Midterm 2

Question 1

What Is artificial intelligence?

Answer 1

A computer program that mimics the intelligence of humans

Question 2

What is machine learning?

Answer 2

technique which a computer can use to learn from data using complex rules

Question 3

What is deep learning?

Answer 3

A technique for machine learning based on the neurons in the brain

Question 4

List the three types of Machine learning and explain them

Answer 4

1. Unsupervised learning - no feedback given to algorithm
2. supervised learning - every example has a label
3. Reinforcement - reward or punishment per action

Question 5

How does supervised learning training work?

Answer 5

Training is a collection of labelled examples {xi, yi} where xi is a feature vector with D dimensions and y is a label

Question 6

What is k-nearest neighbors?

Answer 6

Looks at neighbors closest to the data (similar feature values) and what they are classified as

Question 7

What is linear regression?

Answer 7

Supervised machine learning used on continuous numerical data. It enables us to identify a linear trend and outliers

Question 8

What is binary classification

Answer 8

Supervised learning where the objective is to organize everything into one of 2 classes (logistic regression)

Question 9

what is multi-class classification

Answer 9

supervised learning into 3 or more discrete classes. can be transformed into binary
- one vs all (OvA)

Question 10

Explain one vs all

Answer 10

separate binary classifier for each class. each classifier labels one class as positive and all others as negative. final assignment is based on the classifier with the highest confidence score

Question 11

What is a decision boundary?

Answer 11

A boundary which partitions the underlying feature space into regions corresponding to different class labels

Question 12

What is linearly separable data?

Answer 12

data is linearly separable when 2 classes can be perfectly separated by a single linear boundary (line for 2d, plane for 3d, hyperplane for >3d)

Question 13

what is the difference between a simple decision boundary and a complex one?

Answer 13

simple is when the boundary comes from a polynomial function.
complex is an irregular decision boundary generated by decision trees

Question 14

What is logistic regression?

Answer 14

it is a binary (0,1) classification algorithm which determines the probability that a given instance xi belongs to the positive class

Question 15

Explain the logistic function

Answer 15

maps a real valued input to the open interval 0-1. It is called a squashing function because it maps a wide input domain to a constrained output

Question 16

What is underfitting?

Answer 16

Machine learning concept where the model is too simple to accurately classify the data. It is underfitting if it has poor performance on both training and test data and adding more data doesn’t correct the issue

Question 17

What is overfitting?

Answer 17

When the model is too complex for a given classification problem (tall decision tree, deep and wide neural networks). Too many features creates excellent performance on the training set but poor performance on the testing set

Question 18

Explain learning curves?

Answer 18

Displays the performance of our model by using Root mean square error (RMSE) on both the training and test sets

Question 19

What is the Bias/Variance trade off?

Answer 19

Bias -> error created by overly simplistic models, high bias = underfitting
Variance -> error from overly complex models that is sensitive to fluctuations in the training data. High variance = overfitting
Tradeoff -> aim for a model that generalizes new data well

Question 20

Explain the confusion matrix

Answer 20

A matrix which displays the true positives, false negatives, false positives and true negatives for all labels

Question 21

What is accuracy?

Answer 21

The ratio of correctly predicted instances and the total number of predictions

Question 22

What is precision

Answer 22

ratio of true positives (TP) / total number of positives

Question 23

Explain the holdout method

Answer 23

Allocate roughly 80% of your dataset for training and reserve the remaining 20% for testing
- Training error generally low otherwise there is something wrong
- Generalization error - error rate observed when the model is evaluated on new unseen data

Question 24

What is cross validation?

Answer 24

method to evaluate models and improve performance. Involves partitioning the dataset into multiple subsets

Question 25

Explain k-fold cross validation

Answer 25

1. Divide the dataset into k equally sized folds 2. Training and validation - for each iteration, one fold is used as the validation remaining as training 3. Evaluation - models performance is evaluated in each iteration, resulting in k performance measures 4. Aggregation - stats are calculated based on k performance measures

Question 26

What are the benefits of k fold compared to normal test train split?

