[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"lesson-aws-certified-ai-practitioner-ai-ml-fundamentals-ai-ml-core-concepts-en":3,"cheat-sheet---en":3,"domain-info---en":3,"prev-aws-certified-ai-practitioner-ai-ml-fundamentals-ai-ml-core-concepts-en":3,"next-aws-certified-ai-practitioner-ai-ml-fundamentals-ai-ml-core-concepts-en":3,"topic-info-aws-certified-ai-practitioner-ai-ml-fundamentals-ai-ml-core-concepts-en":4,"course-lesson-metadata-aws-certified-ai-practitioner-en":56},null,{"meta":5,"body":8},{"title":6,"description":7},"AI and ML Core Concepts","What AI and ML actually are, how deep learning and generative AI relate, the data types models consume, and how models learn and serve predictions.",{"type":9,"value":10,"toc":49},"minimark",[11,15,20,42,46],[12,13,14],"p",{},"This topic builds the vocabulary the rest of the AWS Certified AI Practitioner exam depends on. Before you can compare services, reason about prompts, or judge whether a model is fair, you need a clear picture of what AI, machine learning, deep learning, and generative AI are, how they nest inside each other, and how a model goes from raw data to a live prediction. Every term here appears again in later domains, so getting them right early turns much of the exam into reasoning instead of guessing.",[16,17,19],"h2",{"id":18},"what-this-topic-covers","What This Topic Covers",[21,22,23,27,30,33,36,39],"ul",{},[24,25,26],"li",{},"The difference between AI, machine learning, and deep learning, and the core vocabulary of model, algorithm, training, and inference",[24,28,29],{},"What neural networks are, what makes deep learning \"deep,\" and how it compares to traditional ML",[24,31,32],{},"Generative AI, foundation models, LLMs, and agentic AI, and how each differs from predictive ML",[24,34,35],{},"The two independent properties of data (structure and labeling) and common formats like tabular, time-series, image, and text",[24,37,38],{},"The three ways models learn: supervised, unsupervised, and reinforcement learning",[24,40,41],{},"The four ways a trained model serves predictions: real-time, batch, asynchronous, and serverless inference",[16,43,45],{"id":44},"why-it-matters","Why It Matters",[12,47,48],{},"These concepts map almost one-to-one onto Task Statement 1.1 of the exam, which asks you to define AI terms, compare AI, ML, deep learning, generative AI, and agentic AI, and describe the types of data, learning, and inferencing. The exam tests the boundaries between these ideas far more than the definitions themselves: whether you can tell generative from agentic AI, labeled from unlabeled data, or real-time from batch inference in a described scenario. Master the distinctions here and the rest of the course, from foundation model applications to security and governance, rests on solid ground.",{"title":50,"searchDepth":51,"depth":51,"links":52},"",3,[53,55],{"id":18,"depth":54,"text":19},2,{"id":44,"depth":54,"text":45},{"lessons":57,"cheatSheets":236,"mockTestsAvailable":66},[58,63,68,72,76,80,84,88,92,96,100,104,108,112,116,120,124,128,132,136,140,144,148,152,156,160,164,168,172,176,180,184,188,192,196,200,204,208,212,216,220,224,228,232],{"path":59,"title":60,"isFree":61,"description":62},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/01-ai-ml-core-concepts/01-what-is-ai-and-ml","What Is AI and ML?",true,"Define artificial intelligence and machine learning, see how they differ from ordinary programming, and learn how AI, ML, deep learning, and generative AI nest inside each other.",{"path":64,"title":65,"isFree":66,"description":67},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/01-ai-ml-core-concepts/02-neural-networks-and-deep-learning","Neural Networks and Deep Learning",false,"Understand what a neural network is, what makes deep learning 'deep,' and how deep learning differs from traditional machine learning on feature engineering, data needs, and interpretability.",{"path":69,"title":70,"isFree":66,"description":71},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/01-ai-ml-core-concepts/03-generative-and-agentic-ai","Generative and Agentic AI","See how generative AI creates new content with foundation models and LLMs, how agentic AI extends it to plan and act, and how both differ from traditional predictive machine learning.",{"path":73,"title":74,"isFree":66,"description":75},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/01-ai-ml-core-concepts/04-types-of-data-in-ai","Types of Data in AI","Learn the two independent properties of AI data, structure and labeling, and how tabular, time-series, image, and text formats map to real machine learning problems.",{"path":77,"title":78,"isFree":66,"description":79},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/01-ai-ml-core-concepts/05-supervised-unsupervised-reinforcement","Supervised, Unsupervised, and Reinforcement Learning","The three ways a model can learn: supervised learning from labeled examples, unsupervised learning from unlabeled data, and reinforcement learning from reward and penalty.",{"path":81,"title":82,"isFree":66,"description":83},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/01-ai-ml-core-concepts/06-types-of-inferencing","Types of Inferencing","Once