Drag to rotate view. Click on surface to move point.
Arrow keys nudge position. R to reset view.
💡 This bowl-shaped surface has minimum at the origin. Both partial derivatives are zero there!
🤖 ML Insight: x and y represent two neural network weights. The height is the loss. ∂L/∂x tells you how to adjust weight x; ∂L/∂y tells you how to adjust weight y. Gradient descent moves opposite to both!