Skip to content

CS Professional · Artificial Intelligence, Data Analytics and Cyber Security - Laws and Practice · Artificial Intelligence - Introduction and Basics

A company uses a deep neural network with many hidden layers to approve credit limits. A customer asks why her limit was reduced, and the company cannot trace the reasoning behind the output. Which feature of deep learning does this most directly illustrate, and why does it matter for compliance and accountability?

This illustrates the black-box problem of deep learning. Multi-layer neural networks produce outputs whose internal reasoning is hard to interpret, so the company cannot explain or audit the decision, which weakens accountability, fairness and the ability to justify automated decisions to customers and regulators.

  1. AOverfitting; it means the data was encrypted improperly
  2. BThe black-box problem; lack of explainability makes it hard to justify or audit automated decisionsCorrect
  3. CSupervised labelling; it guarantees decisions are always lawful
  4. DClustering; it shows the model has no hidden layers

Explanation

Deep networks with many layers learn complex internal representations that are hard to interpret, called the black-box problem. Without explainability, a company struggles to justify decisions to affected persons, auditors or regulators, which raises accountability and fairness concerns. It is not an encryption or labelling issue.

Did you get it right without looking?

One question tells you little. A timed set on Artificial Intelligence - Introduction and Basics shows your real accuracy, how long you take and where you lose marks.

Practise Artificial Intelligence - Introduction and BasicsFree diagnostic mock

More Artificial Intelligence - Introduction and Basics questions