Cognitive Science - Homepages | The University of Aberdeen

blabbedharborΤεχνίτη Νοημοσύνη και Ρομποτική

23 Φεβ 2014 (πριν από 3 χρόνια και 5 μήνες)

70 εμφανίσεις

Cognitive Science

Definition:

“the scientific study either of mind or of intelligence”


Essential Questions


What is intelligence?


How is it possible to model it computationally?


Takes ideas from


Psychology


Philosophy


Linguistics


Neuroscience


Artificial Intelligence / Computer Science


Maybe also minor contributions from:


Anthropology, Sociology, Emotion studies,

Animal Cognition, Evolution

Origins of Cognitive Science


Psychology of the early 20
th

century was dominated by “behaviourism”


Everything should be treated as a behaviour


“…purely objective experimental branch of natural science.”


-

John B. Watson


Goal: prediction and control of behaviour


“Introspection forms no essential part of its methods


-

John B. Watson


Should not have to describe things in terms of “hypothetical” internals


Such as the “mind”


“Consciousness” not an appropriate question for scientific inquiry



This changed around 1950s


Partly as a result of investigations in Artificial Intelligence,

partly changing trends


People started talking about


Theories of mind


Internal representations


Computational procedures


Term “Cognitive Science” born in 1973


Came out of AI
-

Christopher Longuet
-
Higgins comment on “Lighthill report”

Cognitive Science


Information Processing


Cognitive Science views the mind as an
information processing

system


This is also called the
computational

view


From this perspective: a human mind’s activity consists of


Receive information


Store information


Retrieve information


Transmit information


Transform information


Example: a musician improvising


Listen to many tunes


Remember them


Find similarities


Come up with rules that say what sounds good together


Use those rules in real
-
time while playing

Understanding Information Processing Systems

1.
We attribute non
-
behavioural properties to the system


We say that it has a
purpose
,
goals

or
desires


We say that it has internal
beliefs

and
knowledge

and
competence


We attribute
meaning

to its external behaviour and internal information


We treat other humans like this all the time, call it
folk psychology

2.
Representation: information in the system can represent real things


For example: symbols could represent objects and relationships


This would allow a clear separation of
what

and
how

o
Alternatively: it could be a messy representation

o
what

and
how

tangled together

3.
It has
procedures
for processing information


We call these procedures
algorithms

in computer speak


Describes how it does what it does


A clear set of steps that need to be followed


Like the recipe for making a cake


Like the instructions for long multiplication

Three Levels in Information Processing Systems

(Marr’s three levels)

What

How

Representation ties together

Physical
Implementation

Procedure/Algorithm



clear set of instructions

(how to process the input

潵瑰t琩

坨慴a楮景牭f瑩t渠楳⁣潭楮朠楮g

坨慴a楮景牭f瑩t渠楳畴u畴u敤e

坨慴a楳⁴i攠e敬e瑩t湳桩瀿

(also explains why it’s important)

Must be physically carried out


Man with paper and pen


Mechanical computer


Modern PC


Human brain (neurons)

Three Levels in Information Processing Systems

What

How

Representation ties together

Physical
Implementation

Procedure/Algorithm



clear set of instructions


What information is coming in?

What information is outputted?

What is the relationship?


Interesting:


Unlike other
sciences we can
study top two levels
independently from
the physical level

Must be physically carried out


Man with paper and pen


Mechanical computer


Modern PC


Human brain (neurons)

Caveat:

This is a particular philosophical
position, called “Functionalism”.
Some philosophers do not accept it.


Functionalism: mental states
(beliefs, desires, being in pain, etc.)
are constituted solely by their
functional role; i.e. their causal
relations to other mental states,
sensory inputs, and behavioural
outputs.


Consequence: a mind can be
implemented in lots of different
physical hardware, so long as it
performs the right functions.

Three Levels in Information Processing Systems

What

How

Representation ties together

Physical
Implementation

Procedure/Algorithm



clear set of instructions


What information is coming in?

What information is outputted?

What is the relationship?