a model is trained, inference is how it serves predictions. Compare real-time, batch, asynchronous, and serverless inference and match each to latency, payload, and traffic needs.",{"path":85,"title":86,"isFree":61,"description":87},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/02-practical-ai-use-cases/01-when-to-use-ai-and-when-not","When to Use AI and When Not To","How to tell when machine learning genuinely adds value and when a plain rule, a lookup, or a human is the better answer, using cost-benefit and the prediction-versus-exact-outcome test.",{"path":89,"title":90,"isFree":66,"description":91},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/02-practical-ai-use-cases/02-regression-classification-clustering","Regression, Classification, and Clustering","The three workhorse ML techniques and how to pick the right one for a use case, using the shape of the answer you need: a number, a category, or a set of natural groups.",{"path":93,"title":94,"isFree":66,"description":95},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/02-practical-ai-use-cases/03-real-world-ai-applications","Real-World AI Applications","The application areas the exam expects you to recognize on sight: computer vision, NLP, speech, recommendation systems, fraud detection, forecasting, knowledge bases, and agentic AI, each tied to the problem it solves.",{"path":97,"title":98,"isFree":66,"description":99},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/02-practical-ai-use-cases/04-aws-managed-ai-services","AWS Managed AI Services","The three layers of AWS AI/ML services and what each pre-built service does, so you can match a use case to the right managed service without building a model yourself.",{"path":101,"title":102,"isFree":66,"description":103},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/02-practical-ai-use-cases/05-traditional-ml-vs-foundation-models","Traditional ML vs Foundation Models","How to choose between a task-specific traditional ML model and a large general-purpose foundation model, driven by explainability, regulation, cost, and operational constraints, not just capability.",{"path":105,"title":106,"isFree":61,"description":107},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/03-ai-ml-development-lifecycle/01-anatomy-of-an-ml-pipeline","Anatomy of an ML Pipeline","The stages that carry a model from raw data to monitored production: data collection, EDA, preprocessing, feature engineering, training, tuning, evaluation, deployment, and monitoring, and why the pipeline is a loop, not a line.",{"path":109,"title":110,"isFree":66,"description":111},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/03-ai-ml-development-lifecycle/02-model-sources-and-deployment-options","Model Sources and Deployment Options","Where the model in your application comes from (pre-trained open source, a managed model, or one you train yourself) and how you serve it in production: managed API versus self-hosted, and real-time, serverless, asynchronous, or batch inference.",{"path":113,"title":114,"isFree":66,"description":115},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/03-ai-ml-development-lifecycle/03-aws-services-across-the-pipeline","AWS Services Across the Pipeline","Which AWS and Amazon SageMaker capability handles each stage of the ML pipeline, from data labeling with Ground Truth to drift detection with Model Monitor, so you can match a service to a stage on the exam.",{"path":117,"title":118,"isFree":66,"description":119},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/03-ai-ml-development-lifecycle/04-mlops-fundamentals","MLOps Fundamentals","What MLOps is and why ML systems need it: turning a one-off model into a repeatable, monitored, retrainable production system, and the core practices of experimentation, automation, monitoring, and retraining.",{"path":121,"title":122,"isFree":66,"description":123},"/courses/aws-certified-ai-practitioner/en/domains/01-ai-ml-fundamentals/03-ai-ml-development-lifecycle/05-model-and-business-metrics","Model and Business Metrics","How to measure whether a model works: the confusion matrix and the classification metrics built on it (accuracy, precision, recall, F1, AUC), regression metrics (RMSE, MAE), and the business metrics that decide whether it was worth building.",{"path":125,"title":126,"isFree":61,"description":127},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/01-genai-core-concepts/01-foundation-models-and-llms","Foundation Models and Large Language Models","Go inside the transformer architecture behind foundation models: how self-attention reads a whole sequence at once, what parameters actually store, and why an LLM is one kind of foundation model rather than a synonym for it.",{"path":129,"title":130,"isFree":66,"description":131},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/01-genai-core-concepts/02-tokens-chunking-and-embeddings","Tokens, Chunking, and Embeddings","Follow a document from raw text to a searchable vector index: how text becomes tokens, why long documents get split into chunks, and how embeddings turn meaning into numbers you can compare.",{"path":133,"title":134,"isFree":66,"description":135},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/01-genai-core-concepts/03-multimodal-and-diffusion-models","Multimodal