Discover the Future: AI Foundations

What AI is, where it came from, the careers it creates, and how it changes business and society.

AI Fundamentals

• Define AI, machine learning, and deep learning
• Define different AI careers and job roles
• Define main subdomains of AI
• Define the specialized vocabulary of AI (e.g., agent, entity, POS tagging, AGI)
• Describe how AI can solve problems, including past, present, and future problems
• Describe lessons learned from the history of AI
• Describe the relationship between AI, machine learning, and computer science
• Describe Turing's test
• Explain the range of natural interactions used in AI application development
• Explain the fundamentals of AI
• Identify intelligent and non-intelligent examples of machine behavior
• Identify how tone and speaker intent impact natural language AI systems
• Classify rational agents according to their understanding of the environment

Reasoning

• Describe how a rational agent can deal with contingencies while planning
• Describe how multiple agents coordinate their behavior
• Describe how probabilistic reasoning works
• Explain how logic is used to build reasoning systems
• Explain the difference between propositional and first-order logic
• Understand the basics of fuzzy logic and its use in AI

Social and Business

• Demonstrate how AI is used as an economic driver to provide new services
• Use critical thinking to identify misinformation and deepfakes
• Explain how AI can improve app or website user experiences
• Explain how AI impacts communities and people in different ways
• Explain how AI impacts worker productivity
• Recognize signs of compromised information and data

AI Project Planning

• Describe AI problem identification
• Describe AI system design
• Describe different types of AI deployment models
• Explain the basic mechanism of a planning system
• List the factors that might affect the cost of developing and deploying ML models
• List the risks of preferring a more complex model over a simple one

How to collect, choose, clean, and prepare the data that AI learns from.

• Describe a simple model of the data processing cycle (input, processing, output)
• Describe data gathering to create new datasets
• Describe dataset selection techniques and methods
• Describe the importance of dataset curation
• Explain how to decide between data file formats such as XML, CSV, and JSON
• Explain the importance of feature engineering
• Explain various sampling plans, including subsampling
• Explain what data distribution shift is and its implications in production
• Find and filter out missing or N/A data
• Give an example of the use of multi-modal data in an AI application
• Identify univariate and multivariate outliers in a dataset
• Modify an existing script to clean data

The step-by-step methods AI uses to search, sort, predict, and generate text.

• Define the main algorithms of different machine learning methods
• Describe the classification, approximation, inference, optimization, recognition, and search families of reasoning algorithms
• Describe how generative, pretrained language models generate text
• Describe how parameters like temperature affect the output of large language models
• Distinguish deep learning from other learning algorithms
• Explain Maximum Likelihood Estimation (MLE)
• Explain search algorithms and operators commonly used in AI
• Explain the algorithm for fitting bivariate linear regression models
• Explain the difference between classification and regression
• Explain the effect of computational complexity on solving algorithms
• Explain k-means clustering starting values
• Explain the major differences between algorithms for supervised and unsupervised models
• Identify the differences between informed and uninformed search techniques

Using AI responsibly: privacy, bias, copyright, and telling reliable sources from unreliable ones.

• Describe privacy concerns related to AI
• Describe the role of ethics and philosophy in AI
• Tell the difference between credible and unreliable information sources
• Explain copyright issues from generative models trained on data scraped from websites
• Explain how bias in training data can affect a model's fairness
• Explain the ethical responsibility of AI designers and developers

How computers learn from examples, and how to tell whether a model is actually any good.

• Identify supervised, unsupervised, reinforcement, and transfer learning and the problems they solve
• Compare the complexity of a single decision tree with a random forest
• Describe how AI models make predictions or decisions
• Describe how unstructured observational data can be used to train an AI model
• Describe the issue of "black-box" ML models
• Describe the limitations of AI that supports natural interactions
• Evaluate a prediction model where the outcome is a continuous variable
• Explain the bias-variance trade-off
• Explain how ensemble methods work (e.g., bagging, boosting, random forests)
• Explain how to include a categorical variable in a prediction model
• Explain k-fold cross validation and its purpose
• Explain "agent" and "action" in reinforcement learning
• Explain underfitting and overfitting
• Explain ML vocabulary (e.g., training and testing data, labels, naive Bayes, one-hot encoding)
• Explain when to use logistic rather than linear regression

The math behind AI predictions and how confident we can be in them.

• Define a cost function used to train a neural network
• Distinguish Bayesian and frequentist approaches to probability
• Estimate the mean and standard error of the mean from data
• Explain the definition, purpose, and use of bootstrapping
• Explain null hypothesis significance testing
• Explain the curse of dimensionality
• Explain the sources of uncertainty in a prediction model
• Give examples of continuous, binary, categorical, and ordinal data types

Responsible AI, how generative models work, and the kinds of jobs AI can do.

Principles of responsible AI

• Fairness
• Reliability and safety
• Privacy and security
• Inclusiveness
• Transparency
• Accountability

AI model components and configurations

• Describe how generative AI models work
• Choose an appropriate AI model based on its capabilities
• Identify model deployment options and configuration parameters

AI workloads

• Identify common AI workloads: generative and agentic AI, text analysis, speech, computer vision, and information extraction
• Describe text analysis techniques: keyword extraction, entity detection, sentiment analysis, and summarization
• Identify features of speech recognition and speech synthesis
• Identify features of computer vision and image-generation models
• Identify ways to extract information from text, images, audio, and video

Building real AI apps and agents that work with text, speech, images, and documents.

Generative AI apps and agents

• Write effective system and user prompts
• Deploy a model and interact with it in the Foundry portal
• Create a lightweight chat app with the Foundry SDK
• Create and test a single-agent solution
• Create a lightweight client app for an agent

Text and speech

• Build an app that includes text analysis
• Respond to spoken prompts with a multimodal model
• Build an app with Azure Speech

Computer vision and image generation

• Interpret images in prompts with a multimodal model
• Create new images with generative models
• Build an app that includes vision capabilities

Information extraction

• Extract information from documents and forms with Azure Content Understanding
• Extract information from images
• Extract information from audio and video
• Build an app with information extraction capabilities