Neural Network Basics
What is a neural network?
A system that mimics human thinking.
What is a neuron?
The smallest unit of a neural network.
What is an activation function?
It decides whether a neuron should fire.
What is nn.ReLU()?
Applies the ReLU activation function.
What is forward propagation?
Moves data from the input to the output layer.
What are input layer, hidden layers, and output layer?
The three main parts of a neural network.
What is a hidden layer?
It processes and learns important features.
What is the sigmoid function?
An activation function that outputs values between 0 and 1
What is backward()?
Calculates gradients in PyTorch.
What is backward propagation?
Adjusts weights based on error to improve learning.
What is y = f(WX + b)?
The forward propagation formula.
What is bias?
A fixed value added to the weighted sum.
What is ReLU?
It converts negative values to zero and keeps positive values the same.
What is requires_grad=True?
Enables gradient computation for a tensor.
What is a loss function?
Measures how wrong the predictions are.
What does a neuron do (weights + bias + activation)?
Applies weights, adds bias, and passes through an activation function.
What is a weight?
Represents the importance of the input.
What is the purpose of activation functions?
They determine the neuron’s output and introduce non-linearity.
What is optimizer.step()?
Updates the weights during training.
What optimizer is used (SGD)?
Stochastic Gradient Descent (SGD).
What are the three main parts of a neural network?
Input layer, hidden layers, and output layer
What is the smallest unit of a neural network?
A neuron.
What activation function outputs values between 0 and 1?
Sigmoid.
What function updates the weights?
optimizer.step()
How does a neural network improve its learning?
By adjusting weights using backward propagation.