Exam 3 Flashcards
Well- and Ill-Defined Problems:
What is a problem? What are its 3 components?
- problem: a situation without an immediately apparent standard or routine way of reaching a goal
- 3 components:
1. ) goal state (what you’re working towards)
2. ) initial state (info you have at start of problem)
3. ) operators (action you can take to move from initial to goal state)
Well- and Ill-Defined Problems:
What is the difference between a well-defined and an ill-defined problem?
- well-defined: all 3 components are known
- ill-defined: at least one component is undefined/not known; possibly more difficult to solve
Well- and Ill-Defined Problems:
What are Goel’s four stages of ill-defined problem-solving?
- 1.) problem scoping - what problem entails; use knowledge you have to study problem statement
- 2.) preliminary solution generation - trying to figure out how you’re going to go about finding a solution, path to solution; “lateral transformation”
- 3.) refinement - how it would look if you took a particular path; “vertical transformation”
- 4.) detailing of solutions - final commitment to one path/solution
Well- and Ill-Defined Problems:
What is the difference between a lateral transformation and a vertical transformation through problem space, according to Goel?
- vertical: look down a path to see what would happen next, what steps
- horizontal: comparing choices/options at each move
Well- and Ill-Defined Problems:
What role does the prefrontal cortex play in solving well- vs. ill-defined problems, according to Goel?
- location where the problem mainly occurs
- hemisphere difference in how approach the problem:
- left hem. damage: impairs well-defined problem solving
- right hem. damage: impairs ill-defined problem solving
Well- and Ill-Defined Problems:
What is a problem space? How is it constructed? Describe the tic-tac-toe example
- problem space: visualizing the solving of a problem
- depict stages of a solved problem or turns in a game as a tree
- limited by STM (not enough memory for all possible moves)
- tic-tac-toe: ex. first move of X has 9 choices (branches), second move (next level) is O and has 8 choices
Problem-Solving Strategies and Heuristics:
What is an algorithm? What is a heuristic? What are the advantages/disadvantages of them?
- algorithm: set of procedures that you follow to solve a problem (ex. long division, recipe)
- advantage of algorithm is if you follow it exactly you won’t make a mistake
- heuristic: short-cut for how to do something; help you get to solution more quickly
- advantage of heuristic is it’s quick if you don’t need an exact answer
Problem-Solving Strategies and Heuristics:
What are the heuristics of random guess?
- provides shortcut, gets you kind of closer to the answer
Problem-Solving Strategies and Heuristics:
What are the heuristics of hill-climbing? Discuss how the Rubik’s cube and hobbits and orc problem relate to this type of heuristic
- pick next move to make it look like you’re getting closer to the solution/goal state
- sometimes looks like you’re moving closer but you may be moving further away from the solution
- Rubik’s cube problem utilizes this heuristic
- Hobbits and orcs problem: people have difficulty making a necessary backward move bc it looks like it’s getting them further away from the solution
Problem-Solving Strategies and Heuristics:
What are the heuristics of means-end analysis? How does it relate to the Tower of Hanoi task?
- breaking a big problem down into smaller problems
- Tower of Hanoi - focus on getting the next ring in goal state
- efficient to deal with large problems
- first choice may be significant for next moves but likely don’t realize it
Problem-Solving Strategies and Heuristics:
What is functional fixedness? What are the two classic examples of this?
- functional fixedness: fixated on object’s typical use and cannot see how to use it differently in nontraditional fashion
- ex. Two String Problem - have to hold two strings but they aren’t long enough, have to use pliers to make a pendulum
- ex. Candle on the Wall - have to mount a candle onto the wall and only have the candle and a box of tacks to do so
Problem-Solving Strategies and Heuristics:
According to Simon and Newell what are the basic characteristics of the human information-processing system?
- it’s serial (deals with one thing at a time)
- elementary processes take less than one second
- inputs and outputs are stored in limited STM
- LTM has a higher capacity but it’s slower to store info
Problem-Solving Strategies and Heuristics:
Could a computer model solve problems the same exact way that humans do?
