Concept icon
Concept

Choosing the Weight for Auxiliary Data

When training on a main dataset together with a larger auxiliary dataset, a mixing coefficient such as beta can control how much the auxiliary examples influence the objective. For instance, if 6,000 labeled storefront images are combined with 144,000 web images, a beta value of 1/24 would make their total contributions equal. In practice, beta is then adjusted by checking performance on the development set.

0

1

Concept icon
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