! Session 1 & 2: Intro

Question 1

Algorithm

Answer 1

set of procedures that creates a model when trained. E.g., linear regression

Question 2

Model

Answer 2

fitted algorithm that has been trained. E.g. a linear regression model that has been trained to predict prices of ‘X’.

Question 3

Parameters

Answer 3

internal variables within model that adjust automatically when training model

Question 4

Hyperparameters

Answer 4

variables set by the user to control the algorithm and they define how it learns from the data

Question 5

Data Science

Answer 5

= Collection of statistical & ML model that
- supports info extraction from data
- offers insights, causality, predictions

Question 6

ML

Answer 6

• science of programming computers so they learn from data without being explicitly programmed
• component of data science
• AI technique for sophisticated cognitive task

Question 7

ML Functions

Answer 7

• Descriptive: uses data to explain what happened
• Predictive: uses data to predict what will happen
• Prescriptive: use data to suggest actions

Question 8

ML Subfields

Answer 8

• Natural language processing: machines learn to understand natural language as spoken and written by humans
• Neural networks: modeled on the human brain
• Deep learning networks: neural networks with many layers

Question 9

When does ML work well?

Answer 9

• (Large) data is available
• Problem is dynamic or fluctuating
• Problem requires predictions or discovering patterns

Question 10

ML Flow

Answer 10

Preperation
- Identify Question & Task
- Data Collection
- Data preprocessing
- EDA

Model Development
- Feature & Model Selection
- Splitting Data
- Training with Train Set, validation
- Evaluate with Test set (repeat model development until results satisfiing)

Communicate results & Deploying Model

Question 11

Challenges

Answer 11

• Explainability: What are ML models doing? How decisions made?
• Bias & unintended outcomes

Question 12

Challenge: Bias & unintended outcomes

Answer 12

• insufficient Data (Not enough training, non-representative, irrelevant features)
• Overfitting & underfitting

Question 13

Bias-Variance Tradeoff

Answer 13

• goal: low bias & low variance
• Bias = amount of error by approximating real-world phenomena with simplified model -> underfit
• Variance = how much models test error changes based on variation in training data, prediction error when using data not previously seen by model-> overfit

Question 14

Subcategories of ML Models

Answer 14

1. Based on Training
   - supervised
   - unsupervised
   - reinforcement
   - semi-supervised
2. Based on working
   - instance based
   - model based

Question 15

Subcategories of ML Models - based on training: super and unsupervised

Answer 15

• Supervised = trained with labeled data -> know answers we want
• Unsupervised = looks for pattern in unlabeled data -> goal: find unkonwn structures / trends

Question 16

Subcategories of ML Models - based on working

Answer 16

• Instance based = learns all previous data, compare new data to it & generalizes based on similarity
• model based = learns model based on training data & predict labels according to it

Question 17

Supervised ML Models

Answer 17

• Regression: x -> continous y
• Classification: x -> discrete (binary) y

Question 18

Supervised ML Models - Regression

Answer 18

• Linear Regression
• Neural Networks

Question 19

Unsupervised ML Models - Categories

Answer 19

• Clustering: x -> discrete y
• Dimensionality Reduction: x -> continous y
• Density Reduction: select relevant variables

Question 20

ML versus traditional AI techniques

Answer 20

• trad. AI: static, rule based, no generalization
• ML: dynamic, data driven, generalization

Question 21

e.g. Chess

Answer 21

• Symbolic AI: sit down with best chess player & put knowledge in PC
• Statistical AI: Simulate all possible moves & outcomes & take most likely to win
• ML: Show millions of examples & let program learn

Question 22

Where ML > other AI techniques

Answer 22

• tasks programmers cant describe (handwriting, cognitive reasoning)
• complex multidimensional problems that cant be solved by numerical reasoning (weather forecast, health care outcomes)

Question 23

3 C’s of ML

Answer 23

• Collaborative filtering: technique for recommendations, same algorithm for different objects, e.g. Amazon & Netflix
• Classification
• Clustering

Question 24

Over and underfitting

Answer 24

• Overfitting: learning a function that perfectly explains training data that the model learned from, but doesn’t generalize well - high variance
• Underfitting: (strong correlated features) model is not complex enough to capture underlying trend - high bias

Question 25

Reinforcement learning

Answer 25

trains through try & error with reward system, e.g. Roboter learns to walk (difficult to define task, learning dangeorus e.g. car)

Question 26

semi-supervised learning

Answer 26

Learns with few labels clustering & adopts labels to all instances of cluster

Question 27

Parametic

Answer 27

Any algorithm that learns using a pre-defined mapped function e.g. linear regression

Question 28

Non parametric

Answer 28

Any algorithm that does not make assumptions about form of mapping function e.g. KNN & SVM

Question 29

Pro and con parametric

Answer 29

• pro: simple, fast, less data
• con: constrained to parameters / assumptions, limited complexity, poor fit

Question 30

Pro and con non parametric

Answer 30

• pro: flexibility (large nr of features), high power, performance
• con: more data, slower, overfitting

Question 31

Convex

Answer 31

A shape where given any 2 points in subset, the subset contains the whole line segment that joins them