204: Problem Solving, Reasoning, Decision Making L11

Problem space

various states of the problem

Problem state

representation of problem in some degree of solution

operator

action that transforms problem state into another problem state

problem solving

searching sequence of states in a problem space to go from start state to goal state

influenced by; operators available, how solver selects operator 

acquisition of operators done thorough

random discovery, direct instruction, analogy/imitation

3 criteria to acquiring operators

backup avoidance, difference reduction, means ends analysis

backup avoidance

avoid operators who take us back to previous state

difference reduction

select operators which eliminate differences between current state & end goal

focus on next step over end goal

not always optimal 

means ends analysis

creates new subgoal for operator

identify biggest difference between current and goal state, eliminate 

problem representation

how different states of a problem are represented

successful PS depends on this

incubation effect

solutions of a problem come easier after one has ignored it

Cheap necklace problem

people forget inappropriate ways of solving problem

3 ways irrationality is found in reasoning/decision making

reasoning about conditionals, reasoning about probabilities, subjective utility

reasoning about conditionals

if A, then B

Wason selection task: ony 10% made right combination

when presented with neutral material, people have difficulty seeing negation of the consequent - assume A is true

permission schema, probabilistic interpretation

permission schema

performance can be improved when material is meaningful 

probabilistic interpretation

poor performance on Wason as we select cards using a probabilistic model

randomly soften logic: B will probably occur when A occurs

check headlights of cars with broken taillights - reasonable and realistic inference

reasoning about probabilities

when calculating probability, we have to account for the info we have, which changes according to quality of info/base rate

prior prob: prob that statement is true before evidence is considered (base rate)

posterior prob: prob that statement is true after considering evidence

• prior prob, evidence (structure of event) quality of info (how reliable evidence is)

• Bayes theorem: updating beliefs after receiving new evidence

• base rate neglect

• biased in our estimates as we use memory and similarity

subjective utility

value someone places on something, non-linear, we use subjective > objective values