Home▸Machine Learning & Neural Networks▸Bayesian Optimization: Gaussian-Process Hyperparameter Tuner

🎯 Bayesian Optimization: Gaussian-Process Hyperparameter Tuner

Watch a Gaussian-process surrogate model and an Expected-Improvement acquisition function hunt for the best hyperparameter with far fewer trials than grid search.

Machine Learning & Neural Networks3DAdvanced60 FPS

Details

CategoryMachine Learning & Neural Networks

RenderingThree.js · WebGL

DifficultyAdvanced

Target FPS60

ControlsGP lengthscale slider controls the surrogate's smoothness assumption; exploration ξ slider tunes the Expected Improvement trade-off between exploiting the current best and exploring uncertain regions; Sample next (EI) button evaluates the noisy objective at the current EI-maximizing candidate and adds it to the trial history; Show true function toggle reveals the hidden ground-truth curve; Reset restarts with fresh random trials.

Three.jsBayesian optimizationGaussian processhyperparameter tuningacquisition functionInstancedMesh

📖 Read the articleThe maths behind Bayesian Optimization: Gaussian-Process Hyperparameter Tuner

Related simulations

Decision Tree LiveCART decision tree with Gini impurity splitting; axis-aligned splits emerge level by level…

Reinforcement LearningQ-learning agent navigates an 8×8 grid maze. Heatmap shows max Q-values; white arrows show…

Self-Organising Map24×24 Kohonen SOM (576 neurons, 3D RGB input). Gaussian neighbourhood update as α and σ… ⊞ All 165 in Machine Learning & Neural Networks →

⚙ Under the hood

Tune a Gaussian-process surrogate and an Expected-Improvement acquisition function to hunt for the best hyperparameter value. Adjust the kernel lengthscale and exploration term, sample the next suggested trial, and watch the posterior mean, uncertainty band and acquisition surface update live.

Three.jsBayesian optimizationGaussian processhyperparameter tuningacquisition functionInstancedMesh

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install