FICT_AI_Day2_2024

Who's eye is this?

• Dr Joel

• Dr Lee Sue Han

• Dr Kelvin

• Dr Brian

Answer explanation

What is a MAXPOOLING layer?

• Downsamples the input along its spatial dimensions (height and width) by taking the maximum value over an input window (of size defined by pool_size ) for each channel of the input.

• is used to introduce nonlinearity in a neural network, helping mitigate the vanishing gradient problem during machine learning model training and enabling neural networks to learn more complex relationships in data.

• converts raw output scores — also known as logits — into probabilities by taking the exponential of each output and normalizing these values by dividing by the sum of all the exponentials.

• eminal deep learning model in which the weight layers learn residual functions with reference to the layer inputs

After training a model, you would save the model as model.ipynb format (TRUE OR FALSE)

• TRUE

• FALSE

Answer explanation

model.h5 or model.keras

Typically we test a model before training it (TRUE/FALSE)

• TRUE

• FALSE

Answer explanation

Train first then test

Which one describes a pretrained deep learning network?

• A network that was trained on thousands of images of several classes

• A network that has some randomly initialised weights

• A network with a different classification layer

• A network with a fixed three channel image inputfor dogs

This process results in the updating of the weights in a way that minimizes the loss by giving the nodes with higher error rates lower weights, and vice versa

• Convolution

• Maxpooling

• Activation

• Backpropagation

Which of the statements below is suitable for a multi-class problem?

• Sigmoid activation

• Binary

• Categorical Cross Entropy

• Three Channel

HuggingFace allows you to deploy your AI solutions in specific spaces (TRUE/FALSE)

• TRUE

• FALSE

What date is today?

• 20/7/24

• 19/7/24

• 21/7/24

• 13/7/24