You create a binary classification model. You use the Fairlearn package to assess model fairness.You must eliminate the need to retrain the model.You need to implement the Fairlearn package.Which algorithm should you use?
Answer(s): D
The best algorithm to implement within the Fairlearn package for a binary classification model without retraining is the ThresholdOptimizer.Why ThresholdOptimizer? This algorithm is a post-processing technique specifically designed to mitigate unfairness after a model has already been trained. By setting the prefit parameter to True, it treats your existing model as a "black box," adjusting its decision thresholds for different groups to satisfy fairness constraints (like demographic parity or equalized odds) without changing the underlying model weights.
https://learn.microsoft.com/en-us/azure/machine-learning/concept-fairness-ml
DRAG DROP (Drag and Drop is not supported)A real-time endpoint is deployed in Azure Machine Learning to serve predictions to a web application.Users report intermittent failures and unexpected responses when calling the endpoint.You need to identify the appropriate troubleshooting action for each reported issue.Which troubleshooting action should you perform for each issue? To answer, move the appropriate troubleshooting actions to the correct issues. You may use each troubleshooting action once, more than once, or not at all. You may need to move the split bar between panes or scroll to view content.Note: Each correct selection is worth one point.Select and Place:
Answer(s): A
Box 1: Review Deployment Logs Requests fail after deployment updates.To troubleshoot intermittent failures and unexpected responses in an Azure Machine Learning (AML) real-time endpoint after a deployment update, you should focus on container logs, liveness/readiness probes, and local debugging. Microsoft LearnCore Troubleshooting Steps *-> Review Deployment Logs: Check the AmlOnlineEndpointConsoleLog in Azure Monitor to identify startup crashes or runtime exceptions in your score.py file.Inspect HTTP Status CodesValidate Probe SettingsDebug LocallyBox 2: Validate scoring script and model version. Endpoints responds but returns incorrect results.To troubleshoot an Azure Machine Learning (AML) real-time endpoint that is responding but returning incorrect results: validating the scoring script and model version is the primary step.Priority Troubleshooting Steps 1. Validate the Scoring Script (score.py) The scoring script is the bridge between the raw request and your model. Check for these common issues: Preprocessing Mismatch: Ensure the data transformations (scaling, encoding) in init() or run() exactly match those used during training.Input Parsing: Verify that the JSON payload is being loaded correctly (e.g., json.loads(raw_data)).Exception Handling: Wrap your logic in try-except blocks and log the specific error message to see if a calculation is failing silently.2. Confirm the Model Version Intermittent "incorrect" results often happen during blue/green deployments or when multiple versions are active.Check Tags: Ensure the Model object loaded in your script is the specific version you intended.Traffic Allocation: If using an AML Online Endpoint, check the Deployment weights. You might be accidentally sending 20% of traffic to an old, "degraded" model version.Box 3: Adjusting autoscaling configuration Endpoints times out under increased traffic.When an Azure Machine Learning real-time (online) endpoint experience intermittent failures and timeouts under increased traffic, adjusting the autoscaling configuration is the primary way to ensure the system can handle the load.
https://learn.microsoft.com/en-us/azure/machine-learning/how-to-troubleshoot-online-endpoints https://medium.com/syntonize/deepseek-azure-ml-bridging-innovation-and-security-in-ai-f99d742aa74e
A team manages an Azure Machine Learning workspace and deploys a model to an endpoint.A deployed online endpoint shows inconsistent response times during periods of high traffic.You need to identify potential performance degradation.Which three metrics should you monitor? Each correct answer presents part of the solution. Choose three.Note: Each correct selection is worth one point.
Answer(s): B,C,E
To locate potential performance degradation in an Azure Machine Learning online endpoint during high traffic, you should monitor these three metrics: Requests per minute: This metric tracks the volume of incoming traffic and helps identify if spikes in load correlate with slower response times.Connections active: This monitors the total number of concurrent TCP connections from clients, which can indicate if the endpoint is reaching its capacity limits during peak periods.Request latency: This directly measures the time taken to respond to requests, allowing you to observe exactly when and by how much performance is degrading.
