Azure AI Fundamentals and Machine learning

Learn how machine learning forms the foundation of Azure AI by training on data to make predictions, detect anomalies, and enable computer vision, natural language processing, and dialogue.

Learn about key risks and challenges in AI, including privacy, security, bias, and accountability; explore real-world examples like loan approvals, facial recognition, autonomous vehicles, and medical data protection.

Explore why machine learning underpins everyday tools like search and spam filtering, and how it powers clickstream data mining, electronic medical records into medical knowledge, and genome insights including covid-19.

Explore how machine learning learns from experience through supervised and unsupervised learning, including classification, regression, and clustering. See examples like spam detection, weather forecasting, and Netflix-style clustering.

Azure machine learning offers an end-to-end cloud platform for training, deploying, and predicting with models, using drag-and-drop workflows and Python integration to import data from multiple sources.

Create an Azure machine learning workspace, train and deploy models for bike rental prediction using automated machine learning and the designer, and manage datasets and notebooks.

Learn how to use automated machine learning to predict bike rental demand by selecting target variable, cleaning data, and choosing between regression models like random forest, with primary metric optimization.

Explore evaluating a machine learning model with the best model summary, residuals matrix, and predicted versus actual values, then deploy an endpoint and run notebook predictions.

In this lab, deploy the model, call the endpoint with inputs like weather and temperature, and predict bike rental demand over two years, then adjust scenarios using the designer.

Build a car price prediction pipeline with data cleaning, missing value handling, and column selection; train a linear regression model on a 70/30 split and evaluate results.

Review the regression model's predictions versus the actual prices on the testing data, interpret the mean absolute error and root mean squared error, and publish the evaluated model.

Explore how computer vision enables an AI to see and interpret images through techniques like image classification, object detection, semantic segmentation, and OCR using Azure cognitive services.

Join the lab to build computer vision skills by cloning the Microsoft Docs exercise, running a Python VM, and training a custom vision model to classify fruit (apple, banana, orange).

Publish and configure a Custom Vision project, obtain the project ID, key, and endpoint, then upload test images like an apple to test predictions.

During this lab, configure a custom vision endpoint, use the project ID and key, publish a model named grocery, and run predictions to classify fruit images.

Explore natural language processing with Azure AI, including text analytics, sentiment analysis, key phrase extraction, entity recognition and linking, translation, speech recognition and synthesis, and language understanding.

Analyze hotel reviews to demonstrate text analytics workflows, including language detection, key phrase extraction, sentiment analysis, and named entities, using Azure Cognitive Services.

Build a language understanding model with Azure AI to control home devices by defining light and fan entities, creating switch on/off intents, training, testing, and publishing with endpoint and key.

Learn how conversation ai enables chat and voice interfaces across channels, build a knowledge base with q&a, train natural language processing, and ensure transparent virtual assistant behavior with human handoffs.

Build a customer support bot by creating a knowledge base that interprets questions for a website chat, email, and voice interface, with Q&A, Azure deployment, and telemetry.

Create a bot by configuring a knowledge base with Q&A and chit-chat in English. Test, train, publish the knowledge base, and deploy via the endpoint, subscription, and service provider.

Test your website by running the bot, connect to channels like Slack or Telegram, copy the secret key, and embed the code to enable real-time chat responses.