Week 1 Doing Psych Flashcards
2 main groups of methods
quan and qualitative
features of quan methods
development + test of specific theories
theories –> mathematical predictions can be tested by collecting data + stat analysis
features of qualitative
dev of verbal theories
open-ended/explanatory
what is a theory + specifications
principles that explain a system
specifies relations
explain existing + predict new data
what is NOT a theory
description
set of data
diagram
3 levels of analysis theory of mind
computational
representation and algorithm
physical implementation
2 main model types in psych
mathematical
process: symbolic, connectionist
computational theory
- what problem is it solving (and why)?
- constraints on its solution?
- nature of the problem / the function being computed?
representation and algorithm
- what info does the system represent + how
- what it does with the info
- what algorithm is used to extract useful info
- what is the input to the system, what is its output + stages in between?
physical implementation
how are representations and algorithms realized in the hardware of the device itself
(e.g., in the neurons of the brain, the silicon of the computer, etc.)
what is a symbolic process model
represents knowledge as symbolic data structures
manipulate data with variable-ised rules
what is a connectionist process model
knowledge as nodes in a network
processing carried out by passing activation between nodes
what are representations of symbolic models
basic - atomic elements
rules - composing complex structure i.e. a language
what are processes of symbolic models
operations on data structures
applications of symbolic rules
what are production systems + how many components
prototypical symbolic model
3
3 components of production systems
base of known facts
set of inference rules
executive control structure
5 stages involved in operation of production system
current state - known facts
state space - all possible states
goal state - state want the data base to be in
state transition - moving from one state to another
search - algorithm for traversing state space
3 advantages of symbolic models
computational power
can define variablized rules
represent = reason
5 disadvantages of symbolic models
- rules too rigid for human behaviour
- new representations = combos of existing ones
- how are rules learned
- no graceful degradation with damage
- no obvious neural implementation
what are connectionist models
models composed of networks of interconnected nodes
what are nodes
simple processors that mimic neurons / populations of neurons
what are connections in a connectionist model
weights between nodes
what are representations in a connectionist model
patterns of activation on nodes
what occurs during the processing of connectionist models
nodes pass activation over weighted connections
pos weights are excitatory connections
advantages of connectionist models
- flexible processing (parallel contraints)
- flexible representations: distributed reps capture semantic content + permit auto generalization
- graceful deg with damage
- transparent neural implementation
what are the disadvantages of connectionist models
- not symbolic
- ability to generalize depends on similarity of examples
- cannot represent/use variable-ised rules
- even children generalize to utterly dissimilar examples
7 different vulnerable group categories
marginalized
socially excluded
limited opps
suffer abuse
hardship
prejudice
discrim
8 vulnerable groups defined by SVG ACT 2006
lone parents
disabled
elderly
children
ethnic min
mentally ill
homeless
refugees