and Diffusion Models","Understand how generative AI works beyond text: what makes a model multimodal, how diffusion models build images by removing noise step by step, and how diffusion differs from GANs and from transformer language models.",{"path":137,"title":138,"isFree":66,"description":139},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/01-genai-core-concepts/04-genai-use-cases","Generative AI Use Cases","Work through the use cases the exam names, grouped by the capability behind them: generating new content, transforming existing content, holding a conversation, and finding what matters.",{"path":141,"title":142,"isFree":66,"description":143},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/01-genai-core-concepts/05-the-foundation-model-lifecycle","The Foundation Model Lifecycle","Walk the seven stages the exam names, from data selection to feedback, and see why model selection replaces training as the decision that matters most when you build on a foundation model.",{"path":145,"title":146,"isFree":66,"description":147},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/01-genai-core-concepts/06-token-based-pricing","Token-Based Pricing","Learn how generative AI bills, why output tokens cost more than input tokens, and how on-demand, batch, Provisioned Throughput, and prompt caching change the cost of the same workload.",{"path":149,"title":150,"isFree":66,"description":151},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/01-genai-core-concepts/07-context-engineering","Context Engineering","Treat the context window as a budget you allocate rather than a box you fill: what competes for the space, why more context can make answers worse, and the strategies that decide what earns a place in the prompt.",{"path":153,"title":154,"isFree":66,"description":155},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/01-genai-core-concepts/08-agentic-ai-and-mcp","Agentic AI and the Model Context Protocol","See what turns a foundation model into an agent: the cyclical reasoning loop, tool use, memory, and multi-agent patterns, plus how the Model Context Protocol standardizes the connection between agents and the systems they act on.",{"path":157,"title":158,"isFree":61,"description":159},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/02-genai-business-value/01-what-genai-does-well","What Generative AI Does Well","The four advantages the exam names, grounded in what a foundation model can do that a purpose-trained model cannot: adapt without retraining, respond without a build cycle, hold a conversation, and produce content that did not exist.",{"path":161,"title":162,"isFree":66,"description":163},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/02-genai-business-value/02-genai-limitations-and-risks","Limitations and Risks of Generative AI","Hallucination, inaccuracy, nondeterminism, and interpretability are four different failures with four different fixes. Learn to tell them apart, and see how Amazon Bedrock detects the one you cannot prevent.",{"path":165,"title":166,"isFree":66,"description":167},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/02-genai-business-value/03-selecting-a-genai-model","Selecting a Generative AI Model","Turn the exam's list of selection factors into a working order of operations: eliminate on hard constraints, evaluate the survivors on your own task, then optimize for cost and latency.",{"path":169,"title":170,"isFree":66,"description":171},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/02-genai-business-value/04-measuring-genai-business-value","Measuring Generative AI Business Value","Model scores do not prove business value. Work through the metrics the exam names, from accuracy and cross-domain performance up to ROI, conversion rate, and customer lifetime value, with a full ROI calculation on a real workload.",{"path":173,"title":174,"isFree":61,"description":175},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/03-aws-genai-stack/01-amazon-bedrock","Amazon Bedrock","Understand Amazon Bedrock as the managed API layer for foundation models on AWS: one interface to many models, plus the built-in pieces (Knowledge Bases, Guardrails, evaluations, customization) that turn a model call into an application.",{"path":177,"title":178,"isFree":66,"description":179},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/03-aws-genai-stack/02-sagemaker-ai-and-jumpstart","Amazon SageMaker AI and SageMaker JumpStart","Draw the line between Bedrock and SageMaker AI: what you gain when you take control of training and serving, what you take on in exchange, and where SageMaker JumpStart sits between the two.",{"path":181,"title":182,"isFree":66,"description":183},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/03-aws-genai-stack/03-amazon-q-kiro-and-agentic-tools","Amazon Q, Kiro, and the Agentic Toolset","Separate the AWS tools that ship a finished AI assistant from the ones that give you parts to build with: Amazon Q, Amazon Quick, Kiro, Strands Agents, and Amazon Bedrock AgentCore.",{"path":185,"title":186,"isFree":66,"description":187},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/03-aws-genai-stack/04-why-build-genai-on-aws","Why