Introduction to Artificial Intelligence
What this domain tests
Introduction to Artificial Intelligence checks how AI is defined, how it differs from conventional computing, and how intelligent agents are described. Expect history milestones, the Turing Test, rational-agent thinking, and PEAS descriptions.
Core rules and concepts
- AI is the study of systems that perceive, reason, learn and act. Definitions split into thinking/acting humanly (Turing Test) and thinking/acting rationally (rational agent).
- Conventional computing executes explicit, deterministic instructions; AI often learns patterns from data and handles uncertainty.
- History: the Dartmouth workshop in 1956 named the field; expert systems grew in the 1980s; machine learning revived; deep learning rose after 2012.
- An agent perceives its environment through sensors and acts through actuators. A rational agent chooses the action expected to maximise its performance measure.
- PEAS = Performance measure, Environment, Actuators, Sensors.
- Environment properties: fully/partially observable, deterministic/stochastic, episodic/sequential, static/dynamic, discrete/continuous.
Worked example
Describe a spam filter as an agent. Performance: accuracy and low false positives. Environment: incoming email. Actuators: move a message to inbox or spam, optionally flag it. Sensors: read sender, subject and body. The filter is rational if it maximises expected correct classification, not if it is simply labelled smart.
Traps that cost marks
Do not call every program an AI agent; a calculator has no perception or autonomous action in a meaningful environment. The Turing Test measures indistinguishability in dialogue, not consciousness. Do not omit the performance measure in PEAS. Do not claim AI always beats conventional computing: for fixed arithmetic rules, conventional code is simpler and more reliable.