AI-102 EXAM QUESTIONS 2026 MICROSOFT AZURE AI 102 CERTIFICATION COURSE PART-1

This video focuses on practice questions for the AI 1002 Azure AI Associate exam, aiming to help beginners understand Azure AI concepts. The first question asks which computer vision client library is used to analyze image descriptions. The options provided include `visual_features_type.description`, `visual_features_type.tags`, `visual_features_type.categories`, and `visual_features_type.colors`. The correct answer is `visual_features_type.description`. The presenter explains that some questions are straightforward and can be answered using basic English comprehension. Beyond just answering the question, the video delves into how Azure AI works. For instance, to analyze an image description, you would import `ComputerVisionClient` from `azure.cognitiveservices.vision.computervision` and `VisualFeatureTypes`. You then provide your key and region, create a `ComputerVisionClient` instance, and use `client.analyze_image` with `VisualFeatureTypes.description` (or `.text` for text analysis) to perform the analysis. The results, such as extracted text, are then printed. The next question is similar, asking which computer vision client library is used to analyze image text. Following the previous logic, the answer is `visual_features_type.text`. The presenter encourages viewers to practice these questions. The third question asks how to read a file from a local file system. The options suggest C# syntax. The correct answer is `File.OpenRead()`, which opens an existing file for reading using `System.IO`. This highlights that the exam might include code snippets in C# or Python. The fourth question asks which Azure service is best suited for analyzing visual content in images and videos to categorize content. Options include Azure AI Vision, Azure Cosmos DB, Azure Cognitive Search, and Azure Function. Azure Cosmos DB is a database, Azure Cognitive Search is for searching, and Azure Function is a compute service. Azure AI Vision is specifically designed for visual content analysis, making it the correct choice. The fifth question asks which of the following best describes the prediction made by a machine learning model. The options are: 1. Probabilistic values based on correlations found in training data. 2. Randomly selected values with equal chance of selection. 3. Absolutely correct values based on conditional logic. The presenter explains that machine learning predictions are not random or absolutely correct. They are probabilistic values based on statistical relationships and correlations learned from training data. This is why predictions often come with a confidence score (e.g., 99% match or a score between 0 and 1), indicating the probability of the prediction being accurate. So, option 1 is the best description. The sixth question addresses enhancing security for Azure AI Services by protecting account keys. The options are: 1. Store keys in plain text files on the server. 2. Store keys in a public repository. 3. Store keys in environment variables. 4. Store keys in Azure Key Vault. Storing keys in plain text, public repositories, or environment variables is insecure. Azure Key Vault is the recommended cloud service for securely storing and accessing secrets like API keys, passwords, certificates, and cryptographic keys. The seventh question involves indexing a collection of text documents and searching them from a mobile application. It asks which service should be used to create the index. The options include Azure AI Service, Azure OpenAI Service, Azure AI Search, and Azure AI Index Service. The correct answer is Azure AI Search, which is specifically designed for indexing documents for search purposes, making search faster and more efficient. The eighth question asks how to use a machine learning model, trained with Azure Machine Learning, in an application. The options are: 1. Use Azure Machine Learning to publish the model as a web service. 2. Export the model as an Azure AI Service. 3. Build the application using the Azure Machine Learning designer. The best approach is to publish the trained model as a web service using Azure Machine Learning. This allows applications to consume the model's predictions via its REST interface. The ninth question concerns configuring diagnostic logging for Azure AI services. It asks which actions are necessary. The options are: 1. Define metrics to be collected. 2. Define diagnostic settings and configure log destination. 3. Set up data export schedule. The essential steps are to define diagnostic settings and configure the log destination. This involves enabling the settings, choosing the types of logs to collect (e.g., performance metrics, service-specific metrics), and selecting a destination like Azure Monitor logs, Azure Event Hubs, or Azure Storage. The final question of this session involves a bot built using the Microsoft Bot Framework SDK, started on a local computer, and the need to validate its functionality before connecting to it. The options are: 1. Run the Bot Framework Emulator. 2. Run the Bot Framework Composer. 3. Register the bot with the Azure Bot Service. 4. Run the Windows Terminal. To test and debug a bot locally, the Bot Framework Emulator is the correct and most powerful tool. The presenter emphasizes the importance of covering all aspects of the AI 1002 exam to achieve a good score and encourages viewers to provide feedback and subscribe for more content.