Task 5 not as important Flashcards
How do detect the best model in step 4 ?
- Design a experimental conditions that lead to opposite qualitative predictions from the 2 models
- Qualitive test are often supported by quantitive tests
- the qualitive answer which fits better will tell u which model to choose
What are some adavantaged of rules over logic models?
- rules do not have to be interpreted as universall true (default)
- can better represent strategic information about what to do
- Rules are good at finding the right plan but are bad at decision making
What is meant by default ?
- which is a rough generalisations that admit exceptions
What is the difference between logic and rule based system ?
- logic based system focus on deduction regarding thinking
- rule based system focus on SEARCH regarding thinking
- rule based system have less representational power but more computational
What is meant by representational power ?
- How much knowledge about the world can be represented?
What is meant by computational power ?
- How powerful and efficient are rule-based system
What kind of processing strategie rule based models use ?
- Can be either serial or parallel:
- Conscious thought = serial processing
- unconscious thought parallel processing
What is meant by slow incremental learning ?
- it describes a rule usefuleness via applying a numerical value towards it
- the stronger the rule the more value it has
What kind of processing strategie rule based models use ? (in human being)
- Can be either serial or parallel:
- Conscious thought = serial processing
- unconscious thought parallel processing
What is meant by a buffer ?
- Buffers are an interface between the porcedural memory and the other components of the ACTR
- can be either visual (perception) manuel (motor actions = pressing button) imaginal (working memory) retrieval (declarative memory), goal buffer
- The prdouction rules make use of different buffers to answer to answer the current goal step by step !
- One buffer hold one chunck
Explain in single steps how the ACTR work ?
- 1 Cycle: After the current goal is decided the conflict resolution start (achiving a goal
- It first goes to procedural memory which asks the declarative memory if there is a solution
- If so then it send the retrieval result back to the current goal
2 cycle: Current goal needs new information from declartive memory.(pooed goal) new production rules have to be formed to reach the goal (production complimations)-> since new rules have to be trained first we have to aply to an easier problem to the rules (transform goal)
What is meant by archictectural assumption ?
- Makes general claim about the nature of human cognition
- present in all cognitive model
What is meant by Auxiliary assumption ?
- a claims on how to deal with special task in a case which needs more flexibility
How do we calculate the activation lvl of a chunck ?
- Activation Level = Base-Level Activation (number of times it has been rehearsed and the time that has passed from last rehersal) + Associative Activation (strength of bond btw an item and the required chunk )
What is meant by aductive reasoning ?
- running a rule backward to explain what is currently happening
- risky but good for hypothesis
- Example: if someone walks angry out of the class u would assume he just got a bad grade but there are other possibilities