DL_Module-2_Puzzle1

Edit Answers
123456789
Across
  1. 2. The mathematical rule used to differentiate a composite function.
  2. 3. Regularizer that can force some weights exactly to zero
  3. 4. A mathematical structure representing operations and dependencies between variables.
  4. 7. Regularization technique that randomly disables activations during training.
  5. 9. The process of propagating input values through successive layers to obtain a prediction.
Down
  1. 1. The parameter that controls the additive shift in a neuron
  2. 3. The intermediate value produced by a neuron before applying its activation function.
  3. 5. The derivative of the loss with respect to a parameter.
  4. 6. Optimization method that maintains a moving direction based on previous updates.
  5. 8. The quantity that measures the discrepancy between prediction and target.