Azure AI Fundamentals: AI-900 Concepts & Practice Tests

Explore how artificial intelligence simulates human intelligence, enables data-driven decisions, processes visual input, and uses NLP and speech recognition, plus four AI types: reactive, memory, theory of mind, self-awareness.

Clarifies the differences among data science, machine learning, and artificial intelligence, detailing data modeling and warehousing, supervised/unsupervised/semi-supervised/reinforcement learning, and how artificial intelligence uses data-driven learning and natural language processing.

Explore common artificial intelligence workloads, including machine learning, anomaly detection, computer vision, natural language processing, and conversational AI.

Azure artificial intelligence as a service provides scalable cloud tools: Azure machine learning for training and deploying models, cognitive services for image recognition, and Azure bot service for building bots.

Examine challenges and risks of artificial intelligence, including data bias in loan processes, data quality issues, potential harms from errors, and questions of trust and liability.

Explore six basic principles of responsible artificial intelligence, including fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability, and consider who is responsible for these decisions.

Explore how Azure Machine Learning accelerates the end-to-end lifecycle, enabling time to market for building, training, and deploying models in a cloud platform with responsible machine learning, fairness, and accountability.

Explore Azure cognitive services to add ai capabilities to apps with simple api calls. Discover language, vision, speech, and decision services, plus Azure Bot Service for multilingual conversations.

Explore Azure cognitive search, a cloud service with AI capabilities for vision, language, and speech, plus semantic search that ranks results by user intent, extracting, indexing, and enriching data.

Explore fundamental principles of machine learning on Azure and how computers create predictive models from data. Learn to train models that identify patterns and make decisions with minimal human intervention.

Learn how regression, a supervised machine learning technique, predicts numeric labels from item features, training on data to estimate selling price using engine size, seats, and temperature.

Explore classification as a supervised machine learning technique, using training and validation data to learn the feature and label mappings and predict diabetes risk for new patients.

Explore clustering, an unsupervised machine learning technique that groups items into clusters based on features, with examples like measuring flowers and no predefined labels.

Learn Azure machine learning, a cloud-based platform with a centralized workspace to manage data, models, and experiments, then deploy registered models to container instances, Kubernetes, or REST endpoints.

Explore automated machine learning that selects the best model from your data and model type, and use Azure Machine Learning Designer to train, evaluate, deploy inference pipelines, and manage resources.

Explore how Azure cognitive services enable computer vision to extract text from images and videos, label objects, describe visuals, and analyze video in real time within apps.

Discover computer vision applications such as image classification, object detection, semantic segmentation, and image-to-text description. Explore face detection and recognition, OCR, and health care uses like cancer and COVID-19 diagnosis.

Watch a live demo of Azure computer vision: analyze and describe images with tags and confidence scores, moderate adult content, read text and handwriting, and recognize celebrities and landmarks.

Azure cognitive services bring AI within reach for developers, enabling apps to see, hear, speak, search, and understand via API calls and authentication keys, with resources organized by subscription.

Discover Azure's computer vision service for image and real-time video analysis, including description, tag generation, object and face detection, OCR, content moderation, and edge in containers via APIs.

Train models with the custom vision service for image classification and object detection, then analyze faces with the face service, extracting facial attributes, aging, emotions, similarity, and identity verification.

Learn to extract text with the computer vision service and analyze videos with video indexer, including OCR, real-time transcription, facial recognition, topics, sentiment, and content moderation.

Create and use predefined models for custom insights to recognize language, celebrities, brands, and animated characters, and train facial recognition from sample images using OCR APIs.

Post videos to the video indexer rest api to retrieve indexing results and share insights on web pages; auto-generated video names and descriptions, with English as the source language.

Analyze forms with the form recognizer service to extract data via OCR from scanned images or PDFs, using pre-trained or custom models for semantic recognition of form fields on Azure.

Explore natural language processing (NLP) as an AI field that lets machines read, understand, and derive meaning from human language, with Azure services for text analysis, sentiment analysis, and translation.

Discover Azure cognitive services for natural language processing, including text analytics (language detection, key phrase extraction, entity detection, sentiment analysis), speech services, and entropy-driven tasks like topic detection.

Explore the Azure text analytics service, a cloud natural language processing API offering language detection, key phrase extraction, sentiment analysis, named entity recognition, and entity linking.

Explore text analytics via rest API or client libraries to extract language, sentiment, key phrases, and entities, and learn speech recognition and text to speech synthesis.

Explore the translator service, a powerful Azure AI capability that translates text instantly across 90+ languages with automatic language detection, one-to-many translations, transliteration, and use in call centers or apps.

Learn how to detect language, translate from Japanese to English, and transliterate Japanese to Latin script using Microsoft Translator APIs, returning json results across more than 90 languages.

Explore natural language understanding and how language models interpret user input to determine intent and trigger actions in apps and smart devices such as Cortana, Alexa, Siri, and Google Home.

Learn to configure two Azure resources for language understanding: an authoring resource to train a language model and a prediction resource to host and process requests.

Map natural language utterances to intents to train a language understanding model, then extract entities to trigger app actions like get time or get weather.

Identify and map entities to user intent by extracting time, location, and context. Explore list, regex, prebuilt, pattern, and machine learning entities for robust language understanding.

Explore how patterns and patterns.any() disambiguate intents and entities in language understanding using template utterances like turn on device and identifying the device name.

Explore Azure speech services for speech to text across 85 languages with real-time transcription and custom models, plus translator for live and document translation via a JavaScript SDK demo.

Discover how conversational AI enables human-like dialogue across text and voice channels, using speech recognition, intent understanding, and personalization to deliver natural, bot-powered interactions.

Explore the nine gold guidelines for responsible boards, including transparency about capabilities, clear user communication, human handoff, cultural respect, reliability, privacy, secure data handling, accessibility, and accountability.

Learn how to use the QnA Maker service to rapidly create a conversational layer over data, extract questions and answers from semi-structured content, and publish a bot with no code.

Compare QnA Maker and language understanding services, showing how QnA Maker returns static answers from a knowledge base, while language understanding interprets utterances to detect intents and entities for actions.

Active learning suggests alternative question phrases when scores are similar, with Q&A maker presenting options for review. The client app submits feedback via the train API to improve the bot.

Create a Q&A bot by integrating a knowledge base with the Azure Bot Service, distinguishing Q&A Maker and knowledge base features, and enabling conversations across websites, email, and voice assistants.

Prepare for AI-900 with example questions on building a multi-channel conversational app using Azure Bot Service, Bot Framework, and active learning to refine Q&A Maker responses.

Identify conversational AI workloads and when to apply automated customer interactions. Build self-service bots and knowledge bases with Q&A maker and REST APIs to update content.

Learn to build a knowledge-base chat bot that answers simple questions with predefined answers using Q&A Maker and Azure Bot Service, with text analytics and translator services.

Map text analytics api features to natural language processing tasks, including sentiment analysis, key phrase extraction, and entity recognition for dates, with language detection and translation.

Be prepared for AI-900 questions on form recognizer, custom vision, and computer vision, including when to use OCR, object detection, and classification in practical scenarios.

Be prepared to answer AI-900 questions by mastering train-test split, classification vs regression, supervised and unsupervised learning, automated machine learning, and deploying real-time endpoints with Azure Machine Learning Designer.

Practice mapping machine learning tasks to real scenarios, recognize model evaluation outputs like confusion matrices, and apply feature engineering and feature selection within Azure ML and responsible AI principles.