Interesting:


Unlike other
sciences we can
study top two levels
independently from
the physical level

Must be physically carried out


Man with paper and pen


Mechanical computer


Modern PC


Human brain (neurons)

Caveat:

This is a particular philosophical
position, called “Functionalism”.
Some philosophers do not accept it.


Functionalism: mental states
(beliefs, desires, being in pain, etc.)
are constituted solely by their
functional role; i.e. their causal
relations to other mental states,
sensory inputs, and behavioural
outputs.


Consequence: a mind can be
implemented in lots of different
physical hardware, so long as it
performs the right functions.

What’s special about a mind then?

We know it can do things a computer
can’t do…


A Functionalist claims that the special
thing about the mind is the special
information processing
tasks
,
representations

and
algorithms

it uses


One could implement the same functions
in a computer


don’t need organic
neurons

Important to Study All Three Levels

What

How

Physical
Implementation

Procedure/Algorithm



clear set of instructions


What information is coming in?

What information is outputted?

What is the relationship?


Could have elegant
mathematical theory
which no algorithm
can implement

Must be physically carried out


Man with paper and pen


Mechanical computer


Modern PC


Human brain (neurons)

Important to Study All Three Levels

What

How

Physical
Implementation

Procedure/Algorithm



clear set of instructions


What information is coming in?

What information is outputted?

What is the relationship?


But without top
level…

Lose sight of
what

your information
processing is trying
to achieve

Must be physically carried out


Man with paper and pen


Mechanical computer


Modern PC


Human brain (neurons)

Important to Study All Three Levels

What

How

Physical
Implementation

Procedure/Algorithm



clear set of instructions


What information is coming in?

What information is outputted?

What is the relationship?


Could have a nice
algorithm, but
might take too
much physical
hardware to be
practical

Must be physically carried out


Man with paper and pen


Mechanical computer


Modern PC


Human brain (neurons)

Important to Study All Three Levels

What

How

Physical
Implementation

Procedure/Algorithm



clear set of instructions


What information is coming in?

What information is outputted?

What is the relationship?


Focussing on the
physical
interactions here
gives you no idea
of what their
purpose is

Must be physically carried out


Man with paper and pen


Mechanical computer


Modern PC


Human brain (neurons)

Important to Study All Three Levels

What

How

Physical
Implementation

Procedure/Algorithm



clear set of instructions


What information is coming in?

What information is outputted?

What is the relationship?


Insights from
studying the brain
could give clues
about the
algorithms and
representations
which are

(or are not)

being used

Must be physically carried out


Man with paper and pen


Mechanical computer


Modern PC


Human brain (neurons)

Another Perspective on Cognitive Science


Studying different information processing tasks at different levels

Vision

Language

Memory

Problem
Solving

Learning

What

(Info Proc Task)

How

(Algorithm)

Physical
Implementation

AI and Cognitive Science


Two way interaction between AI and Cognitive Science



AI informs Cognitive Science


Common to implement a cognitive theory in a computer


Run the program and see the ramifications of the theory


(Scientific hypothesis testing)


Running it may be necessary because theory is complicated



Also, existing AI theories may shed light on the way humans do it



Cognitive Science informs AI


Seeking inspiration to solve an AI problem


Study the way humans do it


Copy in computer


…or at least constrain the possible options under consideration


Herbert Simon

“AI can have two purposes. One is to use the power of
computers to augment human thinking,

just as we use motors to augment human or horse power.

Robotics and expert systems are major branches of that.

The other is to use a computer's artificial intelligence to
understand how humans think. In a humanoid way.

If you test your programs not merely by what they can
accomplish, but how they accomplish it, then you're really
doing cognitive science;

you're using AI to understand the human mind.”

Applications of Cognitive Science


Education and Learning


From Cognitive Psychology:


Diagnose and treat children’s reading difficulties


Stroke Therapy


From Linguistics:


Understanding of speech impairments when stroke in left
hemisphere of brain


… better therapy


Legal process


From Cognitive Psychology:


Understanding of reliability of memory


Question reliability of legal witnesses


Computing Technology


From AI:


You know loads of examples by now

Cognitive Science


Different Methods


Psychology


Controlled laboratory experiments


Detailed observations of behaviour


Philosophy


Thought experiments


Investigate consequences, and coherence of theories


Linguistics


Test speakers’ intuitions about “grammatical” sentences


Analyse children’s acquisition and errors


Neuroscience


Study active brain regions when doing something


Study neurons


Artificial Intelligence / Computer Science


Write programs, see where they succeed and fail

Cognitive Psychology


What are the mental processes in between stimulus and response?

