23. Cognitive Architectures Flashcards
(21 cards)
What are cognitive architectures?
They are hypotheses about the fixed structures in the mind that provide a basis for cognitive behavior across tasks.
What are the three key components of cognitive architectures?
Fixed structures (architecture), Task-specific knowledge, Skills.
What are the three layers of the tri-level hypothesis in cognitive science?
Computational, Algorithmic, Implementational.
What is the ‘computational’ level concerned with?
What the system does and why, focusing on goals and logic.
What is the ‘algorithmic’ level concerned with?
How the system does it, focusing on representations and processes.
What is the ‘implementational’ level concerned with?
How the system is physically realized (e.g., neural or hardware level).
What does the Marr’s levels framework help us understand?
The relation between mind (software) and brain (hardware).
What are ‘symbol systems’ in classical cognitive architectures?
Systems that manipulate symbols according to formal rules.
Name a prominent symbolic cognitive architecture.
SOAR or ACT-R.
What characterizes a connectionist cognitive architecture?
Distributed representations and processing units connected in networks.
What is a hybrid cognitive architecture?
An architecture that integrates both symbolic and connectionist approaches.
What are examples of hybrid architectures?
CLARION and Sigma.
What is SOAR’s central hypothesis?
All intentional behavior can be understood as goal-driven problem solving.
What are ‘productions’ in SOAR?
Condition-action rules used to apply knowledge.
What does ACT-R stand for?
Adaptive Control of Thought – Rational.
What are the modules in ACT-R?
Perceptual, Motor, Goal, Declarative Memory, and Procedural Memory.
What is the role of procedural memory in ACT-R?
Contains production rules for decision making.
What is the role of declarative memory in ACT-R?
Stores factual knowledge as chunks.
What is ‘chunking’ in ACT-R?
Creating new knowledge from problem-solving episodes.
Why are cognitive architectures useful?
They offer a framework for integrating findings and modeling cognition across tasks.
What is ‘bounded rationality’ in the context of cognitive architectures?
The idea that decision-making is rational, but limited by cognitive constraints.
