AI developers often find it difficult to explain the processes inside deep learning systems PRIMARILY because:

Question

Options

• ATraining data input for learning is spread throughout the public domain and continues to change
• BGenerated knowledge dynamically changes in memory without being tracked by change history
• CApplied algorithms are based on probability theories to improve system performance
• DNeural network architectures can include statistical methods that are not fully understood

• A
  3% (1)
• B
  3% (1)
• C
  8% (3)
• D
  87% (33)

Explanation

Deep neural networks are often called 'black boxes' because they involve millions or billions of parameters organized in complex hierarchical layers. The internal representations - such as learned features in hidden layers - involve high-dimensional statistical transformations that are not intuitively interpretable, even by the researchers who designed the architecture. This is a well-known challenge in the field of explainable AI (XAI). The other options describe real phenomena (changing training data, dynamic memory changes, probability-based algorithms) but are not the PRIMARY reason for the explainability gap specific to deep learning.

No community discussion yet for this question.