NEW
Font size
WorksheetsFederated Learning Quiz
Total questions: 20
Worksheet time: 10mins
Which feature most effectively differentiates Federated Learning from Centralized Machine Learning?
Higher computational power at server
Centralized data storage
Local model training on client data
Faster convergence rate
Federated Learning primarily emerged due to:
Increased GPU availability
Growth of cloud-only applications
Privacy regulations and edge device proliferation
Decline in distributed systems
Which scenario best justifies using Federated Learning over Distributed ML?
High-speed data center networks
Public benchmark datasets
Sensitive user data across personal devices
Homogeneous server clusters
In comparison to Distributed ML, Federated Learning reduces privacy risks mainly because:
Data is encrypted at rest
Raw data never leaves client devices
Models are smaller
Training is asynchronous
Which limitation is most prominent in Federated Learning but not in Centralized ML?
Model bias
Communication overhead
Overfitting
Feature scaling
Horizontal Federated Learning is best suited when:
Clients share features but not samples
Clients share samples but not features
Clients share neither samples nor features
Clients share labels only
Vertical Federated Learning is most applicable when:
Data samples are identical across clients
Feature spaces are identical
Data belongs to the same users but different features
Training happens on edge devices only
Transfer Learning based Federated Learning is preferred when:
Clients have identical datasets
Data distributions differ significantly across clients
Bandwidth is unlimited
Labels are unavailable
Which factor makes Vertical FL more complex than Horizontal FL?
Larger datasets
Requirement of secure feature alignment
Absence of aggregation
Higher bandwidth
In the Federated Learning ecosystem, the central server’s primary role is to:
Store raw training data
Perform feature extraction
Aggregate client model updates
Label datasets
Which stakeholder directly controls data privacy in Federated Learning?
Central server
Cloud provider
Data owners / clients
Application developer
Malicious clients in Federated Learning mainly threaten:
Storage efficiency
Model integrity and convergence
Feature selection
Training speed only
Which communication factor most affects training latency in Federated Learning?
Batch size
Feature dimensionality
Frequency of model updates
Learning rate
Compared to centralized training, Federated Learning requires more:
Storage
Model layers
Network coordination
Labeled data
Which technique is commonly used to reduce bandwidth usage in FL?
Data augmentation
Model compression
Feature normalization
Dropout
Limited bandwidth in Federated Learning primarily results in:
Higher model accuracy
Faster aggregation
Slower convergence
Reduced model size
Edge Computing improves Federated Learning mainly by:
Centralizing computation
Reducing latency and data transmission
Increasing cloud dependency
Eliminating communication
Why are IoT devices challenging participants in Federated Learning?
They generate less data
They lack network interfaces
They have limited compute and power resources
They do not support ML models
Which property makes Federated Learning suitable for large-scale IoT networks?
Centralized storage
Device-level autonomy
High memory usage
Offline training
Which combination best represents a federated system?
Cloud server + single dataset
Edge devices + local training + central aggregation
GPUs + shared memory
IoT sensors + centralized database
