Fundamentals of AI: Key Concepts and Terminologies

This section aligns with the following exam objectives:

Domain 1: Fundamentals of AI and Machine Learning
Task Statement 1.1: Explain Basic AI Concepts and Terminology

1. Core AI Concepts and Definitions

Artificial Intelligence (AI) refers to the capability of machines to simulate human intelligence, enabling them to reason, learn, and make decisions in order to solve problems or perform tasks autonomously.

Machine Learning (ML) is a subset of AI that focuses on enabling systems to learn from data and improve performance over time without being explicitly programmed.

Deep Learning (DL) is a specialized subset of machine learning that leverages artificial neural networks with multiple hidden layers to process large volumes of data and solve complex problems.

Neural Networks (NNs) are computational models inspired by the human brain. They consist of interconnected layers—input, hidden, and output layers—that work together to extract patterns from data.

Computer Vision is a branch of AI that enables machines to analyze and interpret visual information, such as images and videos, supporting tasks like image classification and object detection.

Natural Language Processing (NLP) focuses on enabling machines to understand, interpret, and generate human language, powering applications such as chatbots, translation systems, and sentiment analysis.

Model refers to a trained mathematical representation that uses learned patterns to make predictions or classifications based on input data.

Algorithm is a defined set of rules or procedures used during training to optimize a model so it can learn patterns effectively.

Training is the process of supplying data to an ML model so it can learn relationships and adjust its internal parameters.

Inference is the phase where a trained model is used to make predictions on new, unseen data.

Bias represents systematic errors introduced into a model due to imbalanced or skewed training data, potentially leading to unfair or inaccurate outcomes.

Fairness refers to efforts made to detect, reduce, and mitigate bias in AI systems to ensure equitable results across different groups.

Model Fit describes how well a model generalizes to unseen data.

• Overfitting occurs when a model memorizes noise in the training data and performs poorly on new inputs.
• Underfitting occurs when a model is too simple to capture meaningful patterns in the data.

Large Language Models (LLMs) are advanced NLP systems trained on massive text datasets, capable of generating context-aware, human-like language outputs (for example, GPT-based models).

📌 Exam Tip: Expect multiple-choice questions that test your ability to distinguish closely related terms such as AI, ML, and Deep Learning.

2. AI vs. Machine Learning vs. Deep Learning

AI is the broadest category and includes any system designed to emulate human intelligence, including rule-based systems and expert systems. Machine Learning is a subset of AI that relies on data-driven learning techniques rather than predefined rules. Deep Learning further extends ML by using deep neural networks, making it especially effective for tasks such as image recognition and natural language processing.

AI systems may function with minimal data and simple logic, while ML typically requires labeled datasets and statistical models. Deep learning systems demand very large datasets and significant computational power, often relying on GPUs or TPUs.

📌 Exam Tip: Scenario-based questions frequently ask you to identify the most appropriate approach—AI, ML, or DL—based on problem complexity and data availability.

3. Inference Approaches

Batch Inference involves generating predictions for large volumes of data at scheduled intervals, such as processing daily or weekly transaction records.

Real-Time Inference produces immediate predictions as new data arrives, enabling use cases such as fraud detection or recommendation systems.

📌 Exam Tip: Be prepared to determine whether batch or real-time inference is more appropriate for a given operational scenario.

4. Data Types Used in AI Systems

AI models work with multiple forms of data depending on the problem domain. Labeled data includes known outcomes and is essential for supervised learning, while unlabeled data is used in unsupervised learning scenarios.

Tabular data is structured into rows and columns, commonly found in databases and spreadsheets. Time-series data consists of sequential observations collected over time, such as sensor readings or financial prices.

Image data supports computer vision tasks, while text data is unstructured and commonly used in NLP applications. Structured data follows a predefined schema, whereas unstructured data lacks a fixed format and includes text, images, audio, and video.

📌 Exam Tip: Expect questions that assess your ability to select the correct data type for a specific AI or ML use case.

5. Machine Learning Learning Paradigms

Supervised learning trains models using labeled input-output pairs and is commonly applied to classification and regression problems such as spam filtering and fraud detection.

Unsupervised learning identifies patterns in unlabeled data and is often used for clustering, anomaly detection, and customer segmentation.

Reinforcement learning (RL) enables models to learn through interaction with an environment by receiving rewards or penalties based on actions taken. This paradigm is widely used in robotics, game AI, and autonomous systems.

📌 Exam Tip: Be able to match real-world scenarios with the appropriate learning paradigm and algorithm type.

6. Key Exam Takeaways

A strong understanding of foundational terminology is critical for success. Focus on clearly differentiating AI, ML, and Deep Learning, and understand how data types influence model selection. AWS exams emphasize practical application, so expect scenario-based questions that test your ability to choose the right learning approach. While this section is concept-focused, basic familiarity with AWS AI services such as Amazon SageMaker, Rekognition, and Comprehend will help contextualize exam scenarios.