← Back to ML Visualizations

t-SNE Visualization

t-Distributed Stochastic Neighbor Embedding reduces high-dimensional data to 2D while preserving local structure.

Cost = KL(P || Q)

Key parameter: Perplexity controls the effective number of neighbors considered.

Watch as the algorithm iteratively minimizes the KL divergence to separate clusters.

Optimization

Iteration: 0
KL Divergence: -
Points: 0
Clusters: 0
200
30
200