Part 4/16:
In virtual environments, NPCs often rely on simpler models like finite state machines (FSMs), switching between states such as hunting, fleeing, or hiding. These discrete states resemble rudimentary cognitive architectures, illustrating how the concept underpins a wide range of AI behaviors.
Core Components of Cognitive Architectures
Shapiro emphasizes that all such systems share three primary components:
Input Processing: Gathering data via sensors, telemetry, or user interactions.
Memory and Planning: Using memory (long-term, short-term, working) to store information, learn from experiences, and plan actions.
Output Generation: Acting upon the processed information through robotics, speech, or other forms of interaction.