Learn Before
Concept icon
Concept

Methods to Reduce High Avoidable Bias

If a model is showing high avoidable bias, several practical fixes can help: increase model capacity, revise the input features using error analysis, reduce or remove regularization, or adjust the architecture so it better matches the task. Simply adding more training examples is usually not the right response for this problem, because extra data often improves variance more than bias.

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