Essay

How More Compute Helps When Training on Very Large Datasets

Question: Explain why increased computational capacity can accelerate progress in deep learning, especially when the training data set is very large.

Sample answer: More computational capacity makes it possible to train larger neural networks or train models for longer. That matters because very large data sets often contain patterns that smaller models cannot capture well. When compute is limited, a model may not be big enough to make full use of the available data. With enough compute, the model can scale to a size that can better learn from the large data set.

Key points:

  • More computational capacity supports training larger models.
  • Larger models can better use very large data sets.
  • Without enough compute, the model may be too small to benefit fully from the data.

Rubric: The response must explain that extra compute enables larger neural networks and that those larger networks are needed to take advantage of very large data sets.

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