About - Fun ML Learning

Our Philosophy

Most machine learning tutorials are filled with symbolic derivations and mathematical proofs that scare away many people curious about AI. We believe that understanding an algorithm doesn't require memorizing its formulas first—you just need to see it running and feel what it does.

Every algorithm on this site runs directly in the browser with real-time parameter tuning and live visualization. We help you build intuition first, then provide concise mathematical descriptions. Interest is the best teacher.

Visualization First

See, Then Understand

Each algorithm comes with real-time animated demos to help you naturally build intuition while watching.

Zero-Formula Onboarding

Formulas Serve Understanding

Formulas appear only when necessary, and every symbol has an intuitive explanation.

Runs in Browser

No Installation Required

All JS code runs directly in the browser. Python code can be copied and run locally.

Dual-Language

JS + Python Dual Versions

Each algorithm provides both JavaScript (browser-side) and Python implementations for side-by-side learning.

Algorithm Directory

Algorithm	Type	Implementation
Gradient Descent	Optimization Basics	Hand-written
Linear Regression	Supervised Learning	Framework
Logistic Regression	Supervised Learning	Framework
KNN	Supervised Learning	Hand-written
Decision Tree	Supervised Learning	Hand-written
SVM	Supervised Learning	Hand-written
K-Means	Unsupervised Learning	Hand-written
PCA	Dimensionality Reduction	Hand-written
MLP	Neural Network	Framework
CNN	Neural Network	Framework
LSTM	Sequence Model	Framework
Word2Vec	Word Embedding	Framework
Attention	Sequence Model	Framework
Transformer	Sequence Model	Framework
Q-Learning	Reinforcement Learning	Hand-written
DQN	Reinforcement Learning	Framework
GAN	Generative Model	Framework
VAE	Generative Model	Framework
Diffusion Model	Generative Model	Framework

Hand-written Pure JS / Pure Python, no framework dependencies Framework TensorFlow.js / PyTorch / TensorFlow

Tech Stack

FrontendAlpine.js + Pico.css (minimalist MPA, no bundler)

VisualizationECharts · PixiJS v7 (high-performance real-time demos)

Code EditorMonaco Editor (VS Code core in the browser)

ML Framework (JS)TensorFlow.js — neural networks run directly on the browser GPU

ML Framework (Python)PyTorch · TensorFlow (downloadable for local use)

BackendGo / Gin — only for analytics; all pages are fully static

FAQ

What math background do I need?

High school math (basic concepts of functions and derivatives) is enough to get started. Each chapter builds from intuition first—formulas are just tools to aid understanding.

Do I need to know Python beforehand?

No. All interactive demos use JavaScript and run in the browser with no installation. The Python versions are provided for readers with programming experience who want to compare.

Can I modify the code?

Yes. The code editor on every page supports direct editing and running. Change parameters, modify logic, and see the effects instantly.

Is completing this site enough?

This site aims to build your intuition and interest, clearing the path for deeper textbooks (like "Deep Learning" or "Dive into Deep Learning"). Once you have a solid foundation, advanced materials will be much easier to approach.

Found a bug or have a suggestion?

Feel free to reach out via the contact links in the top navigation bar, or open an Issue directly on GitHub.