In neural networks, what are the basic processing units that neurons are connected through?

• A. Weights
• B. Biases
• C. Layers
• D. Neurons

Answer: (D)

The essential element being tested is identifying the basic processing unit in a neural network. Each neuron is a processing unit that takes inputs from other neurons, multiplies each input by a weight, sums them, adds a bias, and then passes the result through an activation function to produce an output. This is how computation is actually performed at the smallest scale in a network. Neurons are then connected to other neurons via those weighted connections, and grouped into layers to build the full architecture. The other terms describe parts of the connectivity or organization: weights are the strengths of the connections, biases shift the activation, and layers are collections of neurons.