Short Answer

What risk comes from training on web-scraped photos when the deployment data will differ?

Question: In one to three sentences, explain the main risk of starting with photos collected from public web pages when the examples the system will face later are not similar.

Sample answer: The main risk is that the model learns patterns from the web images that do not match the later real-world inputs, so it may perform poorly once it is deployed. In other words, the training data and the eventual data distribution are too different.

Key points:

  • Web-scraped photos may not match later inputs
  • The central problem is weak generalization to the deployment distribution

Rubric: Full credit: the response identifies mismatch between training and deployment data and names poor generalization. Partial credit: the response mentions mismatch or risk without clearly connecting it to generalization. No credit: the response describes a different issue.

0

1

Updated 2026-08-12

Contributors are:

Who are from:

Tags

Machine Learning

Deep Learning

Machine Learning Strategy

Supervised Learning

Dive into Deep Learning @ D2L

Data Science

Machine Learning Yearning @ DeepLearning.AI