Skip to main content

Section Introduction: Fundamentals of AI, Machine Learning & Deep Learning

Key Takeaways​

This section transitions from service-specific implementations (Amazon Bedrock, Amazon Q) to the conceptual foundations underpinning Domain 1: Fundamentals of AI and ML (20% of scored content) in the AWS Certified AI Practitioner (AIF-C01) exam.

The primary objective is mastering the hierarchical nesting of technologies—from broad Artificial Intelligence down to Generative AI—and understanding when classical rules, machine learning algorithms, deep learning neural networks, or generative models should be applied to solve enterprise business problems.

Main Discussion​

The Nested Hierarchy of Artificial Intelligence​

The four tiers are concentric subsets, not isolated parallel disciplines:

TierCore ConceptDistinguishing CharacteristicRepresentative Business Example
Artificial Intelligence (AI)Any software system or machine capable of performing tasks that typically require human cognitive intelligence.Encompasses symbolic reasoning, deterministic rule engines, expert systems, heuristics, and learning algorithms.Deterministic if-then tax calculation engine, automated chess engine.
Machine Learning (ML)A subset of AI where systems extract statistical patterns from historical data to make predictions rather than being explicitly coded.Algorithms adjust mathematical weights from datasets; performance improves automatically with more training observations.Credit score prediction, customer churn modeling, tabular fraud risk scoring.
Deep Learning (DL)A specialized subset of ML based on multi-layered Artificial Neural Networks (ANNs) inspired by biological neural structures.Automatically performs feature extraction from massive, complex unstructured data (images, audio, video, raw text).Facial recognition, autonomous vehicle computer vision, speech transcription.
Generative AI (GenAI)A subset of Deep Learning powered by large Foundation Models (FMs) and Transformer architectures.Focuses on creating novel synthetic content (text, synthetic data, images, code) rather than merely classifying or predicting labels.Conversational AI assistants, document summarization, code generation.

Traditional Programming vs. Machine Learning Paradigms​

Classical computing relies on deterministic logic, while machine learning shifts the programming paradigm:

  • Traditional Programming: A software engineer writes deterministic rules and conditionals (if/else) that evaluate input data to produce outputs.
  • Machine Learning: An algorithm ingests historical input data alongside ground-truth target outputs to derive the underlying mathematical relationship (the Trained Model), which is subsequently used to evaluate unseen inputs.

Upcoming Section Roadmap​

This section covers the following core AIF-C01 syllabus areas:

Exam Guide​

Exam Tips​

  • Hierarchy Recognition Questions: Exam questions often ask you to categorize a business requirement into its exact conceptual level. Remember:
    • AI is the outermost umbrella.
    • ML is learning patterns from data without explicit rule programming.
    • DL is learning from complex/unstructured data using multi-layered neural networks.
    • Generative AI is generating novel outputs using foundation models.
  • Scored Weighting: Domain 1 constitutes 20% of the scored questions on the exam. The exam focuses on conceptual mapping and service selection, not writing raw Python, PyTorch, or calculus equations.
  • Determinism vs. Probabilities: Classical rule-based systems are deterministic (fixed rule outputs), whereas ML and DL systems are probabilistic (outputting predictions with associated confidence scores).

Practice Test​

Question 1​

An enterprise architect is defining the organization's technology strategy and needs to explain the relationship between Deep Learning, Artificial Intelligence, Generative AI, and Machine Learning. Which of the following correctly describes their relationship from broadest domain to most specific subset?

  • A. Machine Learning →\rightarrow Artificial Intelligence →\rightarrow Deep Learning →\rightarrow Generative AI
  • B. Artificial Intelligence →\rightarrow Machine Learning →\rightarrow Deep Learning →\rightarrow Generative AI
  • C. Deep Learning →\rightarrow Machine Learning →\rightarrow Generative AI →\rightarrow Artificial Intelligence
  • D. Generative AI →\rightarrow Deep Learning →\rightarrow Machine Learning →\rightarrow Artificial Intelligence
Correct Answer
  • B. Artificial Intelligence →\rightarrow Machine Learning →\rightarrow Deep Learning →\rightarrow Generative AI
    • Explanation: Artificial Intelligence is the overarching field. Machine Learning is a subset of AI. Deep Learning is a subset of ML using deep neural networks. Generative AI is a specialized subset of Deep Learning focused on creating new content.

Question 2​

A retail company wants to predict whether a customer will cancel their subscription in the next 30 days based on their last 2 years of browsing and purchasing records. Rather than writing thousands of complex manual if/else business rules, the engineering team wants a system that identifies patterns automatically from historical customer records. Which paradigm best describes this approach?

  • A. Deterministic Rule-Based Programming
  • B. Machine Learning
  • C. Edge Computing
  • D. Static Scripting
Correct Answer
  • B. Machine Learning
    • Explanation: Machine Learning systems train algorithms on historical data and outcomes to learn statistical patterns automatically, replacing manual deterministic rule-writing for complex predictive problems.