← Emma Leonhart

Interactive ML Tutorials

Hands-on visualizers for the math behind machine learning. Drag the inputs, change the weights, watch the gradients flow. Each tutorial is a self-contained page — pick whichever feels closest to what you want to understand.

Vector Math

Embedding Viewer
Voronoi map of 485 word embeddings projected onto custom semantic axes.
Cosine Similarity
A smart sprinkler AI that decides by directional alignment.
Dot Product
Drag vectors to see projection, angle, and both formulas live.
Cross Product
Parallelogram area and rotation direction for 2D cross product.

Neural Network Basics

Multilayer Perceptron
Drag inputs to see weights, sums, and activations flow through a neural network.
Backpropagation
Computational graph with forward and backward pass chain rule.
Linear vs Logistic Regression
Click to add data points and compare how each method fits.
Feature Transforms
Polynomial features turn linear models into curve fitters.

Training Dynamics

Loss Functions
Compare Step, MSE, Cross-Entropy, and Softmax losses interactively.
Regularization
L1, L2, Elastic Net weight penalties plus Dropout and Batch Norm.
Optimizers
GD, SGD, Momentum, AdaGrad, RMSProp, Adam, and AdamW racing on loss surfaces.

Architectures

Convolutional Neural Network
Slide a kernel over input data to see how convolution produces feature maps.
CNN Architecture Timeline
From LeNet to ConvNeXt — what each architecture introduced.
Recurrent Neural Network
Watch hidden state carry memory forward as an RNN processes a sequence.
LSTM
Step through forget, input, and output gates of a Long Short-Term Memory cell.
Keys, Queries & Values
How transformers decide which words matter to each other.