Roy's AI LabAI Engineering
Deep Learning Interview Practice
Practise Deep Learning with multiple-choice questions in the style used in technical screening. Every answer comes with an explanation.
Free sets
Pro sets
Sample questions
1. What does a neuron compute?
- A. A weighted sum of its inputs plus a bias, passed through an activation function
- B. A random number with no inputs
- C. A copy of its input
- D. The whole training dataset
2. Why do networks need activation functions?
- A. They remove the need for weights
- B. They introduce non-linearity, so stacked layers can model more than a straight-line function
- C. They store training data
- D. They always output integers
Answers and explanations appear after you submit the full free set.