • Extract frames using ffmpeg
  4. Labeling Dataset [Successful]

    • Label frames as slide, presenter slide, and others
  5. Model Training [Ongoing]

    • Base Models: Resnet50, YOLO v9, EAST, LLama2-7b
    • Future Scope: Explore state-of-the-art models
  6. Image Tuning Operations [Successful]

    • Crop Transform, Border Removal, Segment Clustering, Image Hashing
  7. Sequential Data Context Search Training [Successful]

  8. OCR, Figure Detection [Successfull]

  9. Transcription (Multimodal) [Successful]

    • Audio to Transcript Generation
  10. Combination and Sync [Successful]

  11. Text Operations [Successful]

    • Spell Check, etc.
  12. Summarization and Quiz Generation

    • Utilize LLama2 finetuning and pre-existing models like BERT-edu
    • Answer Key/Options Generation using Sense2vec word vectors

Comments

{{ gettext('Login to leave a comment') }}

{{ gettext('Post a comment…') }}
{{ gettext('New comment') }}
{{ formTitle }}

{{ errorMsg }}

{{ gettext('No comments posted yet') }}

Hybrid access (members only)

Hosted by

The Fifth Elephant hackathons

Supported by

Host

Jump starting better data engineering and AI futures

Venue host

Welcome to the events page for events hosted at The Terrace @ Hasura. more

Partner

Providing all founders, at any stage, with free resources to build a successful startup.