https://oneuptime.com/blog/post/2026-02-16-how-to-deploy-a-machine-learning-model-as-a-real-time-endpoint-in-azure-machine-learning/view
DRAG DROP (Drag and Drop is not supported)A team deploys a machine learning model to a managed online endpoint. The team monitors model performance and data quality metrics in production.When monitoring thresholds are exceeded, the team requires an automated operational response that notifies downstream systems.You need to configure the monitoring solution to meet the requirements.Which configuration should you associate with each requirement as a first step? To answer, move the appropriate configurations to the correct requirements. You may use each configuration once, more than once, or not at all. You may need to move the split bar between panes or scroll to view content.Note: Each correct selection is worth one point.Select and Place:
Box 1: Configure alert rules Automated responseTo set up an automated operational response for a deployed machine learning model, you should integrate Azure Machine Learning monitoring with Azure Monitor and Action Groups. This ensures that when performance or data drift thresholds are crossed, downstream systems are notified instantly via webhooks, emails, or automated workflows. Core Components for Automated Response Monitoring Signal: Tracks Data Drift, Prediction Drift, or Model Performance (Precision/Recall).*-> Alert Rule: A condition defined in Azure Monitor that triggers when a metric exceeds a threshold.Action Group: A collection of notification preferences and automated receivers (Webhooks, Azure Functions, Logic Apps).Box 2: Trigger retraining pipeline Infrastructure optimizationTriggering the Retraining Pipeline Once the threshold is exceeded, the system must optimize the infrastructure or model. REST API Call: Use the Logic App to call the Azure ML Job/Pipeline endpoint.SDK Trigger: Use an Azure Function with the azure-ai-ml library to submit a new command or pipeline job.Data Prep: Ensure the trigger passes the latest data location to the retraining job to address the detected drift.
https://medium.com/@ajayfrom2k/proactively-predict-azure-vm-failures-using-ai-powered-anomaly-detection-with-azure-monitor-df2598e1005c
HOTSPOT (Drag and Drop is not supported)You use Azure Machine Learning to train models across multiple experiments by using the same workspace.You must record training runs in a centralized location to compare results from different jobs.During training, performance values must be captured so they appear in the experiment run history.You need to configure experiment tracking.What should you configure for each requirement? To answer, select the appropriate options in the answer area.Note: Each correct selection is worth one point.Hot Area:
Box 1: MLflow tracking URI Ensure runs are recorded in the workspace experiment historyTo ensure training runs and performance values are recorded in the Azure Machine Learning workspace experiment history, you must configure and use an MLflow tracking URI or the Azure Machine Learning SDK logging functions.Key Configuration Requirements To capture and centralize your results, follow these steps: Set the Tracking URI: Configure your training script to point to the Azure ML Workspace as the MLflow tracking backend.Initialize an Experiment: Ensure your code specifies an Experiment name; if the name doesn't exist, Azure ML creates it automatically.Enable Logging: Use specific commands within your training script to "send" data to the workspace.Box 2: MLflow metrics logging Record performance values during model trainingTo record performance values during model training in Azure Machine Learning, you must configure Logging within your training script.You should use the MLflow library (specifically mlflow.log_metric) or the Azure ML SDK (run.log) to capture performance values. These values are then stored in the Experiment Run History for centralized comparison.
https://learn.microsoft.com/en-us/azure/machine-learning/how-to-train-mlflow-projects
A data science team trains a classification model that predicts loan approval outcomes.Before registering the model, the team must ensure the following:-Predictions must not disproportionately impact protected groups.-Prediction errors can be evaluated across different data segments.You need to assess whether the model meets Responsible AI expectations.Which two approaches should you use? Each correct answer presents part of the solution. Choose two.Note: Each correct selection is worth one point.
Answer(s): D,E
[D] To evaluate a trained loan classification model for Responsible AI expectations—ensuring no disproportionate impact on protected groups, evaluating error across segments, and verifying prediction transparency—you can employ SHAP (SHapley Additive exPlanations) values to assess feature importance.This approach allows you to identify which variables (e.g., credit history, debt levels) drive the model's predictions, fostering trust and fairness.Feature Importance for Transparency: Use SHAP (model-agnostic) or LIME (local approximations) to explain why the model approved or denied a loan. These techniques identify how each feature contributes to individual predictions.[E] To ensure a trained loan approval classification model meets responsible AI expectations—specifically, that it does not disproportionately impact protected groups and that errors can be evaluated across segments—you should analyze error rates across defined demographic cohorts using Fairness-Aware Machine Learning metrics.
https://urfpublishers.com/journal/artificial-intelligence/article/view/explainable-aiml-testing-ensuring-transparency-accountability-and-compliance https://timvero.com/blog/ethics-in-automated-lending-can-ai-make-fair-credit-decisions
A team deploys a model to a real-time endpoint in Azure Machine Learning. You deploy some updates to the endpoint.The endpoint returns errors after the new deployment is released.You need to restore the service as quickly as possible.What should you do first?