Answer 26

Much more reliable estimate of model performance. Results in better generalization and reduced variability Works very well for hyper parameter tuning

Question 27

Challenges of multi fold

Answer 27

Computationally costly - takes forever to train and doing it a bunch of times increases that Class imbalance - folds may not represent minority classes (if one fold contains a ton of one class it could skew training or validation) Error prone

Question 28

What is a hyperparameter

Answer 28

A hyperparameter is a configuration external to the model that is set prior to the training process and dictates the learning process

Question 29

Grid search

Answer 29

1. Enumerates through all possible hyperparameter combinations 2. train on training set, evaluate on validation set

Question 30

Data augmentation

Answer 30

a technique used to increase the diversity of a dataset by applying various transformations to the existing data

Question 31

What is one-hot encoding?

Answer 31

A technique that converts categorical variables into a binary vector representation where each category is represented with a single 1 and all others as 0 (e.g. instead of something just being labelled 5 it is 0, 0, 0, 0, 1)

Question 32

Explain why one-hot encoding is beneficial

Answer 32

Increases the dimensionality of feature vectors. it helps it avoid bias

Question 33

What is Binning (feature engineering)

Answer 33

placing things into bin categories. e.g. ages into: child, teen, adult and senior

Question 34

What is normalization

Answer 34

A scaling technique which accelerates optimization -> algorithms perform optimally when feature values are within similar ranges and this helps with it

Question 35

What is standardization?

Answer 35

Transforms each feature to have a normal distribution with a mean of 0 and a standard deviation of 1

Question 36

Standardization or Normalization?

Answer 36

-> standardization for unsupervised learning or if features resemble a normal distribution -> standardization handles outliers better otherwise use normalization

Question 37

What is data imputation

Answer 37

Data imputation -> the process of replacing missing values in a dataset using statistics or machine learning

Question 38

Data imputation strategies

Answer 38

- mean, median or mode replacement - special value method -> value outside normal range as a notifier of a missing value

Question 39

What is a class imbalance

Answer 39

A scenario where the number of instances in one class significantly outnumbers the instances of another one -> the model becomes biased towards the dominant majority class

Question 40

Explain the solutions to class imbalance

Answer 40

Oversampling the minority class -> can lead to overfitting and poor performance on the test data Undersampling the majority class -> loss of info about majority class can lead to underfitting Synthetic data -> generate fake minority data

Question 41

Deep learning - how it works

Answer 41

Machine learning technique that can be applied to supervised learning, unsupervised learning and reinforcement learning It is inspired by neurons and uses layers of them connected to classify things

Question 42

Explain the layers of a neural network

Answer 42

Input: where the data is input it corresponds to the number of attributes in the data Hidden layer: the process for which the computer sorts the data Output layer: where the data is classified completely

Question 43

What is an activation function in relation to neural networks

Answer 43

Activation function is applied to the entire neural network and introduces non-linearity into the neural network -> a neuron is fired or activated when the requirement passed to the node exceeds the value stored within the node

Question 44

Explain the three common activation functions?

Answer 44

Sigmoid -> (sin based function) produces outputs inbetween (0, 1) Tanh -> (tan based function)outputs in between (-1, 1) ReLU ->(rectified linear unit) outputs values in the interval [0, infinity)

Question 45

Explain the universal approximation theorem

Answer 45

A neural network with a single hidden layer can approximate any continuous function

Question 46

What is back-propagation

Answer 46

A learning procedure which repeatedly adjusts the weights of the connections in the network so as to minimize a measure of the difference between the actual output vector of the net and the desired output vector.

Question 47

Explain the backpropagation steps

Answer 47

1. Initialization 2. Forward pass 3. compute loss 4. backpropagation 5. repeat 2 to 5 The algorithm stops either after a number of epochs or when the convergence criteria is satisfied

Question 48

Explain the initialization step

Answer 48

Initializing the weights and biases of the neural network 1. Zero initialization - all weights are 0 - doesn't work well for symmetry as all neurons produce identical outputs 2. Random initialization - weights are initialized randomly using uniform or normal distributions

Question 49

Explain the forward pass step

Answer 49

Data is passed to the first layer -> for each of the hidden layers, compute the activations by applying the weighted sum of inputs plus bias -> followed by an activation function

