Short Answer

What makes combining training data from multiple sources risky?

Question: In one to three sentences, what factor mainly determines whether combining training data from different sources, such as store photos and studio photos, is risky?

Sample answer: The main factor is how flexible the learning algorithm is. Less expressive models with fixed feature engineering can lose accuracy when mixed data introduces mismatch, while highly expressive modern models usually tolerate mixed sources with much less risk.

Key points:

  • Flexibility of the learning algorithm is the key factor
  • Older, rigid approaches were more vulnerable to harm from mixing datasets
  • Modern flexible models are usually less affected

Rubric: Full credit identifies model flexibility and contrasts older rigid methods with modern flexible methods.

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