引導式測驗 · 中等 等級
兒科人工智慧測驗
測試為什麼兒童資料會讓 AI 變得更加困難、NICU 心率監測如何預測敗血症,以及為什麼成人訓練的影像模型會誤讀兒童的 X 光片。
問題 1 的 8
為什麼接受成人 X 光訓練的人工智慧可能會誤讀 10 歲兒童的骨骼影像?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- 為什麼接受成人 X 光訓練的人工智慧可能會誤讀 10 歲兒童的骨骼影像?
- 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
- HeRO 監測儀分析什麼來評估早產兒敗血症風險?
- Heart rate characteristics such as variability and decelerations
- Blood glucose readings every hour
- Parent-reported feeding diaries
- Chest X-ray images
- Cognoa 於 2021 年獲得 FDA 授權的 Canvas Dx 在自閉症護理中發揮什麼作用?
- 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
- 該指南給出了兒科數據集較小的原因是什麼?
- 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
- 為什麼從業者要按年齡層報告兒科模型的準確性?
- 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
- 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
- Which approach lets several children's hospitals train a shared model without pooling raw patient data?
- Bone age estimation
- 數據增強
- Alert fatigue reduction
- Federated learning
- 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