Chapter 22 Quiz

Question 1

sweeping protection to those in EU including a “right to be forgotten”

Answer 1

general data protection regulation

Question 2

decisions or actions that are illegal for a human in a company (discrimination in lending based on gender or age, for example) are also illegal for an algorithm to do in automated fashion

Answer 2

protected groups

Question 3

best practices for data science

Answer 3

non-maleficence, fairness, transparency, accountability, data privacy and security

Question 4

avoiding harm, ensuring models do not cause foreseeable harm or harms of negligence

Answer 4

non-maleficence

Question 5

equal representation, anti-discrimination, dignity, just outcomes

Answer 5

fairness

Question 6

user consent, model interpretability, Explanations for other modeling choices made by the creators of the model

Answer 6

transparency

Question 7

legal compliance, acting with integrity, responding to concern of individuals who use/are affected by machine learning

Answer 7

accountability

Question 8

other gathering necessary information, storing data securely, ensuring data is de-identified once used in model, ensuring other aspects of user data cannot be inferred from the model results

Answer 8

data privacy and security

Question 9

responsible data framework

Answer 9

justification, assembly, data preparation, modeling, auditing

Question 10

gain understanding of problem in a business context

Answer 10

justification

Question 11

team assembles various elements needed for project

Answer 11

assembly

Question 12

standard data exploration and prep

Answer 12

data preparation

Question 13

trying out multiple models and settings to find best one

Answer 13

modeling

Question 14

reviewing modeling process and resulting model performance

Answer 14

auditing

Question 15

impact statements, model cards, datasheets, audit reports

Answer 15

documentation tools

Question 16

interpretability methods apply to all types of models

Answer 16

model agnostic

Question 17

true positive rate

Answer 17

true positive/positive

Question 18

true negative rate

Answer 18

true negative/negative

Question 19

false positive rate

Answer 19

false positive/negative

Question 20

false negative rate

Answer 20

false negative/positive

Question 21

partial dependence plot

Answer 21

global interpretability

Question 22

individual conditional expectation plot, Shapley values, local interpretable model-agnostic explanations

Answer 22

local interpretability

Question 23

understanding how a model behaves on average across the entire dataset

Answer 23

global interpretability

Question 24

explaining individual predictions made by a model

Answer 24

local interpretability