Short Answer

Irrelevant Training Data and Model Capacity

Question: If a training set contains examples that do not help solve the target task, what harm can they do to a neural network beyond increasing compute time?

Sample answer: They can consume model capacity by pushing the network to allocate parameters to patterns that are not useful for the real task.

Key points:

  • They use up model capacity.
  • They can cause the network to spend parameters on irrelevant patterns.
  • Those patterns may not match the distribution the model will face at evaluation time.

Rubric: The response must state that the irrelevant data uses up the model's capacity.

0

1

Updated 2026-08-12

Contributors are:

Who are from:

Tags

Machine Learning

Deep Learning

Supervised Learning

Dive into Deep Learning @ D2L

Data Science

Machine Learning Strategy

Machine Learning Yearning @ DeepLearning.AI