Build Generative AI on AWS","Turn the exam's two lists, the advantages of AWS generative AI services and the benefits of AWS infrastructure, into arguments you can defend, and learn which word in a scenario points at which one.",{"path":189,"title":190,"isFree":66,"description":191},"/courses/aws-certified-ai-practitioner/en/domains/02-genai-fundamentals/03-aws-genai-stack/05-genai-cost-tradeoffs-on-aws","Cost Tradeoffs of AWS Generative AI Services","Work through the choices that decide a generative AI bill on AWS: on-demand versus batch versus Provisioned Throughput, what a custom model really costs, and how responsiveness, redundancy, and Regional coverage each carry a price.",{"path":193,"title":194,"isFree":61,"description":195},"/courses/aws-certified-ai-practitioner/en/domains/03-foundation-model-applications/01-fm-application-design/01-choosing-a-foundation-model","Choosing a Foundation Model","The exam names eight criteria for choosing a pre-trained foundation model. This lesson walks each one, what it means for your design, and the exam cue that points at it.",{"path":197,"title":198,"isFree":66,"description":199},"/courses/aws-certified-ai-practitioner/en/domains/03-foundation-model-applications/01-fm-application-design/02-inference-parameters","Inference Parameters","Same prompt, same model, different answers: the inference parameters that control how a model chooses each token, when to reach for each one, and the settings that quietly break a feature.",{"path":201,"title":202,"isFree":66,"description":203},"/courses/aws-certified-ai-practitioner/en/domains/03-foundation-model-applications/01-fm-application-design/03-retrieval-augmented-generation","Retrieval Augmented Generation (RAG)","Give a model access to your private, current data without retraining it: how RAG retrieves relevant passages at query time, how Amazon Bedrock Knowledge Bases automates it, and where the boundary with fine-tuning sits.",{"path":205,"title":206,"isFree":66,"description":207},"/courses/aws-certified-ai-practitioner/en/domains/03-foundation-model-applications/01-fm-application-design/04-vector-databases-on-aws","Vector Databases on AWS","The store that makes retrieval possible: what a vector database does, why RAG needs one, and the AWS services the exam expects you to name for holding embeddings.",{"path":209,"title":210,"isFree":66,"description":211},"/courses/aws-certified-ai-practitioner/en/domains/03-foundation-model-applications/01-fm-application-design/05-fm-customization-tradeoffs","Customization Tradeoffs","Prompt engineering, RAG, fine-tuning, and continued pre-training form a ladder of rising cost. This lesson maps what each one changes, what problem it solves, and why you start at the cheap end.",{"path":213,"title":214,"isFree":66,"description":215},"/courses/aws-certified-ai-practitioner/en/domains/03-foundation-model-applications/01-fm-application-design/06-ai-agents-in-applications","AI Agents in Applications","When a single prompt cannot finish the job: how agents reason across multiple steps and take actions, how Amazon Bedrock Agents orchestrate that loop, and where an agent fits versus plain RAG.",{"path":217,"title":218,"isFree":61,"description":219},"/courses/aws-certified-ai-practitioner/en/domains/03-foundation-model-applications/02-prompt-engineering/01-anatomy-of-a-prompt","Anatomy of a Prompt","Why the same model gives one team a useful answer and another team junk: the parts of a prompt, what each one does, and how instruction, context, input, and output indicator fit together.",{"path":221,"title":222,"isFree":66,"description":223},"/courses/aws-certified-ai-practitioner/en/domains/03-foundation-model-applications/02-prompt-engineering/02-prompting-techniques","Prompting Techniques","From zero-shot to chain-of-thought: the named prompting techniques, how much scaffolding each one adds, and which task each is built for.",{"path":225,"title":226,"isFree":66,"description":227},"/courses/aws-certified-ai-practitioner/en/domains/03-foundation-model-applications/02-prompt-engineering/03-prompt-engineering-best-practices","Prompt Engineering Best Practices","The concrete habits that turn a vague prompt into a reliable one: be specific, place the instruction well, use separators, give a fallback answer, and say what you do not want.",{"path":229,"title":230,"isFree":66,"description":231},"/courses/aws-certified-ai-practitioner/en/domains/03-foundation-model-applications/02-prompt-engineering/04-prompt-attacks-and-risks","Prompt Attacks and Risks","How attackers turn your own prompt against you: prompt injection, jailbreaking, and prompt leaking, plus the layered defenses including Amazon Bedrock Guardrails.",{"path":233,"title":234,"isFree":66,"description":235},"/courses/aws-certified-ai-practitioner/en/domains/03-foundation-model-applications/02-prompt-engineering/05-prompt-versioning-and-management","Prompt Versioning and Management","Treating prompts like code: Amazon Bedrock Prompt Management, the difference between a mutable draft and an immutable version, and how versioning gives you safe rollout and rollback.",[237,238],"ai-ml-fundamentals","genai-fundamentals"]