Skip to content

Add interactive Vietnamese ML lesson on Perceptron fundamentals - #2

Merged
LMK89 merged 1 commit into
mainfrom
claude/intelligent-davinci-4h9mh3
Aug 21, 2026
Merged

Add interactive Vietnamese ML lesson on Perceptron fundamentals#2
LMK89 merged 1 commit into
mainfrom
claude/intelligent-davinci-4h9mh3

Conversation

@LMK89

@LMK89 LMK89 commented Aug 21, 2026

Copy link
Copy Markdown
Owner

Summary

This PR introduces a comprehensive, interactive educational resource on Perceptron neural networks in Vietnamese. The lesson is delivered as a self-contained HTML file with embedded CSS and JavaScript, requiring no external dependencies or build process.

Key Changes

  • New lesson file: 2026-08-20-perceptron-cd2c084e.html — A complete interactive tutorial covering:

    • Perceptron definition and mathematical formulation
    • Real-world problem: orange ripeness classification using two features (Brix sweetness, color)
    • Four distinct training states visualized with interactive plots showing decision boundaries
    • Live training demo where users can adjust learning rate and step through epochs
    • Prediction demo with step-by-step calculation display
    • Data quality guidance (normalization, label noise, class imbalance)
    • Practice exercises and self-grading quiz
  • New README: Claude-lession/README.md — Documents the lesson series structure and naming convention (YYYY-MM-DD-<topic>-<random-hex>.html)

Implementation Details

  • Responsive design: Uses CSS custom properties for theming with automatic dark mode support
  • Canvas-based visualization: Decision boundary plots render dynamically based on model weights, showing correctly/incorrectly classified points
  • Interactive training loop: Implements the perceptron learning algorithm in JavaScript with epoch-by-epoch visualization
  • Data normalization: Includes z-score standardization to handle feature scaling issues
  • Educational progression: Moves from theory → concrete example → visualization → hands-on interaction → reflection on limitations (XOR problem, linear separability)
  • Accessibility: Semantic HTML, ARIA labels, keyboard navigation support, reduced-motion preferences respected

The lesson emphasizes practical understanding through visualization and interaction rather than pure theory, making it suitable for beginners new to neural networks.

https://claude.ai/code/session_01NDr1jZCzdzHU9MH6cNZKSc

Tạo thư mục Claude-lession cho các bài giảng AI/ML tiếng Việt dạng
HTML tương tác, kèm README mô tả quy ước đặt tên và cấu trúc bài.

Bài đầu tiên (mã cd2c084e) — Perceptron:
- 4 trạng thái huấn luyện (khởi tạo → 1 epoch → hội tụ → XOR), mỗi
  trạng thái có biểu đồ canvas vẽ đường ranh giới và điểm đoán sai
- Bảng so sánh Input/Xử lý/Output/Đánh giá + pros/cons
- Demo huấn luyện trực tiếp: chạy từng epoch, chỉnh learning rate,
  tự dừng khi hội tụ
- Demo dự đoán: nhập độ ngọt + độ vàng, hiện đủ các bước tính z
- Mục dữ liệu tốt vs xấu (lệch thang đo, nhiễu nhãn, mất cân bằng,
  rò rỉ đặc trưng) kèm sơ đồ ASCII
- Bài tập tính tay, bài tập code, quiz 4 câu tự chấm có giải thích
- Dark mode theo theme người xem, responsive, tôn trọng
  prefers-reduced-motion

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NDr1jZCzdzHU9MH6cNZKSc
@LMK89
LMK89 merged commit 51647da into main Aug 21, 2026
@LMK89
LMK89 deleted the claude/intelligent-davinci-4h9mh3 branch August 25, 2026 07:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants