BGE (Print) Machine learning and bias

What is essential for machine learning to function effectively?

• High quality and large quantities of training data

• Expensive computers

• Simple algorithms

• Minimal data

Where does training data for machine learning often come from?

• Automatic collection from user activities

• Government databases

• Manual collection by researchers only

• Direct user inputs only

What can cause a computer to make biased predictions?

• Learning from biased data

• Having an excess of data

• Using advanced algorithms

• Using unbiased data

What is a potential problem if training data is collected only from men?

• There will be no noticeable effect

• The data will be considered more comprehensive

• The predictions can be universally accurate

• The system may not work well for women

What should you ask when evaluating training data for machine learning?

• Is the data expensive?

• Is the data colorful and visually appealing?

• Does the data represent all possible scenarios and users without bias?

• Can the data be stored easily?

What does biased data in machine learning typically favor?

• All scenarios equally

• Random selection of scenarios

• Certain scenarios over others

• None, it is always unbiased