From grade school basics to cutting-edge AI research. A complete journey through neural networks, residual connections, and the mathematics behind stable large-scale training.
Start LearningFive levels spanning from middle school foundations to graduate-level research. Switch between casual and formal explanations anytime. All visualizations work in both 2D and 3D.
What is AI? How do computers learn patterns? Start with the absolute basics and build intuition for machine learning.
Building artificial brains from simple math. Learn about neurons, vectors, and how information flows through networks.
Why deep networks break and how ResNet saved the day. Explore vanishing gradients and residual connections.
Spaces that curve and bend. Discover manifolds, topology, and why constraint matters in high-dimensional learning.
The complete paper explained. Manifold-Constrained Hyper-Connections: motivation, method, mathematics, and results.
Designed for learners at every level. Switch modes instantly to match your preferred learning style.
Read in Casual mode with everyday analogies and conversational explanations, or switch to Formal mode for precise academic language.
Play with neural networks, manipulate manifolds, and watch gradients flow. Choose between clean 2D Canvas or immersive 3D Three.js views.
Every concept includes working Python code and clear pseudocode. Run examples in your browser or copy to your IDE.
Start with intuitive explanations, then dive deeper. Every topic has both simple (ELI5) and complex (full mathematical) views.
All 8 figures from the mHC paper are explained interactively. Hover for annotations, click to explore, understand the research visually.
Jump between levels anytime. No gates, no prerequisites. Your progress is tracked but never blocks exploration.