Question 1
How does gradient descent minimize the cost function?
Question 2
What is the key advantage of using Stochastic Gradient Descent over traditional Gradient Descent?
Question 3
What happens to weights between neurons when they frequently activate together, according to Hebbian theory?
Question 4
In a neural network designed for handwritten digit recognition, how many neurons are in the input layer?
Question 5
How does backpropagation update the weights in a neural network?
Question 6
Which of the following best describes mini-batches in the context of Stochastic Gradient Descent?
Question 7
Which concept suggests that 'neurons that fire together, wire together'?
Question 8
What does the cost function measure in a neural network?
Question 9
What is the effect of a higher sensitivity in the cost function on backpropagation?
Question 10
What is the primary purpose of backpropagation in neural network learning?
Question 11
How is the cost for a single example calculated in a neural network?
Question 12
What is the purpose of averaging changes across all training examples during backpropagation?
Question 13
What type of neurons are present in the output layer in a neural network configured for digit recognition?
Question 14
Why do changes propagate backwards through layers in backpropagation?
Question 15
What dataset is often used as an example for training neural networks in digit recognition?