WorksheetsFAM_M10_FinalsQ1
Total questions: 25
Worksheet time: 13mins
Deploying a model into a live business system is known as:
Model Training
Model Deployment
Model Testing
Data Cleaning
Which type of deployment processes data in groups or intervals?
Real-time deployment
Batch deployment
Edge deployment
Cloud deployment
Which deployment type provides instant predictions?
Batch
Real-time
Periodic
Offline
A self-driving car using onboard sensors uses which deployment type?
Cloud
Batch
Edge
On-premises
Hospitals usually choose which deployment type for privacy reasons?
Cloud
On-premises
Real-time
Batch
What is the main purpose of model monitoring?
Increase cost
Detect drift and ensure reliability
Stop model updates
Reduce accuracy
Which factor ensures a system can handle more data over time?
Latency
Scalability
Compatibility
Simplicity
Real-time fraud detection requires:
Batch deployment
Edge deployment
Real-time deployment
Cloud-only deployment
The term “model drift” refers to:
Model improvement
Changes in data or relationships that reduce accuracy
Software update
Hardware issue
Data drift happens when:
The input data’s distribution changes
The output labels disappear
Hardware fails
Nothing changes
Concept drift occurs when:
Input-output relationship changes
Data becomes larger
Accuracy improves
System crashes
Retraining a model means:
Deleting it
Training it again with new data
Ignoring old data
Stopping it permanently
Which tool is commonly used for model tracking and deployment?
Power BI
MLflow
Excel
Tableau
Which monitoring tool is designed for TensorFlow models?
PyTorch
TensorFlow Serving
DataRobot
MLflow
Which platform offers commercial tools for deployment and monitoring?
TensorFlow
DataRobot
Jupyter Notebook
Excel
The main advantage of cloud deployment is:
High privacy
Scalability and flexibility
Data isolation
Manual updates
The main advantage of edge deployment is:
Long response times
Low latency and independence
Heavy computation
High network usage
What ensures smooth integration with existing systems?
Compatibility
Latency
Accuracy
Cost
In model lifecycle, deployment comes after:
Data collection
Model testing
Model development
Model validation
Which is NOT a model monitoring activity?
Drift detection
Performance tracking
Retraining
Data cleaning
Model deployment is the process of building a model.
True
False
Batch deployment happens instantly.
True
False
Real-time deployment provides immediate responses.
True
False
On-premises deployment gives better data privacy.
True
False
Cloud deployment is less flexible than on-premises.
True
False
