Gids-gekoppelde quiz · Middelmatig Niveau

AI in kindergeneeskundequiz

Test waarom gegevens van kinderen AI moeilijker maken, hoe hartslagmonitoring op de NICU sepsis voorspelt, en waarom door volwassenen getrainde beeldvormingsmodellen röntgenfoto's van kinderen verkeerd kunnen interpreteren.

Gerelateerde gidspadenAI in de kindergeneeskunde
Vraag 1 van 8

Why might an AI trained on adult X-rays misread a ten-year-old's bone image?

Bekijk elke vraag8 vragen

Antwoorden en uitleg verschijnen terwijl je de quiz doet.

  1. Why might an AI trained on adult X-rays misread a ten-year-old's bone image?
    • Children's X-rays are always in color
    • Normal growth plates in children can resemble fracture lines in adults
    • Children's bones do not appear on X-rays
    • Pediatric X-rays use no radiation
  2. What does the HeRO monitor analyze to estimate sepsis risk in premature infants?
    • Heart rate characteristics such as variability and decelerations
    • Blood glucose readings every hour
    • Parent-reported feeding diaries
    • Chest X-ray images
  3. What role does Cognoa's Canvas Dx, authorized by the FDA in 2021, play in autism care?
    • It replaces the need for any clinician
    • It is a genetic test for autism
    • It treats autism with behavioral games
    • It is an aid that supports clinicians in diagnosing autism in young children
  4. Which reason does the guide give for pediatric datasets being small?
    • Children's data is never stored electronically
    • Hospitals are banned from using children's data in research
    • Children are healthier on average and rare diseases are spread thin
    • Pediatric images are too large to store
  5. Why do practitioners report pediatric model accuracy by age band?
    • Regulators forbid pooled numbers
    • A pooled figure can hide poor performance in the youngest group
    • Age bands make the model train faster
    • Older children cannot be imaged
  6. What is the risk of fine-tuning an adult-pretrained network on a small pediatric dataset?
    • It can import adult assumptions, so external validation is still essential
    • It always makes the model worse than random
    • It deletes the pediatric data
    • It cannot run on hospital computers
  7. Which approach lets several children's hospitals train a shared model without pooling raw patient data?
    • Bone age estimation
    • Gegevensvergroting
    • Alert fatigue reduction
    • Federated learning
  8. What misconception about adult AI tools does the guide warn against?
    • That they are too expensive for children's hospitals
    • That they cannot read digital images
    • That they can simply be used on children without separate validation
    • That they require parental consent to run