Quiz połączony z przewodnikiem · Średni Poziom

Quiz dotyczący sztucznej inteligencji w pediatrii

Testuje, dlaczego dane dzieci utrudniają sztuczną inteligencję, jak monitorowanie tętna OIOM-u pozwala przewidzieć sepsę i dlaczego modele obrazowania przeszkolone przez dorosłych mogą błędnie odczytać zdjęcia rentgenowskie dzieci.

Powiązane ścieżki prowadząceAI w pediatrii
Pytanie 1 z 8

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

Podgląd każdego pytania8 pytania

Odpowiedzi i wyjaśnienia pojawiają się w miarę rozwiązywania quizu.

  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
    • Zwiększanie danych
    • 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