The AI Study Guide: Azure Machine Learning Edition

NatalieM

Microsoft

Feb 21, 2024

Hi! It's me, Natalie, your Azure AI and ML learning guru. Want to learn something about Azure AI or Machine Learning? I've got something for that!

The AI Study Guide: Discover Machine Learning with these free Azure resources

Welcome to the February edition of the Azure AI Study Guide. Every month I’ll be spilling the tea on the best and newest tools for skilling up on Azure AI. This month we’re putting on our thinking caps to investigate Azure Machine Learning (ML). I’ll give you a quick breakdown of what it is, then we’ll explore a four-week roadmap of our top FREE resources for you to continue your AI learning journey! And as a bonus, stay tuned to the end to see what makes machine learning and generative AI a dynamic duo.

First things first: What is Machine Learning?

Short and sweet, Machine Learning is what happens when a computer learns from data without explicit programming. Algorithms help the computer improve its performance over time by analyzing data and identifying patterns, enabling it to make predictions, classifications, and decisions based on learned information.

Meet Robo

Imagine you have a pet robot named Robo. He’s cuddly and eager to please, but he didn’t come fully pre-programmed. So to make his owner—you—happy, he’s going to use machine learning to learn and adapt to his environment.

Isn't Robo cute?

Say you want Robo to recognize different fruits so he can bring you a snack upon request. He has sensors that collect data about his surroundings, so to help you show him an apple or a banana. Robo examines each, and the algorithms within his software analyze the data. These algorithms are like Robo’s "brain," capable of recognizing patterns and making connections. Now, when you show Robo a banana, he’ll say, “That looks like a banana!” Robo uses what he has learned from the seeing the different fruits to make decisions. It’s like Robo’s brain is figuring things out on its own! 

The learning process is continuous. As Robo interacts with you and the world, he gathers more data, feeds it back into the algorithms, and further refines his understanding. This creates a feedback loop that allows Robo to become more and more personalized and responsive over time. Eventually he’ll learn to not only distinguish produce, but also learn new tricks you teach him.

Go from Azure Machine Learning novice to expert in 4 weeks.

Now that we have a basic grasp on ML, let’s jump into the FREE 4-week ML roadmap I’ve built out for you below. This study guide is going to empower developers and data scientists alike to build, deploy, and manage high-quality models faster and with confidence. 

Week 1: Fundamentals 

Introduction to Machine Learning: Unleash the power of Machine Learning, zero coding required! This interactive journey dives into the heart of ML without the intimidating jargon.
Explore the Azure Machine Learning Workspace: Build the foundation for your Azure Machine Learning adventures and navigate the workspace like a pro with these powerful instruments.
Quick start tutorial: Get started with Azure Machine Learning: Master the core features of Azure Machine Learning with this hands-on tutorial. Build, train, and register your AI brainchild, then deploy your model to the cloud and make predictions in real-time.

Week 2: Get your hands dirty

Build Classical Machine Learning Models with Supervised Learning: Train models to see the future? Yes, please! Supervised learning lets you do just that. Feed it examples, and it learns to make predictions, constantly refining its knowledge based on feedback. Explore the individual components of the learning process, and exactly how this process can improve a model.
Tutorial: Upload, access and explore your data in Azure Machine Learning: Let’s dive into prototyping, the secret weapon of early ML exploration. Learn how to upload your data to cloud storage, create an Azure ML data asset, and more.

Video: Collaborate on machine learning assets across teams and workspaces with Azure ML registries: Collaborate like never before with Azure ML registries. Share, discover, and reuse valuable ML assets like models, pipelines, and environments across your entire organization.

Week 3: Approaching expert status

Deploy and consume models with Azure Machine Learning - Training | Microsoft Learn: You’ve trained your model—now, deploy it in your app! Enhance user experience with a service that enables real-time predictions for individual or small sets of data points.
Video series: Scaling your AI/ML practices with MLOps and Azure Machine Learning: Join Seth Juarez and an amazing lineup of guests for a three-part series on powering AI and ML with Azure.

Episode 1: Abe Omorogbe gives an overview of MLOps and how to utilize AzureML MLOps capabilities to streamline the process of moving ML experiments from training to inference.
Episode 2: Setu Chokshi introduces the Azure MLOps (v2) Solution Accelerator and its value proposition. He also shares some customer use-cases!
Episode 3: Scott Donohoo demos how to use the Azure MLOps Solution Accelerator to securely train, deploy and manage ML models in production environments.

Week 4: Start training and earn a badge!  

Train and evaluate deep learning models: Strap in for some cutting-edge ML/AI. “Deep learning” is an advanced form of ML that tries to emulate the human brain, using artificial neural networks that process numeric inputs rather than electrochemical stimuli.
Train a model and debug it with Responsible AI dashboard: Machine learning models should not only excel in accuracy but also adhere to ethical principles. In this module you’ll learn how to create a responsible AI dashboard, mitigate biases in your data, and meet compliance regulation requirements.
Assessment: Train and deploy a machine learning model with Azure Machine Learning: OK, hotshot, time to show off! Demonstrate your ability to train and deploy ML models to earn a Microsoft Applied Skills credential (and a cool badge).
Microsoft Certified: Azure Data Scientist Associate: This exam measures your ability to accomplish the following technical tasks: design and prepare a ML solution; explore data and train models; prepare a model for deployment; and deploy and retrain a model. (Note: Price for certification exam is based on the country or region in which the exam is proctored.)

Extra goodies

Gen AI + ML = BFFs 

Generative AI and Azure Machine Learning are powerhouses individually, but together they create a synergy that unlocks even greater potential. For example, a common issue with training ML models is a lack of sufficient data. But generative AI can create synthetic data, allowing the ML model to learn from broader and more diverse datasets, improving its effectiveness and generalizability.

To learn more about this dynamic duo, check out this series of videos produced by Azure experts:

A Beginners Guide to AI: Azure Machine Learning: Passionate about building bespoke AI solutions? Join us for a comprehensive introduction to Azure's Machine Learning service, guiding you on when and why to train your unique models. Elevate your AI prowess and start tailoring solutions that resonate with your vision.
Build and maintain your company Copilot with Azure ML and GPT-4 : Large AI models are transforming the way we live and work. Harnessing these technologies for real-world applications requires purpose-built tooling to enable effective prompt engineering, experimentation, and safety mechanisms that deliver great customer experiences.
Prompt flow: An end-to-end tool to streamline prompt engineering: By streamlining the development, evaluation, and continuous integration and deployment of prompt engineering projects, prompt flow empowers data scientists and LLM application developers with an experience that combines natural language prompts, templating language, a list of built-in tools and Python code.
Streamline Your AI Application Development with Prompt Flow in Azure Machine Learning: Prompt flow also provides a streamlined experience to quickly create prompt workflows that connect large language models to your organizations data to create intelligent applications.
Deploy and fine-tune large AI models with your data with Foundation Models in Azure Machine Learning: Learn how to provide native capabilities to fine-tune and deploy foundation models from multiple open-source repositories using Azure Machine Learning components and pipelines.