Task 5 Adaptive Control of Thought

Question 1

ACT-R

Answer 1

Cognitive model of rational analysis that states that each component of the cognitive system is optimized with respects from demands from the environment

Question 2

Goal Stack

Answer 2

Part of the ACT-R model that contains the cognitive goals, which are popped to the current goal, and to which goals can be added at anytime.

Question 3

Chunks

Answer 3

A set of elements that symbolize a memory in declarative memory

Question 4

Declarative memory

Answer 4

A type of memory of facts and events, represented in chunks

Question 5

Procedural memory

Answer 5

A type of memory that stores information about how to perform certain tasks, that is represented as production rules

Question 6

Retrieval request

Answer 6

In the ACT-R model, a rule that is fired in procedural memory may need chunks of the declarative memory and this request retreives it.

Question 7

Production compilation

Answer 7

Chunks from the declarative memory are sent to the procedural memories, where they are used to create new production rule

Question 8

Production rule

Answer 8

A rule that has a condition and an action, structured as and IF -> THEN statement; they have psychological plausability and almost as good as formal logic

Question 9

Symbolic level

Answer 9

Basic step in information processing, the learning of new chunks and production rules in ACT-R

Question 10

Sub-symbolic level

Answer 10

Parallel process in which learning as change of connections between chunks & production rules takes place in ACT-R

Question 11

Fan effect

Answer 11

• The greater the number of facts related to a concept, the slower the recall of one fact
• Due to limited resource “activation”: Depends on level of association of single chunks

Question 12

Computational Power ACT-R

Answer 12

1. Problem solving –> Good
2. Planning –> Good
3. Decision-making –> Not very good! Need other supplementary processes
4. Explanation –> Good
5. Learning –> Good
6. Language –> Depends on view if language is made of rules or of loose associations (connectionist view)

Question 13

Psychological plausibility

Answer 13

Most psychologically plausible from all computational-representational approaches

Question 14

Neurological Plausiblity

Answer 14

Crude analogy between rules and neurons connected by snapses (If one neuron fires, it can then cause the firing of neuron connected to it)

Question 15

Practical Applicability

Answer 15

• Expert systems in industry and government

- Education