Categorisation

Attention


Memory


Knowledge representation

Numerical cognition

Thinking

Learning


Language

Sight

Hearing

Taste

Smell

Touch

Balance

Heat/cold



Voice

Limbs

Fingers

Head



sensory
input

Motor
output

Sensory
systems

Central
systems

Motor
systems

(rough model
-

Boundaries are not clear in reality)

Cognitive Science


Different Methods

Focus on central unit…


Thinking


Draw conclusions from facts, solve problems, plan actions…


In many diverse domains


Attention


Helps us focus on some task


Has limited capacity


Memory (includes Knowledge Representation)


Seems to be huge


Seems to be no limit on how well it retrieves relevant information


Learning


Acquire new knowledge and sensorimotor skills



How does this central unit work?

Physical Symbol System Hypothesis

“A physical symbol system has the necessary and
sufficient means of general intelligent action.”









therefore…


human thinking = symbol manipulation

Newell & Simon,

1963.

Physical Symbol System Hypothesis

“A physical symbol system has the necessary and
sufficient means of general intelligent action.”


Their symbols are taken to mean
high level symbols


Directly correspond to objects in the world,


such as “monkey” and “table”.


…but the weights and connections in a neural network
could also be represented as symbols


Use this to make a “scruffy” representation of “monkey”


but that’s not considered to be what they meant

Physical Symbol System Hypothesis

“A physical symbol system has the necessary and
sufficient means of general intelligent action.”



Most AI people nowadays would not accept the idea of

high level symbols
being sufficient



Seems to work well for


playing chess, problem solving (if problem well defined)


but doesn’t work so well for some “easy” problems


Vision, moving around in the world



But most AI people would accept the computational
theory of mind (i.e. Functionalism)

Universal Computing Machine





Turing machine:


Actions:


Head can move left and right over the tape


Can read and write symbols on the tape


Can overwrite symbols on tape


Machine has an internal state


Takes Action depending on state


Turing’s thesis: “If an algorithm exists then there is an equivalent Turing
Machine”


Turing machine is the simplest possible description of a computer that
can do anything


All modern computers can be simulated by a Turing machine


Only real difference:

Turing machine has
infinite

tape, real computers have finite memory

Universal Computing Machine


How many symbols and

states do you need?









Interesting…


If you make some really fancy machine…


Loads of states


Loads of possible symbols


Multiple tapes


Multiple stacks for storing things


Many heads working in parallel


You end up with something equivalent to the Turing machine

States

Symbols

24

2

10

3

7

4

5

5

2

5

4

6

3

10

2

18

Universal Computing Machine


The Turing machine has a set of rules


These determine how it acts


Can make a Universal Turing machine


Encode the rules you want it to use on the tape


The first thing it does is to read the rules


Then follow them…


Could also reprogram its rules as it goes along


Important ability for learning


Behaviour must change given experience

Universal Computing Machine


We said


“If an algorithm exists then there is an equivalent Turing Machine”


i.e. a (different) Turing machine is available to do any job we want to do


Now we can say


“If an algorithm exists then it can be simulated on a Universal Turing Machine”


i.e. all we need is a single Universal Turing Machine


This can do anything


This is the idea behind modern computers


Program instructions stored in memory just like any other data


Download a program off the web, and start running it


You don’t need a different computer for different jobs


One computer can do everything


Games, spreadsheet, database, music, movies, photo editor, word processor…

Is the Brain a Universal Computing Machine?


Warren McCullogh and Walter Pitts showed


Small collections of neurons can act as “logic gates”

(building blocks of computers)


Brain could be viewed as a computing device, just like Turing machine


i.e. a brain can do what a computer can do



Other direction is a stronger claim


Can a computer do what a brain can do?


Can’t be proved


But universality of Turing machine suggests… maybe