Adhering to principles like Microsoft’s responsible AI guidelines requires extra steps when using automated machine learning. Learn how explaining models and tuning for fairness can lead to more transparent, ethical AI.
Table of Contents
Question
You build a machine learning model by using the automated machine learning user interface (UI). You need to ensure that the model meets the Microsoft transparency principle for responsible AI. What should you do?
A. Set Validation type to Auto.
B. Enable Explain best model.
C. Set Primary metric to accuracy.
D. Set Max concurrent iterations to 0.
Answer
B. Enable Explain best model.
Explanation
Model Explain Ability.
Most businesses run on trust and being able to open the ML “black box” helps build transparency and trust.
In heavily regulated industries like healthcare and banking, it is critical to comply with regulations and best practices. One key aspect of this is understanding the relationship between input variables (features) and model output. Knowing both the magnitude and direction of the impact each feature (feature importance) has on the predicted value helps better understand and explain the model. With model explain ability, we enable you to understand feature importance as part of automated ML runs.
The correct answer is B. You should enable Explain best model to ensure that the model meets the Microsoft transparency principle for responsible AI.
The transparency principle for responsible AI means that people who create and use AI systems should be open about how and why they are using AI, and open about the limitations of the system. This principle involves providing clear explanations and justifications for the decisions and actions of AI systems, and enabling users to understand and interact with AI systems in meaningful ways.
By enabling Explain best model, you are following the transparency principle. Explain best model is a feature of the automated machine learning UI that allows you to generate and view model explanations for the best performing model in your experiment. Model explanations are visualizations and metrics that help you understand how the model makes predictions, what features are important for the model, and how the model behaves for different data points. Model explanations can help you debug and validate your model, as well as communicate your model to stakeholders and customers.
The other options are not related to the transparency principle, but to other aspects of the automated machine learning UI.
Setting Validation type to Auto is related to the reliability and safety principle, which means that AI systems should perform reliably and safely. Validation type is a parameter that determines how the automated machine learning UI splits your data into training and validation sets, and how it evaluates the performance of the models. Setting Validation type to Auto allows the automated machine learning UI to choose the best validation method based on your data size and characteristics. This can help you ensure that your model is robust and generalizable to new data.
Setting Primary metric to accuracy is related to the fairness principle, which means that AI systems should treat all people fairly. Primary metric is a parameter that determines how the automated machine learning UI ranks and compares the models. Setting Primary metric to accuracy allows the automated machine learning UI to choose the model that has the highest proportion of correct predictions. However, accuracy is not always the best metric for fairness, as it can mask the performance of the model for different subgroups of the data. You may need to consider other metrics, such as precision, recall, or F1-score, to evaluate the fairness of your model.
Setting Max concurrent iterations to 0 is related to the performance and efficiency of the automated machine learning UI, but not to any responsible AI principle. Max concurrent iterations is a parameter that determines how many iterations (model and hyperparameter combinations) can run in parallel on your compute target. Setting Max concurrent iterations to 0 disables parallel execution and runs the iterations sequentially. This can slow down the experiment and increase the training time, but it can also reduce the compute cost and resource consumption.
References
Microsoft Docs > Browse Identify guiding principles for responsible AI > Identify guiding principles for responsible AI
Microsoft Azure AI Fundamentals AI-900 certification exam practice question and answer (Q&A) dump with detail explanation and reference available free, helpful to pass the Microsoft Azure AI Fundamentals AI-900 exam and earn Microsoft Azure AI Fundamentals AI-900 certification.
