Artificial Intelligence and Machine Learning: Core Concepts and Vocabulary Guide

Artificial Intelligence and Machine Learning subject snapshot

Studies search, representation, supervised and unsupervised learning, neural methods, evaluation, and responsible use. This core concepts and vocabulary guide helps a U.S. college learner build a connected concept map and explain essential terms instead of memorizing isolated definitions. Course titles, depth, notation, laboratory rules, and assessment weights differ; the instructor syllabus remains the controlling local source.

Cluster: Computing and Data. Useful preparation: programming, linear algebra, probability, statistics, calculus, and data management. Target evidence: a model experiment with baseline, protocol, results, error analysis, risk review, and model card.

How to use this resource

Start with the current syllabus, calendar, required materials, accessibility information, academic-integrity policy, safety rules, and grading method. Mark which ideas are already secure, which require prerequisite repair, and what the instructor accepts as evidence. Use the guide to organize learning—not to guess undisclosed exam questions or replace assigned work.

Core concept 1: search and optimization

Define search and optimization in the language used by the course, then add a representation, valid example, near-miss, governing relationship, and a question that distinguishes it from feature and representation choices. Connect the concept to this subject-wide method: define task and metric, create a baseline, separate training from evaluation, analyze errors, test robustness, and document use limits.

Core concept 2: feature and representation choices

Define feature and representation choices in the language used by the course, then add a representation, valid example, near-miss, governing relationship, and a question that distinguishes it from supervised learning. Connect the concept to this subject-wide method: define task and metric, create a baseline, separate training from evaluation, analyze errors, test robustness, and document use limits.

Core concept 3: supervised learning

Define supervised learning in the language used by the course, then add a representation, valid example, near-miss, governing relationship, and a question that distinguishes it from unsupervised learning. Connect the concept to this subject-wide method: define task and metric, create a baseline, separate training from evaluation, analyze errors, test robustness, and document use limits.

Core concept 4: unsupervised learning

Define unsupervised learning in the language used by the course, then add a representation, valid example, near-miss, governing relationship, and a question that distinguishes it from evaluation and responsible AI. Connect the concept to this subject-wide method: define task and metric, create a baseline, separate training from evaluation, analyze errors, test robustness, and document use limits.

Core concept 5: evaluation and responsible AI

Define evaluation and responsible AI in the language used by the course, then add a representation, valid example, near-miss, governing relationship, and a question that distinguishes it from search and optimization. Connect the concept to this subject-wide method: define task and metric, create a baseline, separate training from evaluation, analyze errors, test robustness, and document use limits.

Build the concept network

Draw arrows among the five concepts and label every connection with a verb such as causes, constrains, represents, measures, transforms, supports, or contradicts. Add Compare a simple baseline and a complex classifier, then explain when the added complexity is not justified. and mark which concepts are necessary at each decision. Rebuild the map from memory two days later and correct it against approved course sources.

Accessibility, integrity, and safety

Use approved accommodations and accessible formats early. Follow laboratory, clinical, field, studio, technology, privacy, human-subject, copyright, and professional-scope rules. Cite source and tool use as required. Never use this guide to bypass assessment rules, perform unauthorized security testing, make a diagnosis, or provide individualized legal, medical, financial, or safety instructions.

Instructor or tutor conference questions

• Which two concepts create the greatest prerequisite bottleneck in this section?
• What does a complete explanation include beyond the final answer?
• Which errors require immediate correction before the next unit?
• What practice best matches the actual assessment format without revealing protected questions?
• Which approved source or office should resolve a changing rule, safety issue, or accommodation need?

Related Computing subject guides

Artificial Intelligence and Machine Learning — Core Concepts and Vocabulary Guide · Programming Fundamentals — Core Concepts and Vocabulary Guide · Python Programming — Core Concepts and Vocabulary Guide

Open-learning reference and editorial boundary

Use the Computing open-learning catalog, OpenStax subject library, and MIT OpenCourseWare only when they fit the instructor’s scope and license terms. This is an original independent Exams.fit study guide; it does not reproduce textbooks, guarantee a grade, predict protected exam questions, or replace the current syllabus, instructor, laboratory manual, clinical protocol, institutional policy, or qualified professional. Reviewed August 2, 2026.