PSYC 213 Lecture Notes - Lecture 20: General Problem Solver, Alphago, Logicism
11 May 2018
School
Department
Course
Professor
PSYC 213 (April 3rd) Reasoning & Risk – how we make judgments & choices
GPS – The general problem Solver
• The first AI programs designed to describe problem solving, can use to solve all problems
• Works only for well-defined problems : tells how we move to a problem space that has specific set goals.,
constraints & operators
• Used to figure out how we solve all problems
• Used the means-end analysis heuristic
o We always consider the ultimate goal & create conditions to get closer to it
o These condions are sub-goals that are defined by the operators in the problem space
o This program was thought to mimic how people solve tasks
Recent simulations: Alpha-Go
• A computer program, simulation that was thought to play the game Go & that beat the world champion at the
game of Go : relies on strategies, logics, a very complex & ill-defined game of strategy
• It takes many years for humans to be good at this game
• Alpha-Go was not given explicit instructions or rules for how to play this game. Instead it was showen examples
o It developed its own strategies in under 40 days (far faster than any human could)
o A major sign of creativity
Maybe we can use this computer programs to find solutions to ill-defined problems that we have (cures for diseases?)
Expertise – experts define, view & encode a problem in their area of expertise differently than novices.
• There might not be any anatomical differences in the brain but you do see differences in how people use their
brains – functional differences.
• What makes an expert? – practice or 10 000 hour (10 years rule)
Creativity – see something differently
• There are ways that can helps us or stops or from being creative
• Habitual thinking can stop us from being creative – when we rely on habits
o Strong but wrong effect – the tendency to default to ‘pre-programmed’ sequences bc they don’t require a lot of
conscious effort. This can be thoughts or behaviours when we do something
▪ We do this bc this doesn’t require a lot of effort or processing. The problem is that if we over use this, it
stops us from incorporating new info and see things differently
▪ This can stop you from seeing things in a new way.
▪ Ex. after moving to a new apartment, the next day you walk to your old apartment on your way ‘home’
from work.
▪ Don’t incorporate new info
o Can think about it as a mental slip which can lead to faulty, illogical behavior, decisions that we will make.
Reasoning, Logic & Risk “do people follow, understand logic?”
• 2 ways to research this question by examining deductive inference
o Syllogisms
o Wason’s 4 card task
• Heuristics & biases in ‘everyday’ thinking – ex when we do not follow logic
• Risky decision making – risking loosing something vs. not gaining smh T the relation to emotion.
Reasoning – complex cognitive construct that involves every other cognitive processes
• Influenced by factors that capture our attention, memory, language that’s used to frame what we reason about, our
ideas about problem solving & how problems are best solved
• Deductive inferences – the thought process that leads us to some conclusion from info that we’re given
The mental act of drawing conclusions from a set of premises (percepts, thoughts)
• Logicism – philosophical idea that when we reason, we do so with logic
• There is some support for the idea that we are logical, we as humans can spot errors in logical deductions.
Humans may be inheritably logical
• People with no logical reasoning can still make logical deductions (infants do not have logical reasoning yet)
o Infants (12-19 months) watched animations that included 2 objects with similar tops
o During the animation, the 2 objects were placed behind a screen that hid these 2 objects: one was ‘scooped’ up
and placed in a bucked with only the top visible. Could only see the the top of the objects. Bc the top of the 2
objects looked the same, the baby doesn’t know what object was scooped up and don’t know which object was
scooped out.
find more resources at oneclass.com
find more resources at oneclass.com
o The screen is lowered & reveals the other object that wasn’t hided.
o Infant can use logic, logical deduction to infer what object is in the bucket
o The researchers monitored the baby’s eye movements during the final test phase : the baby was shown the
object in the bucket.
▪ In one condition – the object that was showed was consistent with what logical deduction could have
been made. In the other condition, the object was inconsistent with the logic.
o Infants looked longer at the outcome during the test phase that was inconsistent with logical deduction. When
the object did not match with what their logic would be, they spent more time looking at it.
o This suggests they are puzzled by what they’re looking at. These babies who cant speak yet, they reasoned on
what object was hiden & use logical deduction.
• We don’t always use our logic. The types of errors that we make are common and consistent across people.
Syllogisms – we use something in addition to logic to come to conclusions & to reason.
• Logical arguments, structured so that you have this 2 premises (statements) & your job is to determine if these 2
premises fit with a given conclusion state that you have. You have info from 2 premises & you use that info to
deduce a new truth that’s presented to you in a conclusion statement.
• Statements with premises & conclusion
o Major premise
o Minor premise
o Conclusion
o Sometimes, these syllogisms can be valid or invalid. This difference will depend on whether these premises
logically predict a given outcome.
o Valid syllogism – where the conclusion statement follow the premise statement
Ex. All A are B, all C are B so then all C are A
All coffee is a beverage
All lattes are coffee
Therefore lattes are beverages
Universal affirment statements, easy for us to reason with.
o Invalid syllogisms – conclusion statement given is not correct based on the premises that you have in terms of
the logic.
Ex. all A are B, all C are B & all C are A
Coffee is a beverage
Milkshake is a beverage
Milkshakes are coffee.
• It is more difficult to reason with ‘particular’ statements: the statements that have “some” instead of “all”. The
ones that have “all” are universal statements. The particular statements (some) are open to interpretations, which is
hard for us to decide which one to take.
• It’s difficult to reason with negative statements in the concluding statement – we have trouble with it bc these
statements are difficult for us to imagine. Its hard to have negative premises.
• Syllogistic reasoning – some of the situations that make us difficult to reason with.
• There are situations, or forms of syllogistic reasoning that aren’t as difficult for us. This shows us that people
don’t just use their logical pattern of them to make sense of them. When we reason logically with syllogisms,
often we make inferences based on what info is contained in the premises or the conclusion that we have in this
syllogisms.
• Prior knowledge & beliefs will inform how we deduce logic from syllogism. It leads to :
• Believability effect – if a syllogism conclusion is believable, people will accept it as true despite logic, even if the
premises don’t logically support conclusion.
Ex. 1. No millionaires are hard workers (no A are B) Invalid if think with logic
Some rich people are hard workers ( Some C are B)
Therefore some millionaires are not rich people ( Some A are not C)
2. No addictive things are expensive
Some cigarettes are expensive
Therefore some addictive things are not cigarettes.
People are more likely to judge 2) as valid as compared to 1) .What people believe about addictive items will
dictate how they interpret logic. We bring assumptions of truth when deducing inferences, which will bias our
logical thinking.
• Syllogisms measure deductive inferences – we deduce if some statement logically explains a given conclusion.
Wason’s 4 card task – assess deductive reasoning.
• Look at how we reason/deduce info from conditional statements.
• A general problem that requires conditional reasoning
• Judge if the statement “If P then Q” is true. P is the condition. Shown cards with a P statement on one side & Q on
the other side.
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
Psyc 213 (april 3rd) reasoning & risk how we make judgments & choices. It takes many years for humans to be good at this game: alpha-go was not given explicit instructions or rules for how to play this game. It developed its own strategies in under 40 days (far faster than any human could: a major sign of creativity. Maybe we can use this computer programs to find solutions to ill-defined problems that we have (cures for diseases?) Practice or 10 000 hour (10 years rule) This can be thoughts or behaviours when we do something: we do this bc this doesn"t require a lot of effort or processing. Reasoning complex cognitive construct that involves every other cognitive processes. Humans may be inheritably logical: people with no logical reasoning can still make logical deductions (infants do not have logical reasoning yet) Could only see the the top of the objects.