- No
- humans are limited by short-term memory
Problem-Solving Strategies and Heuristics:
What is the difference between a “scan and search” strategy and a “progressive deepening strategy”? Which one do people typically use and why?
- “scan and search” - identify most promising node and look ahead before committing to solutions; helps to avoid problems; most efficient strategy
- “progressive deepening” - look for the shorter path to take through a space, may not look at all nodes or branches
- ppl use progressive deepening more due to limited STM bc can’t remember all possible outcomes in scan and search
Problem-Solving Strategies and Heuristics:
What info is used to define a problem space?
- task instructions
- previous experience with similar tasks
- programs in LTM for such tasks
- programs in LTM to construct problem spaces
- information learned while solving the problem
Problem-Solving Applications:
Describe what the Carraher at al study did and found. What mistakes did the children make?
- explore how child street vendors did math on the streets
- informal task: researchers asked prices on the street in “natural” environment
- formal task: formal test of mathematical procedures, operation problems, word problems
- performance overall on formal task wasn’t as good as informal; however did better with word problems then the straight up math problems
- mistakes: swapped order of operations, incorrect operator, didn’t use operator at all
Problem-Solving Applications:
What does it mean to reason by analogy when it comes to problem solving? How does the Gick and Holyoak study show this?
- reason by analogy: make connections/adapt or modify previous solution to current situation
- Gick and Holyoak: given problem with solution (fortress/dictator) and given another problem without solution (tumor/doctor)
- see similarities between the two problems’ elements and adapt existing solution to fulfill current needs
Problem-Solving Applications:
Describe the two examples of real-world problem solving discussed in class
- ex. adapt plastic bottles to act like a light bulb
- ex. nail polish that detects some date-rape drugs
Conditional Reasoning:
What is the difference between conditional, deductive, and inductive reasoning?
- conditional: given conditional statement and extra info
- deductive: determining whether a conclusion is valid using rules of logic
- inductive: conclusions you draw are based on probability
Conditional Reasoning:
What is a conditional statement?
- “if then” statement that tells us something about the world
- antecedent = “if”
- consequent = “then”
- only tells us what will happen if antecedent occurs, don’t know what happens if antecedent doesn’t occur
Conditional Reasoning:
What is modus ponens, denial of antecedent, affirmation of consequent, modus tollens?
- modus ponens: told “if” is true, conclude that “then” is true (ex. told he did clean up his toys (“if”), conclude that he got a cookie)
- denial of antecedent: told “if” is false, conclude “then” is false; invalid conclusion
- affirmation of the consequent: told “then” is true, conclude if is true, but really an invalid conclusion
- modus tollens: told “then” is false, conclude “if” is false
Conditional Reasoning:
What is the traditional Wason Selection Task? What cards are people supposed to turn over,? What cards do they actually turn over and why?
- 4 cards, one side number, one side letter
- “if card has D on one side, then it has a 3 on the other side”
- should pick D and 7
- mistake: chose 3 (affirmation of consequent) or D (modus ponens)
- also task with squiggle cards
Conditional Reasoning:
What is the modified Wason Selection Task? Describe the 3 examples from class
- modified tasks tap into existing schemas so easier to reason with (relevance to real world experiences)
- ex. “if a person is drinking beer, then they are over 21”
- ex. “if an envelope is sealed, then it has a 20 cent stamp”
- ex. “if entering the country, then must have cholera vaccine”
Deduction:
What is a syllogism?
- syllogism: 3 statement set
- first two statements are premises
- third statement is a conclusion
- goal to evaluate if 3rd statement is valid
- if involves categories of items, then can assess validity using venn diagrams
Deduction:
What mistakes do people make when performing deductive reasoning tasks?
- tend to focus on whether it’s true in the real world, rather than validity
- some things might be true IRL but invalid
- some things may be false IRL but valid
Deduction:
How did Rips study deductive vs. inductive reasoning? What were the results?
- 4 conditions ranging in whether valid/invalid ‘
A.) deductively correct, causally consistent w/ reality
B.) deductively correct, casually inconsistent
C.) deductively incorrect, casually consistent
D.) deductively incorrect, causally inconsistent - C and D did not include the premises so not logical
- had to rate valid/invalid (deductive) and strength of argument (inductive)
- results: don’t process validity the same way as strength; deductive and inductive reasoning go through separate processes
Induction:
What’s the difference between inductive and deductive reasoning?
