ตรวจข้อสอบ > ปัณณทัต ปลื้มปรีดาพร > Medical & Health Sciences (Secondary Level) | สาขาการแพทย์และสุขภาพ ระดับมัธยมศึกษา > Part 2 > ตรวจ

ใช้เวลาสอบ 24 นาที

Back

# คำถาม คำตอบ ถูก / ผิด สาเหตุ/ขยายความ ทฤษฎีหลักคิด/อ้างอิงในการตอบ คะแนนเต็ม ให้คะแนน
1


How does the concept of “model as a dataset” reshape traditional data-sharing practices in medical imaging?

3. It enables sharing of learned model weights instead of sensitive raw images.

It allows for multicenter collaborations due to being able to share model weights and not reveal patient information. Under "Summary" : "enables privacy-preserving multicentre collaborations" 7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

2


Which analytical conclusion can be drawn about the trade-offs between physics-informed and statistical models?

2. Physics-informed models are more interpretable but computationally intensive.

The writers states that physics informed model can make higher fidelity data but needs lots of computational power and thus being expensive Under "Generative models" : " physics informed models offer high fidelity" and "require extensive domain expertise and computational resources" 7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

3


Why is “mode collapse” considered a critical problem in GAN-based medical image synthesis?

2. It reduces image realism and variety by producing repetitive outputs.

Mode collapse happen when not all data variation are covered and the model outputting similar images over and over again Under "Generative models" : "might not always cover all data variation" and "leading to mode collapse" 7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

4


Why are healthcare-specific metrics preferred over general-purpose metrics such as FID or SSIM?

2. They better capture clinical accuracy and diagnostic relevance.

Healthcare-specific metrics are purpose built for the job the images are going to do and thus having more relevance in the field Under "Health-care-specific metrics" : "tailored to health-care needs" 7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

5


What does the article identify as the key tension between privacy preservation and image fidelity?

1. Higher realism may risk reproducing identifiable patient data.

When a model is trained on a set of data and is able to recreate the data with high realism, it could also recreate an identifiable part of a patient. But removing identifiable features could also remove variation and diversity from the training sample Under "Patient privacy and data copying" : "trained on a specific data set", "reveal sensitive patient information", "raises concern about the degree of anonymisation" 7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

6


Why is the FDA’s approval of synthetic MRI technology significant for future AI-generated data?

1. It establishes a framework for validating synthetic data equivalence in clinical use.

It is a proof of performance that synthetic images in medicine works and does not degrade the performance of the radiologist. Under "Future directions" : "proof of performance", "performance of radiologist remained equivalent" 7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

7


Which strategy would best mitigate demographic bias in generative models according to the article?

2. Applying diversity-aware training and fairness constraints

The writer states that targeting the minorities is able to close the fairness gap thus raising awareness is able to mitigate the demographic bias. Under "Increased dataset size and diversity" : "targeted oversampling of minoritised ... close the fairness gap by 40%" 7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

8


How do DDPMs exemplify versatility in healthcare image synthesis?

2. They can perform multiple tasks such as denoising, inpainting, and anomaly detection without retraining.

They can be repurposed for other tasks without needing to retrain the model to that specific task. Under Versatility across talks" : "can be adapted and repurposed", "without any further training can also be used for inpainting" 7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

9


What analytical insight does the article provide about integrating AI-generated medical images into education and research?

2. It enhances training by providing diverse, realistic datasets without ethical breaches.

It increases the sample size and thus the diversity of medical images for education. They are also more anatomically accurate than their real counterpart Under "Modelling complex biological phenomena" : serving as virtual surgical planning tools and educational resources", "depict the potential progression of a brain tumor over time" 7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

10


Why is regional calibration essential when applying risk prediction models across countries?

2. To adjust for population-specific incidence and lifestyle differences

The need for calibration is because countries don't have the same lifestyle ignoring factors that the existing model accounted for could lead to over or underestimation of risk. 7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

11


What analytical conclusion can be drawn when comparing the China-PAR and Framingham models?

1. Both overestimate CVD risk in East Asians.

The article states that both of the models overestimate the risk in china Under "ASCVD risk predictions in China" : "poor prediction for chinese men" 7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

12


Based on CVD mortality data, what analytical inference can be made about Japan’s position compared to neighboring countries?

4. Japan’s diet increases risk compared to Mongolia.

7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

13


What analytical limitation arises when using Western-derived coefficients in East Asian models?

2. It introduces systematic overestimation of ASCVD probability.

7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

14


What policy implication can be derived from country-specific risk models?

1. They allow for targeted national prevention programs.

7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

15


If a model excludes socioeconomic variables, what analytical consequence might occur?

2. Ignored non-biological determinants of disease

7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

16


How might AI improve next-generation ASCVD risk prediction in East Asia?

2. By integrating multimodal data, including imaging and lifestyle information

7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

17


What conclusion can be drawn from comparing Mongolia’s and South Korea’s CVD mortality rates?

2. Both have identical age-adjusted mortality rates.

7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

18


What is the most logical future direction for improving ASCVD models across East Asia?

3. Removing local variability from analysis

7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

19


According to the “image generation trilemma” shown in the figure, what analytical conclusion can be drawn about the relative strengths of VAEs, GANs, and DDPMs in medical image synthesis?

2. GANs provide a balance between image quality and diversity but may suffer from mode collapse.

The writer states that GANs is the middle ground for image quality and diversity but could suffer from a mode collapse which reduces the diversity. Under "Figure 2" : "strike a balance, providing good quality and diversity but can suffer from a mode collapse" 7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

20


Based on Figure, what analytical conclusion can be drawn regarding the distribution of cardiovascular disease (CVD) subtypes across East Asian countries?

2. Stroke dominates as the primary cause of CVD death in all East Asian countries equally.

The pie chart shows that stroke dominates as the leading cause of death in the cardiovascular disease category in east asian countries the stroke part of the pie chart takes up 48%, 39%, 47% ,and 47% respectively showing a high value 7

-.50 -.25 +.25 เต็ม 0 -35% +30% +35%

ผลคะแนน 76.1 เต็ม 140

แท๊ก หลักคิด
แท๊ก อธิบาย
แท๊ก ภาษา