Question 7Which of the following are true about bottleneck layers? (Check all that apply)

To answer your question about bottleneck layers, let's identify the general characteristics of bottleneck layers in neural network architectures.

1. Definition of Bottleneck Layer: A bottleneck layer is typically a layer that has fewer neurons than the layers before and after it, effectively "bottlenecking" the information flow. This is often used to force the network to learn a compressed representation of the data.

2. Characteristics:

   • They help in reducing dimensionality.
   • They can reduce the computational load.
   • They lead to capturing important features effectively.

3. Common Usage: Bottleneck layers are often used in architectures like ResNet and autoencoders.

From this explanation, the true characteristics of bottleneck layers usually include reduced dimensionality, the efficiency of feature extraction, and computational benefits.

Please check the provided options and evaluate each one based on these characteristics to determine which are true about bottleneck layers.