Advanced topics and ethics Flashcards
What is MLOps?
a) A data visualization technique
b) The process of managing the machine learning lifecycle from development to deployment and monitoring
c) A method for manual machine learning workflows
d) A clustering algorithm for data mining
b) The process of managing the machine learning lifecycle from development to deployment and monitoring
What is a key goal of MLOps?
a) Optimizing test datasets
b) Bridging the gap between development, deployment, and operation to ensure consistency and reliability
c) Manual tracking of model versions
d) Eliminating automation in ML workflows
b) Bridging the gap between development, deployment, and operation to ensure consistency and reliability
Which principle is NOT part of MLOps?
a) Continuous integration and delivery
b) Version control
c) Manual experimentation
d) Model governance
c) Manual experimentation
What are information systems in the context of machine learning?
a) Systems focused solely on storing data
b) Systems that integrate ML models into business workflows for decision-making
c) Tools used only for manual data entry
d) Static systems without automation
b) Systems that integrate ML models into business workflows for decision-making
What is a primary role of information systems in ML?
a) Simplifying neural network architectures
b) Enabling organizations to derive insights from data-driven decisions
c) Minimizing data processing steps
d) Manually managing data pipelines
b) Enabling organizations to derive insights from data-driven decisions
What is PlantUML used for?
a) Creating dynamic web applications
b) Generating UML diagrams from simple textual descriptions
c) Optimizing machine learning algorithms
d) Designing neural networks
b) Generating UML diagrams from simple textual descriptions
Which type of diagrams can PlantUML create?
a) Sequence diagrams
b) Data processing pipelines
c) Neural network architectures
d) None of the above
a) Sequence diagrams
Why are case studies important in MLOps?
a) They provide a way to test unrelated hypotheses
b) They offer real-world insights into implementing MLOps best practices
c) They eliminate the need for monitoring ML models
d) They focus on theoretical ML workflows only
b) They offer real-world insights into implementing MLOps best practices
In an MLOps case study, what is a typical outcome?
a) An abstract ML algorithm without application
b) A complete lifecycle showcasing model development, deployment, and monitoring
c) A fixed solution to all ML problems
d) A single static model version
b) A complete lifecycle showcasing model development, deployment, and monitoring
What is machine learning ethics?
a) A set of rules for training models
b) The study of moral implications and responsibilities in designing and using ML systems
c) A branch of neural network optimization
d) Guidelines for coding in Python
b) The study of moral implications and responsibilities in designing and using ML systems
Why is machine learning ethics important?
a) To improve model training accuracy
b) To ensure ML models are used responsibly and do not cause harm
c) To automate all decision-making processes
d) To reduce training costs
b) To ensure ML models are used responsibly and do not cause harm
What does fairness mean in machine learning?
a) Ensuring that all models are highly accurate
b) Ensuring that ML models do not disproportionately favor or disadvantage any group
c) Ensuring models run equally fast on all hardware
d) Ensuring training data is perfectly balanced
b) Ensuring that ML models do not disproportionately favor or disadvantage any group
Which of these is an example of fairness in ML?
a) A hiring model that treats all candidates equally regardless of gender or ethnicity
b) A model that only predicts for the majority group
c) A system designed to focus on profitability over equity
d) A training algorithm that ignores outliers
a) A hiring model that treats all candidates equally regardless of gender or ethnicity
What does transparency mean in the context of machine learning?
a) Making all datasets available to the public
b) Providing clear explanations of how models make decisions
c) Ensuring all code is written in plain text
d) Disclosing the financial costs of model training
b) Providing clear explanations of how models make decisions
Why is transparency critical in machine learning?
a) To make models more complex
b) To build trust and allow stakeholders to understand how decisions are made
c) To ensure all data is equally weighted
d) To avoid sharing model performance results
b) To build trust and allow stakeholders to understand how decisions are made
