Question

What is the terminology used to describe the connection where a set of neuron in one layer is connected to one neuron in the next layer?

(a) Fully connected

(b) Partially connected

(c) Splitting

(d) Pooling

I got this question by my school teacher while I was bunking the class.

My doubt stems from Techniques topic in portion Cognitive Techniques of Cognitive Radio

Answer 1

The correct answer is:

(b) Partially connected

Explanation: In neural networks, particularly in the context of feedforward networks or convolutional neural networks (CNNs), a partially connected architecture refers to a scenario where a set of neurons in one layer is connected to only a subset of neurons in the next layer, rather than all neurons being interconnected (which would be a fully connected scenario).

• Fully connected refers to each neuron in one layer being connected to every neuron in the next layer (i.e., a dense connection).
• Partially connected refers to some neurons in one layer being connected to a subset of neurons in the next layer (this is common in architectures like CNNs, where only certain neurons in one layer are connected to certain neurons in the next layer).
• Splitting and Pooling are different concepts used primarily in CNNs, with pooling referring to downsampling operations, and splitting referring to dividing data or operations across different parts of a network.

So, the terminology used to describe the connection where a set of neurons in one layer is connected to a subset (not all) neurons in the next layer is partially connected.