AI-102 EXAM QUESTIONS 2024 MICROSOFT AZURE AI 102 CERTIFICATION COURSE PART-12

The video provides a series of questions and answers related to Azure AI services, aimed at preparing individuals for the AI-102 certification exam. The host begins by congratulating viewers who have successfully passed the exam using their playlist, encouraging others to share their success stories. The first question addresses a chatbot scenario using OpenAI, where company data needs to be uploaded via chat playground to ensure the chatbot uses this data for user questions. The task is to complete a blank in the provided code. The correct option is "chat completion option." This class is used to configure chat completion, including setting up messages and integrating additional extensions like chat extensions. Other options like "completion option" are for text completion, not chat, and "streaming chat completion" is for streaming responses, not basic configuration and extension integration. The second question continues the previous scenario, asking about blank two in the code, which involves integrating a new chat extension and defining search endpoint, search key, and index name. The correct answer is "Azure cognitive search Chat extension configuration." This option is specifically designed to configure Azure Cognitive Search as a chat extension, enabling the chatbot to use uploaded data. "Azure chat extension option" is too general, and "Azure chat extension configuration" lacks the specificity for cognitive service integration. "Azure key credential" is for providing credentials, not configuring the extension itself. The third question focuses on using the content moderator service (now Azure AI Content Safety) to identify sexually explicit language in messages. The user needs to identify which section in the response will contain the category score and which category will be assigned. The correct section is "classification," as it includes category scores for different content types and identifies offensive or explicit content. The correct category is "category one," which represents content likely to be sexually explicit. "PII" is for personally identifiable information, and "terms" identifies specific keywords but doesn't provide sexually explicit category scores. Category two is for potentially offensive language not necessarily sexual, and category three is for threatening or violent language. The fourth question involves building an app to generate tags for uploaded images in a user's preferred language (English, French, Spanish) while minimizing development effort. The recommended Azure service endpoint is "computer vision image analysis." This service offers robust image tagging features, supports multiple languages, and is a pre-built service, thus minimizing development effort. "Content moderator" is for moderation, not tagging. "Custom Vision image classification" requires more development effort. "Custom translator" translates text but doesn't tag images. The fifth question is about an app sharing user images, requiring scanning, text extraction, analysis for profane language, and minimizing development effort. The specific part of the question asks what to use for "text extraction." The optimal choice is "Azure AI document intelligence." While "Azure AI computer vision" can perform OCR, Document Intelligence is specifically designed to extract text from scanned documents, providing structured information and handling various document formats, making it a better fit for scanned images than general image processing. "Azure AI language" is for natural language understanding, not text extraction from images. "Content moderator" is for profanity detection, not extraction. "Azure AI custom Vision" requires more development effort. The sixth question follows up on the previous scenario, now asking specifically for "profane language detection." The answer is "content moderator" (Azure AI Content Safety), which is designed to detect and filter offensive content in text and images. The seventh question involves developing an app using both decision and language APIs, requiring a single endpoint and credential for accessing both services. The correct resource type to provision is "Azure AI Services." This service combines multiple AI services, including decision-making and language, under one umbrella, allowing unified access via a single endpoint and credential. This is crucial when multiple cognitive services need to be accessed efficiently. The eighth question describes an internet-based training solution that monitors a user's video stream to verify they are alone and not collaborating, while minimizing development effort. The recommended solution is "spatial analysis in Azure AI vision." This feature analyzes video streams to understand human behavior and the number of people in a defined space, making it suitable for verifying user presence and ensuring they are alone, with minimal custom development. "Speech to text in Azure AI speech service" is for verbal communication, not visual monitoring. "Object detection in Azure AI custom Vision" requires more development effort. The ninth question is about a language learning solution that analyzes teacher-submitted lesson plans to extract key fields like lesson times and required text, with minimal development effort. The recommended service is "Azure AI document intelligence." This service is ideal for analyzing documents and extracting structured information from them. The tenth question, also for a language learning solution, requires analyzing learning content and providing students with pictures representing commonly used words or phrases in the text. The recommended service is "immersive reader." This service is specifically designed to enhance text readability and comprehension, including showing pictures for common words, thus minimizing development effort. If the requirement was for every word, "custom Vision" might be considered, but for commonly used words, Immersive Reader is more appropriate.