All MCQs

The correct answer is backpropagation through time unfolds an rnn across time steps to compute gradients. This matches the Deep Learning course topic: BPTT.

The correct answer is standard benchmarks help compare architectures on common tasks and datasets. This matches the Deep Learning course topic: Benchmarks.

The correct answer is pooling reduces spatial dimensions while keeping important features. This matches the Deep Learning course topic: CNN pooling.

The correct answer is batch normalization normalizes intermediate activations to stabilize and speed up training. This matches the Deep Learning course topic: Batch normalization.

The correct answer is a feature map represents filter responses across spatial locations. This matches the Deep Learning course topic: CNN feature maps.

The correct answer is dropconnect randomly drops weights rather than neuron outputs during training. This matches the Deep Learning course topic: DropConnect.

The correct answer is underfitting occurs when a model is too simple to capture important patterns. This matches the Deep Learning course topic: Underfitting.

The correct answer is deep learning learns layered representations from raw data. This matches the Deep Learning course topic: Basics of deep learning.

The correct answer is cnn computational cost depends on filter size, input size, channels, and number of filters. This matches the Deep Learning course topic: CNN complexity.

The correct answer is resnet uses skip connections to help train very deep networks. This matches the Deep Learning course topic: ResNet.

The correct answer is a multilayer perceptron uses layers of neurons with nonlinear activations. This matches the Deep Learning course topic: MLP.

The correct answer is deep models usually contain multiple hidden layers. This matches the Deep Learning course topic: Shallow vs deep learning.