Deep Learning

What is an important aspect of deep learning that involves automatically constructing meaningful features of the data?

• A)   Data encryption

• B) Feature extraction

• C) Manual labeling

• D) Data compression

2. Who developed the first mathematical model of a neural network?

• A)   Henry J. Kelley

• B) Stuart Dreyfus

• C) Warren McCulloch and Walter Pitts

• D) Alexey Grigoryevich Ivakhnenko

What do deep learning neural networks mimic to make a series of calculations and reach a conclusion?

• A)   The data storage process

• B) The decision-making processes of the human brain

• C) The mechanical movements of robots

• D) The communication patterns of networks

What is one of the key advantages of using machines with deep learning over humans?

• A)   Ability to work without any errors

• B) Ability to process massive amounts of data

• C) Ability to make ethical decisions

• D) Ability to mimic physical tasks

How can deep learning positively impact businesses if used effectively?

• A)   It can automate customer service entirely

• B) It can reduce the need for data

• C) It can improve productivity, increase retention and drive revenue

• D) It can eliminate the need for human workers

What technique is introduced to address the vanishing or exploding gradients problem in Recurrent Neural Networks (RNNs)?

• A)   Convolutional layers

• B) Dropout

• C) Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU)

• D) Feedforward architecture

Which technique is commonly used in deep learning to prevent overfitting by controlling the complexity of the model?

• A)   Stochastic Gradient Descent (SGD)

• B) Dropout and L1/L2 Regularization

• C) Transfer Learning

• D) Feedforward Network Architecture

Which of the following is a key task in computer vision that deep learning architectures are commonly used for?

• A)   Natural language translation

• B) Object detection

• C) Speech synthesis

• D) Data encryption

Which model uses a triplet loss function to cluster face representations of the same person during training?

• A)   Google's DeepFace

• B) Facebook's FaceNet

• C) Microsoft's Face API

• D) Apple's Face ID

What type of deep learning architecture is used in a multimodal multistream framework to tackle egocentric activity recognition using both video and sensor data?

• A) Single Convolutional Neural Network (CNN)

• B) Feedforward Neural Network

• C) Dual CNNs and Long Short-Term Memory (LSTM)

• D) Recurrent Neural Network (RNN)