- inductive: use info to make a judgement
- deductive: use logic to make decision
Induction:
Describe the Mastermind inductive reasoning task
- take in evidence w/ each round
- don’t know initially correct color or position, so make a hypothesis
- use the feedback to reformat and work out alternatives
- keep track of evidence
Induction:
Describe the Bruner et. al task rules
- simple rule: one feature
- conjunctive rule: “and” rule, 2 features, each from a diff dimension
- disjunctive rule: “or” rule, only need one of 2 parts to be a yes
- negative rule: “not” rule, pick one feature
- relational rule: ex. number of shapes matches number of borders
Induction:
What rules in the Bruner et. al task are easiest and which are hardest? What strategies are available?
- simple rule is easiest
- conjunctive is easier than disjunctive
- strategies:
- successive scanning: go from yes card and change one feature
- focus gambling: keep only one feature from something you already know
“Rationality” and Reasoning Applications:
What does it mean to be rational or have reason?
- rational thought is:
- intelligent (based on knowledge)
- orderly, regular, complete (organizes all info)
- objective (based on facts not opinion)
“Rationality” and Reasoning Applications:
Are people rational? What is the notion of bounded rationality?
- people can be rational in certain circumstances with low cognitive load (and irrational when high cog. load)
- “bounded rationality” - we can be rational to the extent that our cognitive limitations allow
- these limitations include attention, working memory, executive functions
“Rationality” and Reasoning Applications:
Discuss the first two problems in Tversky and Kahneman’s study
- one group of subjects choose between A and B
- options were mathematically almost the same except A is absolute/sure while B is risky
- another group chooses between C and D
- mathematically same as A and B but framed in terms of lives lost
- when framed in terms of gains - people pick sure thing
- when framed in terms of losses - people pick risky choice
- emotion might play a role
“Rationality” and Reasoning Applications:
Discuss problems 3 and 4 in Tversky and Kahneman’s study
- failure to combine decisions
- 2 sets of 2 options
- one option sure gain w/ risk, other option sure loss w/ risk
- should be thinking about combinations
- people usually treat the sets as separate and choose sure gain then choose loss; not the best outcome
- problem 4 tells to subjects to think about them together but still treat them as separate
“Rationality” and Reasoning Applications:
Discuss problems 5-7 in Tversky and Kahneman’s study
- failure to account for contingencies
- problem 5: sure gain or risky gain - people choose sure gain
- problem 6: 25% likelihood of reaching stage to choose C or D - so really no longer a sure thing
- problem 7: choose between options having already advanced
for prob. 6 and 7 people still prefer guaranteed wins but don’t take into account the low probability of that win
“Rationality” and Reasoning Applications:
Discuss problems 8 and 9 in Tversky and Kahneman’s study
- failure to account for similarity in outcomes
- problem 8: lost $10 bill, will you still buy $10 ticket; people do
- problem 9: lost ticket, will you buy $10 ticket; people don’t
- lose same amount of money either way
- difference in that in problem 8 the $10 hasn’t been assigned to anything yet
“Rationality” and Reasoning Applications:
Discuss problem 10 in Tversky and Kahneman’s study
- $5 savings is dependent on proportion of cost
- calculator $15 regularly, $10 on sale
- or $125, regular, $120 on sale
- people willing to drive 20 minutes to buy the $10 one, they don’t treat $5 as $5
“Rationality” and Reasoning Applications:
What is the strawman logical fallacy?
- intentionally misrepresent someone’s argument to make it easier to attack
“Rationality” and Reasoning Applications:
What is the slippery slope logical fallacy?
- if we allow this to happen, them this other thing will happen, and other thing will happen
“Rationality” and Reasoning Applications:
What is the ad hominem logical fallacy?
- attack on opponent is based on personal characteristics rather than their argument
“Rationality” and Reasoning Applications:
What is the gambler’s logical fallacy?
