Kvíz spojený s průvodcem · Snadno úroveň

In-House vs Outsourced Data Labeling Quiz

Compare internal, vendor, and hybrid annotation models across expertise, capacity, privacy controls, labor practices, and quality assurance.

Související vodicí cestyIn House Vs Outsourced Data Labeling
Otázka 1 z 8

Why might a company building a niche medical imaging model choose in-house labeling over outsourcing?

Náhled každé otázky8 otázky

Odpovědi a vysvětlení se zobrazují, jak budete v kvízu.

  1. Why might a company building a niche medical imaging model choose in-house labeling over outsourcing?
    • The specialized domain expertise required is hard to source cheaply from a general labeling vendor
    • In-house labeling is always cheaper per label
    • Outsourced vendors are legally barred from labeling any images
    • In-house teams never need quality control
  2. What can an external labeling vendor offer on a large, clearly specified task?
    • Guaranteed lower cost and higher accuracy in all cases
    • Potentially added capacity, depending on the provider and contract
    • Automatic compliance without review
    • The right to ignore the task guidelines
  3. Why might a project restrict who can annotate sensitive or classified data?
    • Because outsourcing is always illegal
    • Because annotation cannot be done on encrypted systems
    • To meet applicable access, confidentiality, residency, or classification requirements
    • To avoid writing guidelines
  4. According to the guide, what is a gold-standard task in an outsourced labeling quality assurance process?
    • A task that pays labelers a bonus rate
    • A task reserved only for the most senior in-house engineers
    • A task that requires no review at all
    • A pre-labeled example with a known correct answer, seeded invisibly into a labeler's queue to measure accuracy
  5. How does consensus labeling work as a quality mechanism?
    • The same item is sent to multiple independent labelers, and the majority or weighted-agreement answer is kept
    • It relies on a single labeler's judgment being trusted completely
    • It removes the need for any labeling guidelines
    • It only applies to audio annotation tasks
  6. Why is it a misconception that outsourcing always means lower quality?
    • Outsourcing is technically impossible for any task
    • A well-managed vendor may produce reliable labels when expertise, instructions, calibration, and QA fit the task
    • In-house teams never make mistakes
    • Outsourced vendors are legally required to guarantee perfect accuracy
  7. According to the guide, what is the hybrid approach described for combining in-house and outsourced labeling?
    • Using only in-house labelers forever and never outsourcing
    • Outsourcing the entire project with no in-house involvement at all
    • Refining ambiguous labeling guidelines with a small in-house team first, then scaling the finalized instructions out to an outsourced vendor
    • Alternating randomly between in-house and outsourced labelers with no plan
  8. Do privacy or sector-specific duties automatically require annotation to remain in-house?
    • Yes, all regulated information must be labeled by employees
    • No; applicable rules may restrict access, transfer, and safeguards, but the sourcing choice depends on the actual requirements
    • Yes, but only for every medical image
    • No, privacy rules never apply to data vendors