Metadata
Technology & Computer Science Graduate Create Hard
Metadata
  • Subject

    Technology & Computer Science

  • Education level

    Graduate

  • Cognitive goals

    Create

  • Difficulty estimate

    Hard

  • Tags

    federated learning, differential privacy, byzantine robustness, IoT, edge computing, scalability

  • Number of questions

    5

  • Created on

  • Generation source

  • License

    CC0 Public domain

  • Prompt

    Assess students' ability to create a novel, scalable federated learning architecture for heterogeneous IoT edge devices that provides provable differential privacy and formal robustness against Byzantine clients. The quiz should test specification of system components (edge, aggregator, hierarchy), algorithms for DP (mechanisms, privacy accounting, composition), Byzantine-robust aggregation strategies and proofs/sketches of robustness under a stated threat model, methods for handling device heterogeneity and resource constraints (partial participation, personalization, compression), scalability techniques (communication reduction, hierarchical/federated aggregation), formal statements and proofs or proof sketches of guarantees, evaluation metrics, and an experimental validation plan.
Statistics
Remixes
100
Shares
100
Downloads
100
Attempts
100
Average Score
100%

Mock data used for demo purposes.