Part 1: Core Concepts (2 Marks)
- Artificial Intelligence: The branch of computer science dedicated to engineering computational systems capable of performing tasks typically requiring human intelligence (reasoning, learning, perception, problem-solving).
- Turing Test: An operational test of machine intelligence proposed by Alan Turing (1950) where a human interrogator in blind communication tries to distinguish whether responses come from a human or a computer.
- Intelligent Agent & PEAS Framework: An entity that perceives its environment through sensors and acts upon that environment via actuators. Characterized by PEAS: Performance measure, Environment, Actuators, and Sensors.
Part 2: Structural Frameworks (10 Marks)
Intelligent Agent Architecture & Environment Classification
ENVIRONMENT
| SENSORS | ---> | AGENT BRAIN | --> | ACTUATORS |
|---|---|---|---|---|
| (Camera, | (State, Goals, | (Motors, | ||
| Keyboard) | Utility Model) | Display) |
Environment Taxonomies: Fully vs Partially Observable, Deterministic vs Stochastic, Episodic vs Sequential, Static vs Dynamic, Discrete vs Continuous, Single-Agent vs Multi-Agent.
Part 3: Distinction Tables
| Basis of Comparison | Goal-Based Agent | Utility-Based Agent |
|---|---|---|
| Decision Metric | Binary satisfaction: whether a desirable goal state is achieved or not. | Continuous utility function: quantifies how good/happy a state is. |
| Trade-Off Handling | Struggles when multiple competing goals exist. | Handles conflicting goals by optimizing trade-offs (e.g. speed vs safety). |
| Mathematical Model | Goal state test: |
Real-valued utility score: |
Part 4: Worked Problems & Exam Traps
Formulating PEAS for an Automated Taxi Driver
- Performance Measure: Safety, speed, legal compliance, passenger comfort, fuel efficiency, maximum profit.
- Environment: Urban roads, pedestrian traffic, variable weather conditions, fellow motorists, traffic signals.
- Actuators: Steering wheel, accelerator pedal, brake calipers, horn, LCD display, turn signals.
- Sensors: LiDAR, video cameras, ultrasonic sonar, GPS transceiver, speedometer, engine sensors.
An agent’s rationality is NOT omniscience or clairvoyance! Rationality depends strictly on expected performance given the historical percept sequence, not perfection or knowing the future.