CSE 6740 · Lecture 11
Backpropagation Lab
Edit weights
Reset
Full screen
Example
Sigmoid network
Saturated sigmoid
Inactive ReLU
Activation
Sigmoid
Tanh
ReLU
Input x₁
Input x₂
Target
Show values
Select a weight or neuron
Chain rule
Select a weight or neuron to trace its gradient. Blue arrows point forward; orange arrows point backward.
Prediction
Target
Loss
← Back
Next →
↻ Replay
▶ Play
Learning rate
0.05
0.5
2
10
Apply 1 update
Apply 20 updates
Undo
Loss convergence
· single example
Weights & biases
Done
W connects source i to destination j. Manual edits start a new loss history.
This interactive lab requires JavaScript.