NeuroNomixer logoNeuroNomixer
  • Home
  • Blog
  • Visual Guides
  • Authors
  • Contact
Sign InSign Up
NeuroNomixer logo
HomeBlogAuthorsContactPrivacy Policy

© 2026 NeuroNomixer — Built with Next.js & Tailwind CSS

Visual Guides/Gradient Descent: Rolling Down the Hill
Machine Learning

Gradient Descent: Rolling Down the Hill

Click anywhere on the loss landscape to drop a ball, then watch it roll toward the minimum. Experiment with different learning rates and landscapes to see convergence, oscillation, and getting stuck.

Exploration Progress0/3 landscapes · 0/5 runs

Sign in to save your progress.

Loss landscape(click to set starting position)
w (weight)b (bias)min

Loss over Steps

Click on the landscape to start

Landscape

Learning Rate α

0.100

Too small → slow. Too large → oscillates or diverges.

Controls

Key Insight

The gradient points uphill, so subtracting it moves downhill. The learning rate controls step size: too large and you overshoot the valley.

← All GuidesNext Guide →