AI / ML VisualizerRuns in your browser · No API keyOpen Model
← All topics

Foundations

Backpropagation

Inspect the chain rule and apply a real gradient update.

Weighted neural networkWeighted neural networkWeight 0.24770.2477Weight -0.1267-0.1267Weight 0.86340.8634Weight 0.41580.4158Weight -0.9198-0.9198Weight 0.07520.07520.7Layer 0, neuron 1; activation 0.7; bias 00.3Layer 0, neuron 2; activation 0.3; bias 00.1345Layer 1, neuron 1; activation 0.13453764023190118; bias 00.6225Layer 1, neuron 2; activation 0.6225030870771254; bias 00.4808Layer 2, neuron 1; activation 0.48077911982248117; bias 0Solid: positive · Dashed: negative · Width: |weight| · Select an edge to inspect
L = ½(a − target)²
½(0.4808 − 1)²
Loss = 0.1348

Inputs

Hidden z · 1 × 2
w(new) = w − η ∂L/∂w
0.2477 − 0.1 × 0.0819
New weight = 0.2395
Step 0 / 6

Continue exploring

Artificial Neuron →Activation Functions →Forward Propagation →

No account required. Local images and model files stay in your browser. Educational models demonstrate mechanisms; they are not production predictors.