AI Bot
Azure Machine Learning Fundamentals Flashcards
Microsoft Azure AI Fundamentals AI-901 Flashcards
Study our Azure Machine Learning Fundamentals flashcards for the Microsoft Azure AI Fundamentals AI-901 exam with 33+ flashcards. View as flashcards, a searchable table, or as a fun matching game.
| Front | Back |
| Define dataset in Azure ML | A reference to data stored in a datastore or external source used for training and inference |
| How do you monitor deployed models | Use Azure Monitor and Application Insights to track health performance and latency |
| How do you register a model | Upload model files and metadata to the workspace model registry after a successful run |
| How do you secure an endpoint | Use authentication keys role based access network isolation and SSL |
| What are deployment targets in Azure ML | Options like Azure Container Instances and Azure Kubernetes Service for hosting models |
| What are pipeline steps | Modular tasks in a pipeline such as data preparation training evaluation and registration |
| What is a Compute Instance | A managed VM for development use with notebooks and IDEs |
| What is a datastore in Azure ML | A connection to storage such as Azure Blob or Azure Data Lake used to access data |
| What is a real time inference endpoint | A hosted HTTP endpoint that serves live predictions from a deployed model |
| What is a run in Azure ML | A single execution of a training job that records metrics artifacts parameters and logs |
| What is a scoring script | A Python file that loads the model and defines input output logic for inference |
| What is a training job | A submitted process that runs code on compute using a specified environment and dataset |
| What is AmlCompute | A managed scalable cluster for distributed training and large jobs |
| What is an Azure Machine Learning workspace? | A top level resource that organizes assets like experiments datasets models and compute |
| What is an environment in Azure ML | A reproducible package specification containing dependencies and Docker configuration |
| What is an experiment in Azure ML | A logical container for runs that tracks training jobs and metadata |
| What is attached compute | External compute resources such as Databricks on which you can run Azure ML jobs |
| What is Automated ML in Azure ML | A service that automates algorithm selection feature engineering and hyperparameter tuning |
| What is Azure Container Instances ACI used for | Quick low scale testing and development deployments |
| What is Azure Kubernetes Service AKS used for | Production scale low latency model deployments with autoscaling |
| What is Azure ML Pipelines | A way to orchestrate reusable steps for data preparation training and deployment |
| What is batch inference | Running predictions over large datasets asynchronously without a live service |
| What is CI CD for ML | Automating model build test and deployment using pipelines and repository integration |
| What is Concept Drift | A change in the relationship between input features and target that causes model degradation |
| What is Data Drift | A change over time in the input data distribution that can degrade model performance |
| What is feature store | A centralized repository for storing and serving feature values for training and inference |
| What is MLflow in Azure ML | An open source tracking and model format integrated with Azure ML for experiments and artifacts |
| What is model deployment roll back | The process of returning to a previous model version to recover from a bad release |
| What is model explainability in Azure ML | Tools and SDK features to surface feature importance and explanations for predictions |
| What is model versioning | Tracking multiple versions of a model so you can manage and deploy specific iterations |
| What is pipeline publishing | Creating a REST endpoint from a pipeline so it can be triggered programmatically |
| What is the model registry | A centralized store for versioned models with metadata and lineage |
| What problem types does Automated ML support | Classification regression and time series forecasting |
About the Flashcards
Prepare effectively with these comprehensive Flashcards for the Microsoft Azure AI Fundamentals exam. This deck helps students master essential cloud machine learning concepts and terminology. You can review core components like workspaces, datastores, and scalable compute resources. The collection also highlights crucial orchestration features, ensuring you understand how to build robust training pipelines and utilize automated machine learning tasks.
These flashcards also focus heavily on model lifecycle management and deployment operations. Students will test their knowledge of model registries, deployment targets, and inference endpoints. By reviewing complex topics like data drift, model explainability, and security configurations, learners can build the confidence needed to succeed on test day.
Topics covered in this flashcard deck:
- Workspaces and storage datastores
- Scalable compute instances and clusters
- Automated machine learning pipelines
- Model registries and versioning
- Inference endpoints and deployment targets
- Data drift and model explainability