Basic AI Concepts and Terminologies

Welcome. This lesson introduces core artificial intelligence (AI) concepts and terminology to help you prepare for the AWS Certified AI Practitioner exam. We cover why AI matters to organizations, common application areas, differences between rule-based and ML systems, and the relationships between AI, machine learning (ML), and deep learning (DL). This course emphasizes practical business drivers for AI adoption: increasing efficiency, lowering operational costs, and accelerating data-driven decision-making. AI systems act as force multipliers—automating complex tasks and interactions such as customer support, fraud detection, and data analysis. AI adoption improves customer experience while reducing human workload. As demand for AI skills grows, businesses use AI to gain deeper insights and make faster, more accurate decisions. Common real-world examples include AI-powered chatbots, automated code generation, and analytics pipelines that surface actionable trends from large datasets. AWS provides a broad set of services to support ML and AI deployments. For example, Amazon SageMaker is a managed platform for building, training, and deploying models at scale, while Amazon Rekognition offers image- and video-analysis capabilities such as face and object detection.

​

Rule-based systems vs. machine learning

Rule-based systems follow explicit, hand-crafted logic—for example, “approve loans when credit score > 750.” They are deterministic, simple to implement, and cost-effective when inputs and outputs are consistent and requirements are stable. Machine learning systems, by contrast, are probabilistic. They learn patterns from data and can generalize to unseen inputs, making them better suited to prediction, recommendation, and forecasting tasks where explicit rules would become complex or brittle. ML systems can be retrained with new data to adapt to changing conditions. Table: Rule-based vs ML comparison

​

What is Artificial Intelligence?

Artificial intelligence is the broad field of computer science focused on creating systems that perform tasks typically requiring human intelligence. The scope of AI ranges from simple automation to advanced capabilities such as perception, reasoning, and natural-language processing. Common AI capabilities include:

• Visual perception: image and video analysis
• Speech recognition and transcription
• Decision-making and planning
• Natural-language understanding and translation

​

Narrow AI vs General AI

• Narrow AI (specialized): Systems built for a single task or domain—e.g., voice assistants, recommendation engines, or image classifiers.
• General AI (broad): A theoretical system able to perform any intellectual task a human can. Despite advances in generative AI, true general AI remains a research goal and is not yet realized.

Practical examples: AI can help radiologists read X-rays faster or personalize retail product recommendations based on previous customer behavior.

​

Machine learning (ML) fundamentals

Machine learning is a subset of AI that uses algorithms, mathematical models, and statistics to learn patterns from data—structured and unstructured, labeled and unlabeled. After training, an ML model applies learned patterns to new inputs. Key concepts:

• Training data: examples used to teach the model (often labeled).
• Validation/test sets: used to evaluate generalization and tune hyperparameters.
• Generalization: the model’s ability to perform well on unseen data.

Examples:

• Spam filters learn features of spam emails from labeled examples.
• Chatbots improve responses by learning from conversation logs and user feedback.

​

Deep learning and neural networks

Deep learning is a subset of ML that uses multi-layered neural networks to learn hierarchical representations. Layers of artificial neurons transform input features through weighted connections and nonlinear activations. Deep networks automatically discover increasing levels of abstraction—useful for image recognition, speech recognition, and language modeling. Conceptual example: Given an image of a cat, early neural layers detect edges and textures, mid-layers combine these into shapes like ears and eyes, and later layers integrate shapes to predict “cat” with a confidence score.

​

AI, ML, and Deep Learning — concise comparison

Thanks for reading this lesson.