To restore the service as quickly as possible, you can roll back traffic to the previous deployment by updating the traffic allocation settings of your Azure Machine Learning online endpoint.Azure Machine Learning managed online endpoints support multiple deployments under a single endpoint, allowing for blue-green deployment strategies where you can shift traffic between versions instantly. Key Benefits of This Approach Instant Recovery: Traffic shifting is a routing change and does not require redeploying the previous model's code or environment, making it the fastest recovery method.No Downtime: Because the previous deployment remains "warm" (active but receiving no traffic), the switch happens without interrupting the service.Isolation for Debugging: You can keep the failing deployment at 0% traffic to inspect its logs using az ml online-deployment get-logs without affecting end usersHow to Roll Back Traffic If your new deployment (e.g., "green") is returning errors, you can reallocate 100% of the traffic back to the known stable deployment (e.g., "blue") using the following methods:* Azure CLI: Use the az ml online-endpoint update command to set the traffic percentage:az ml online-endpoint update --name <your-endpoint-name> --traffic "blue=100 green=0"* Azure Machine Learning Studio: Navigate to Endpoints in the left menu. Select your specific real-time endpoint. Go to the Details or Live Traffic tab. Adjust the traffic percentages so the previous deployment receives 100% and the failing deployment receives 0%. Select Update or Save to apply the changes immediately
https://learn.microsoft.com/en-us/answers/questions/1275110/azure-ml-v2-yaml-code-for-live-traffic-allocation
HOTSPOT (Drag and Drop is not supported)You train a model in Azure Machine Learning.You plan to capture experiment details for later comparison. The training code must log parameters and metrics for each run.You review the following training script.You need to verify whether the training script meets the experiment tracking requirement. For each of the following statements, select Yes if the statement is true. Otherwise, select No.Note: Each correct selection is worth one point.Hot Area:
Box 1: Yes Yes - The code ensures that training parameters and metrics are tracked for each run.To ensure parameters and metrics are tracked for each run, your code should follow this structure: Initialize a Run: Use mlflow.start_run() to define the beginning of your tracking session.Log Parameters: Use mlflow.log_param("name", value) for hyperparameters or configuration settings (e.g., learning rate, batch size).Log Metrics: Use mlflow.log_metric("name", value) for performance data (e.g., accuracy, loss).Box 2: Yes Yes - The code uses the experiment name to define the artifact path of the logged model.To ensure that your Azure Machine Learning training code captures experiment details and uses the experiment name to define the artifact path of the logged model, you must include the following code statement: mlflow.set_experiment("experiment-name")Why this is the correct statement: Defining the Artifact Path: In Azure ML, the experiment acts as the logical container for your runs. By setting the experiment name using MLflow, you explicitly define where the run artifacts (including the model) will be stored in the workspace.Artifact Organization: When you subsequently call a logging function like mlflow.sklearn.log_model(model, "model_artifact_path"), the "artifact path" is relative to the specific experiment run. Setting the experiment name ensures the model is not logged to the "Default" experiment, but rather to a path associated with your specific project name.Parameter & Metric Logging: This statement typically precedes mlflow.start_run(). Once the experiment and run are active, functions like mlflow.log_param() and mlflow.log_metric() will correctly attribute those values to that specific experiment.Box 3: Yes Yes - The code associates the run with a named experiment.To name the experiment to group runs together use mlflow.set_experiment("my-titanic-experiment")
https://learn.microsoft.com/en-us/azure/machine-learning/how-to-use-mlflow-cli-runs
Post your Comments and Discuss Microsoft AI-300 exam prep with other Community members:
Our website is free, but we have to fight against AI bots and content theft. We're sorry for the inconvenience caused by these security measures. You can access the rest of the AI-300 content, but please register or login to continue.