Learn Before
Case Study

Determine the labeling workflow for a new medical imaging dataset.

Case context: A machine learning team is building a dataset of medical images. They have access to one junior doctor and a panel of three highly experienced senior doctors, whose time is very expensive.

Question: Based on Andrew Ng's advice, how should the team assign the labeling tasks among the junior doctor and the senior doctors to balance quality and cost?

Sample answer: The team should have the single junior doctor review and label the vast majority of the medical images. If the junior doctor encounters particularly difficult or ambiguous cases, those specific images should be brought to the panel of senior doctors for discussion and labeling.

Key points:

  • Assign the vast majority of standard cases to the junior doctor.
  • Identify hard cases during the initial junior doctor review.
  • Escalate only the hard cases to the expensive senior doctors.

Rubric: The answer must clearly propose assigning the bulk of the labeling to the junior doctor and reserving the panel of senior doctors exclusively for the hard cases.

0

1

Updated 2026-05-27

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