Weights and biases are the paremeters of a neural network

$y=x\ast w+b$ In neural networks, ws are the called the weights of the neural network and the bs are the biases of the network.

Together, they are the parameters of the network.

The biases need to exist because the function $y=x\ast w$ is not flexible enough. If $x=0$, the result will always equal zero.

mlai