5 questions · pass at 70% to earn the journey badge. You can retake it anytime.
01 What is the basic computational unit of a neural network?
A database row An artificial neuron that computes a weighted sum of inputs and applies an activation function A single if/else rule A pixel
02 What is the purpose of an activation function?
To store the training data To introduce non-linearity so the network can model complex relationships To slow training down on purpose To encrypt the weights
03 What does backpropagation do during training?
Generates new training data Propagates the output error backward to compute gradients and update weights Deletes neurons that are unused Chooses the model architecture automatically
04 What key mechanism makes the transformer architecture so effective?
Recurrence that processes one token at a time only Self-attention, which lets the model weigh the relevance of all tokens to each other Hard-coded grammar rules Storing the entire internet in memory
05 Where do neural networks sit relative to machine learning and deep learning?
They are unrelated to machine learning Deep learning is machine learning using multi-layer (deep) neural networks Machine learning is a subset of neural networks Deep learning replaced machine learning entirely