NEW
Font size
WorksheetsAI in Healthcare Unit 2 Review
Total questions: 26
Worksheet time: 13mins
What type of AI would be most appropriate for the task of analyzing symptoms and vital signs to prioritize care in the ER?
Strong AI - because it needs to think like a human doctor
Weak AI - designed for the narrow task of symptom analysis
General AI - to handle all possible medical situations
Virtual Assistant - like Siri but for hospitals
How would you classify the X-ray images in terms of data structure?
Structured Data - they're organized in folders
Semi-Structured Data - they have some metadata
Unstructured Data - free-form visual information
Patient Admission Records - they're medical data
What must be done before using patient data for research in an AI study on diabetes prediction?
Just remove the names and it's fine to use
Get patient consent and properly anonymize all PHI
Only use the data if it's stored in the cloud
Ask the hospital's IT department for permission
What's the most important consideration when the AI system notices that patients who drink more coffee tend to have fewer heart attacks?
This proves coffee prevents heart attacks - recommend it immediately
This shows correlation, but we need more research to prove causation
Coffee is always bad for heart patients regardless of data
The AI system must be malfunctioning
What type of analytics is the hospital implementing to forecast which patients are likely to be readmitted within 30 days?
Descriptive Analytics - describing what happened before
Diagnostic Analytics - figuring out why readmissions occur
Predictive Analytics - forecasting future readmissions
Prescriptive Analytics - telling doctors what to do
What's the most likely cause of an AI diagnostic tool misdiagnosing conditions in patients from different ethnic backgrounds?
The AI needs more processing power
Algorithmic bias due to unrepresentative training data
The patients aren't describing symptoms correctly
The AI system needs to be updated to the latest version
What's a key advantage of cloud storage for a small rural clinic's new AI system?
Cloud storage is always more secure than local storage
Remote access and automatic backups without local IT staff
Cloud storage is free and has no ongoing costs
Patient data is automatically anonymized in the cloud
In a situation where a patient is anxious about their cancer diagnosis and needs emotional support, what's the best approach?
Let the AI handle everything - it has all the medical data
Ignore the AI analysis and rely only on human judgment
Use AI for data analysis while humans provide emotional support
Replace all doctors with AI systems for consistency
Which of the following best describes the role of AI in automating medical image analysis?
AI is not used for medical image analysis
AI only stores images without analysis
AI assists radiologists by highlighting potential abnormalities
AI replaces radiologists entirely
What is a potential risk when deploying AI systems in healthcare without proper validation?
AI systems will always make correct decisions
Patient privacy is automatically protected
Incorrect diagnoses due to untested algorithms
Healthcare costs will immediately decrease
For a machine learning process, how should you split data for training and evaluation?
Use features for training and labels for evaluation
Randomly split the data into rows for training and rows for evaluation
Use labels for training and features for evaluation.
Do not split, use the same data for training and evaluation
What does HIPAA primarily regulate?
Data visualization methods
Patient privacy and healthcare data security
Types of medical imaging devices
Healthcare provider licensing
Which step is essential to ensure fairness when training an AI model on healthcare data?
Ignore demographic information
Use only data from one hospital
Train the model without any validation
Include diverse patient populations in the training data
What is a primary benefit of using AI for early disease detection in medical records?
AI can instantly cure diseases
AI can identify patterns that may be missed by humans
AI replaces all medical staff
AI only works with structured data
Why is it important to validate AI predictions with clinical experts before implementation?
AI systems never make mistakes
Validation ensures AI recommendations are medically sound
Validation slows down the adoption of AI
Clinical experts are not needed if the AI is accurate
Which type of learning uses labeled data to train models?
Unsupervised learning
Supervised learning
Reinforcement learning
Transfer learning
This model predicts if this is a shoe or not a shoe.
regression
classificiation
What is the main goal of unsupervised learning?
To maximize rewards
To classify labeled data
To find patterns in unlabeled data
To transfer knowledge between tasks
Which learning type does NOT require labeled data?
Supervised learning
Unsupervised learning
Transfer learning
Reinforcement learning
Which learning type is most suitable for clustering similar items together?
Supervised learning
Unsupervised learning
Transfer learning
Reinforcement learning
Which learning type is commonly used for image classification tasks?
Unsupervised learning
Reinforcement learning
Supervised learning
Transfer learning
Which learning type is most appropriate for grouping customers based on purchasing behavior without prior labels?
Supervised learning
Unsupervised learning
Transfer learning
Reinforcement learning
A teacher provides students with answers to math problems, and students learn to solve similar problems. Which learning type does this resemble?
Supervised learning
Unsupervised learning
Transfer learning
Reinforcement learning
Explain why supervised learning might not be suitable for tasks where labeled data is scarce. Choose the best reason.
It requires a lot of computational power.
It cannot handle large datasets.
It relies on labeled data for training.
It is only used for clustering.
A student wants to build a model to predict house prices but only has data about house features, not prices. Which learning type should they use and why?
Supervised learning, because it works with features only.
Unsupervised learning, because there are no labels.
Transfer learning, because it uses knowledge from other tasks.
Reinforcement learning, because it uses rewards.
Predicting the breed of a dog is an example of a _________ machine learning problem.
Regression
Gradient Descent
Classification
Generalized
