deep networks

Edit Answers
12345678910
Across
  1. 4. – Popular activation that outputs zero for negatives
  2. 6. – Algorithm used to update neural network weights
  3. 7. – Measures prediction error
  4. 9. – Model memorizes training data
  5. 10. – Activation function with S-shaped curve
Down
  1. 1. – Optimization method using gradients
  2. 2. – Direction of steepest increase
  3. 3. – Mathematical mapping used in networks
  4. 5. – Core operation in CNNs
  5. 8. – Function applied to a neuron’s output