2.4 Bias

Question 1

Bias (in AI-based systems)

Answer 1

Statistical measure of the distance between the outputs provided by the system and ‘fair outputs’ which show no favoritism to a particular group

Question 2

Inappropriate biases can be linked to attributes such as

Answer 2

Gender, Race, Ethnicity, Sexual orientation, Income level, Age

Question 3

Examples of reported inappropriate bias in AI-based systems

Answer 3

Systems used for making recommendations for bank lending, Recruitment systems, Judicial monitoring systems

Question 4

Much discussion relating to bias takes place in the context of Make decisions and predictions like in

Answer 4

ML systems

Question 5

Components that can introduce bias in ML system results

Answer 5

Learning algorithm, Collected data

Question 6

Algorithmic bias occurs when

Answer 6

The learning algorithm is incorrectly configured (e.g., overvalues some data compared to others)

Question 7

Algorithmic bias can be caused and managed by

Answer 7

Hyperparameter tuning of ML algorithms

Question 8

Sample bias occurs when

Answer 8

The training data is not fully representative of the data space to which ML is applied

Question 9

Inappropriate bias is often caused by

Answer 9

Sample bias

Algorithmic bias