Skip to content

CHYYYYYYYY/VidTailor

Repository files navigation

VidTailor

Personalized Tutoring System Based on Video Content Understanding

Project Duration: June 2024 - June 2025

Overview

VidTailor is an intelligent video learning platform that generates personalized practice questions by analyzing students' viewing behaviors, providing precise learning feedback and a social learning experience.

Core Features

1. Intelligent Question Generation

  • Behavior-Driven Analysis: Real-time identification of learning difficulties based on viewing behaviors (fast-forward, slow-motion, pause, replay, etc.)
  • Personalized Questions: Automatically generates practice questions tailored to individual learning levels and weak points
  • Instant Feedback: Provides real-time feedback to help students test and reinforce their understanding immediately

2. Smart Error Book & Note System

  • Video Segment Linking: Precisely associates incorrect answers with corresponding video segments for one-click review
  • Knowledge Point Highlighting: Automatically records errors and highlights key concepts to build a personal knowledge gap database
  • Synchronized Notes: Smart note-taking feature addresses cognitive blind spots, improving review efficiency and knowledge system completeness

3. Social Learning Interaction

  • Smart Matching: Recommendation algorithm pairs learners with similar questions or challenges
  • Topic Forums: Guides students into relevant discussion groups to share insights and problem-solving approaches
  • AI Assistant: Intelligent chat assistant helps discuss difficult problems and pushes relevant knowledge points
  • Learning Community: Creates a collaborative learning atmosphere to boost motivation and engagement

Design Methodology

User Research & Iteration

  • Multi-Round Prototype Testing: Scenario demonstration videos → Low-fidelity prototype interviews → Questionnaire surveys
  • A/B Testing: Compares different interaction prompt schemes to optimize the balance between timeliness and distraction
  • Heuristic Evaluation: Continuously collects user feedback to iterate and optimize features and interface design

Validation Results

  • Significantly improved system usability (SUS satisfaction scores, task completion rates, etc.)
  • Validated the effectiveness of personalized learning paths and social learning models
  • Provided clear direction for future feature iterations

Problems Solved

Pain points of traditional video learning platforms:

  • ❌ Lack of supporting practice questions, unable to test learning outcomes promptly
  • ❌ Cannot provide personalized targeted exercises
  • ❌ Learning discussions scattered across various chat apps, lacking systematization

VidTailor's solutions:

  • ✅ Intelligent question generation system based on behavior analysis
  • ✅ Personalized learning paths with instant feedback
  • ✅ Integrated social learning community

Technical Highlights

  • Video content understanding and knowledge point extraction
  • Learning behavior analysis algorithms
  • Personalized recommendation engine
  • Intelligent Q&A and collaboration system

Project Lead: Yue Li

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages