Which of the following steps in deep learning are automatically completed by the model?

Question

Options

• AModel training
• BFeature selection
• CAnalysis and positioning task
• DFeature extraction

• A
  15% (5)
• B
  79% (27)
• C
  6% (2)

Explanation

In deep learning, the neural network automatically learns to extract meaningful features from raw data through its layered architecture (D), and simultaneously learns which of those features matter most - effectively performing feature selection (B) - via backpropagation and weight optimization during training. This automatic feature learning is precisely what distinguishes deep learning from classical machine learning, where engineers must manually engineer and select features.

Why A is wrong: Model training is not automatic - a human must define the architecture, set hyperparameters, supply labeled data, and initiate/monitor the process. The model doesn't train itself unprompted.

Why C is wrong: Analyzing the problem and positioning the task (i.e., deciding what to solve and how to frame it) is a human responsibility requiring domain expertise and business judgment - no model can define its own objective.

Memory tip: Remember the phrase "Deep learning does the Feature Work (eXtraction + Selection) automatically." Anything that requires human judgment - framing the problem (C) or setting up training (A) - stays with the person. Anything that involves learning from data internally - finding and prioritizing patterns (B, D) - is handled by the model.

No community discussion yet for this question.