- events occur in “runs,” or that something is due bc it’s been a while, but really events are statistically independent
“Rationality” and Reasoning Applications:
What is the bandwagon logical fallacy?
- popularity is taken as a measure to indicate if something is good
“Rationality” and Reasoning Applications:
What is the black or white logical fallacy?
- presenting a situation as if there is no middle ground or “gray area”
“Rationality” and Reasoning Applications:
What is the appeal to authority logical fallacy?
- trusting an authority figure on a topic of which you know little about
“Rationality” and Reasoning Applications:
What is the appeal to nature logical fallacy?
- it it’s natural it’s assumed it is good
“Rationality” and Reasoning Applications:
What is the anecdotal logical fallacy?
- using personal stories rather than actual scientific evidence, or giving equal weight to personal stories and evidence
“Rationality” and Reasoning Applications:
What is the appeal to emotion logical fallacy?
- using emotion to win an argument rather than actual logic or reason
“Rationality” and Reasoning Applications:
What is the tuquoque logical fallacy?
- criticizing the criticizers instead of the argument; answering criticism with other criticism
- ex. they did it too
“Rationality” and Reasoning Applications:
What is the Texas sharpshooter logical fallacy?
- pick bits of data that support point and ignore parts that don’t support it
Bayes’ Theorem and the Expected Utility Model:
What is Bayes’ Theorem? What does it allow us to do? Do people follow it when making decisions?
- people don’t follow this
- allows us to figure out exact probability by combining base rate (prior probability, “H”) with evidence (conditional probability, “E”)
“Rationality” and Reasoning Applications:
What are framing effects? What is the certainty effect? The pseudocertainty effect?
- framing effects: how something is worded
- certainty effect: if framed in terms of gains, people will want the sure thing
- pseudocertainty effect: no longer a sure thing
Bayes’ Theorem and the Expected Utility Model:
What is the Expected Utility Model? Do people follow it when making decisions?
- allows you to figure out which of two options is mathematically better
- people are motivated to win and may choose the higher prob. of winning
- gains: prob. of winning x amount to be won
- loss and gains: (prob. winning x amount to win) - (prob. lose x amount to lose)
Decision-Making Heuristics:
What is base-rate neglect?
- ignore given info about statistical probability/likelihood
Decision-Making Heuristics:
Describe De Neys et. al’s study
- given description of fictional person and given statistic
- 4 diff conditions:
- incongruent: heuristic and probability lead to different choices
- congruent (control): heuristic and probability lead to same choice ‘
- neutral (control): cannot use heuristics, must be probability only
- heuristic (control): cannot use probability, must use heuristic only
- results: high accuracy for control conditions and in incongruent condition longer response time when used base-rate
Decision-Making Heuristics:
What role do the anterior cingulate cortex and right lateral prefrontal cortex play in heuristic decision-making?
- ACC: involved in conflict monitoring; more active in incongruent trials, indicates that subjects detected conflict
- RLPFC: differences in activity depending on whether selected base-rate or heuristic; more activity with base-rate response
Decision-Making Heuristics:
What is the representativeness heuristic?
- how well something fits our idea of something else
- description much more representative of a certain subset of the group and increases your idea of the likelihood while ignoring the base-rate
Decision-Making Heuristics:
What is the availability heuristic?
- assume something happens more often if we can retrieve many memories of that something happening
- we assume the statistical probability is higher than it actually is
- ex. media replaying events
Decision-Making Applications:
Discuss the Hammerton study. Why did people make those mistakes? How does Bayes’ Theorem help us evaluate the results of medical screenings?
- examined how people use base-rate info to make medical judgments
- fictional disease and given statements
- 90% chance person w/ disease will be recorded positive
- if no disease 1% chance for false positive
- roughly 1% of pop has disease
- subjects median response was 85% chance of having disease
- really should be about 50%
- confidence ratings very high, even when just guessing without all the given info
- overall people weren’t using base-rate info - instead maybe used their memory to try to recall instances
Bayes’ theorem takes into account all statistical probabilities of having the disease, false pos., etc.
Decision-Making Heuristics:
Which heuristic is involved when we use media coverage to determine risk or base rate probability?
- availability heuristic
