KDELengthScaleView = slider({
title: "KDE Parameter",
min: 0.01,
max: 1,
value: 0.3, // this can be from dataset
step: 0.01,
precision: 4,
description: html`RBF Kernel Length Scale (${tex `l_\text{KDE}`})`
})
KDELengthScale = Generators.input(KDELengthScaleView)
score = (p) => {
const px = kde(p)
const mean = mathjs.mean(
trainingData.map(d => rbf_for_kde.grad(p, d)),
0
)
return mathjs.multiply(mean, 1 / px)
}
{
chart.update(current_epoch)
chart.showParticleTrajectory(isShownParticleTrajectory)
}
SVGDLengthScaleView = slider({
min: 0.01,
max: 1,
value: 0.3, /// default from dataset
step: 0.01,
precision: 2,
description: html`Kernel Length Scale (${tex `l_\text{SVGD}`})`}
)
lrView = slider({
min: 0.0001,
max: 0.8,
value: 0.3,
precision: 2,
step: 0.01,
description: html`Learning Rate (${tex `\gamma`})`,
title: "SVGD Parameters",
})
numberofEpochsView = slider({
min: 10,
max: 200,
value: 100,
step: 10,
description: html`Number of Epochs (${tex `T`})`
})
totalParticlesView = slider({
min: 10,
max: 200,
value: 50, // default from dataset
step: 10,
description: html`Number of Particles (${tex `M`}); the larger large value the higher latency.`}
)
mutable particles = {
resetParticles;
dataset;
return Array.apply(null, Array(totalParticles)).map(() => [uniformSamplingInRange(-maxVal, maxVal), uniformSamplingInRange(-maxVal, maxVal)])
}
showParticleTrajectoryView = checkbox(
{label: html`Show Particle Trajectory`, checked: false}
)
isShownParticleTrajectory = Generators.input(showParticleTrajectoryView)
showDataView = checkbox({
title: "Visualization",
label: "Show Data",
checked: false
})
showData = Generators.input(showDataView)
showScoreView = checkbox({
label: html`Show Score Vector Field ${tex `\nabla_{\mathbf x}\log \hat p (\mathbf x)`}`,
checked: false
})
showScore = Generators.input(showScoreView)
showContour = Generators.input(showContourView)
showContourView = checkbox({
label: "Show Contour",
checked: true
})
height = 400
width = 400
One platform to build and deploy the best data apps
Experiment and prototype by building visualizations in live JavaScript notebooks. Collaborate with your team and decide which concepts to build out.
Use Observable Framework to build data apps locally. Use data loaders to build in any language or library, including Python, SQL, and R.
Seamlessly deploy to Observable. Test before you ship, use automatic deploy-on-commit, and ensure your projects are always up-to-date.