Question 50

Explain the compute loss step

Answer 50

Calculate the error (loss) using a suitable function by comparing predicted values with actual target values -> A smaller loss indicates that predicted values are closer to the actual target values

Question 51

Explain the Backwards pass step

Answer 51

Output layer: Compute the gradient of the loss with respect to the output layers weights and biases using the chain rule of calculus Hidden layers: propagate the error backwards through the network layer by layer. For each layer compute the gradient loss with respect to the weights and biases Update the weights and biases: adjust the weights and biases using the calculated gradients and a learning rate

Question 52

What is gradient descent

Answer 52

An optimization algorithm used to minimize the loss function by iteratively moving towards the steepest descent as defined by the negative gradient

Question 53

Explain weight initialization and how vanishing gradients were solved

Answer 53

Random initialization breaks the symmetry and allows for effective learning Glorot initialization fixes it for sigmoid and tanh' He initialization is optimal for ReLU and its variants

Question 54

Explain how learning rate is important to optimization

Answer 54

Learning rate determines the step size during optimization

Question 55

Explain the hierarchy of concepts

Answer 55

Each layer detects patterns from the output of the layer preceding it -> in other words the network uncovers patterns of patterns

Question 56

What is a Convolutional Neural Network

Answer 56

Crucial pattern info is often local (e.g. top left edge) convolutional layers, reduce parameters significantly because neurons are not fully connected to the preceding layer but rather their receptive fields

Question 57

What is a kernel in the context of machine learning

Answer 57

A kernel is a small matrix that slides over input data such as an image to perform convolution. The kernel is moved through the entire image one n pixels at a time (kernel is nxn) the values in the kernel are multiplied by the value in the input matrix region the overlap and then all the values are summed to make a single scalar value -> output matrix is the feature map

Question 58

What is a receptive field

Answer 58

Each unit is connected to neurons in its receptive fields -> unit i, j in layer l is connected to the units (i to i + fh -1) and (j to j+fw-1) of the layer l-1

Question 59

What is padding?

Answer 59

Zero padding -> to have layers of the same size the grid can be padded with zeroes -> allows it to recognize edges

Question 60

Explain the stride

Answer 60

Stride -> it is possible to connect a larger layer (l-1) to a smaller one (l) by skipping units. The number of units skipped is called the stride

Question 61

what are filters

Answer 61

A window of size fh x fw is moved over the output layers l-1 referred to as the input feature map - For each location, the product is calculated between the extracted patch and a matrix of the same size known as the convolution kernel or filter

Question 62

Explain the kernel parameters and where they originate

Answer 62

The parameters of the kernel are learned through backpropagation allowing the network to optimize its feature extraction capabilities based on the training data

Question 63

what is a feature map

Answer 63

in CNN the output of a convolution operation is the feature map

Question 64

What is the bias term

Answer 64

a single bias term is added uniformly to all entries of the feature map -> this bias helps adjust the activation level

Question 65

what is pooling

Answer 65

basically a convolutional layer except there is no weights instead there is aggregating function normally max or mean-> each neuron in a pooling layer is connected to neurons in the receptive field

Question 66

Advantages to pooling

Answer 66

Dimensionality reduction -> reduces spatial dimensions of input feature maps decreasing the # of parameters and computational load Feature extraction -> essentially summarizes the region discarding less important details Translation invariance -> network becomes less sensitive to small changes Noise reduction -> smooths noise through aggregation

Question 67

What are the environmental characteristics

Answer 67

1. Observability: partially or fully 2. agent composition: single or multiple 3. Predictability: deterministic or nah 4. State dependency: stateless or stateful 5. temporal dynamics: static or dynamic 6. state representation: discrete or continuous

Question 68

Explain the search problem

Answer 68

A collection of states (state space) -> an initial state where the agent begins -> one or more goal states -> a set of actions available in the state -> a transition model that determines the next state based on the current state and action

Question 69

What is informed search

Answer 69

searching with heuristic functions involved to estimate costs

Question 70

Explain best-first search

Answer 70

Breadth first improvement -> uses heuristics to prioritize nodes that seem closer to the goal -> it uses a priority queue sorted by estimated cost

Question 71

Manhattan distance heuristic

Answer 71

specifically for 8-tile problem -> calculates the sum of the distance of tiles from their goal position