Varieties of Learning - PowerPoint PPT Presentation

Loading...

PPT – Varieties of Learning PowerPoint presentation | free to download - id: 6d3de9-ZGQyM



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Varieties of Learning

Description:

Varieties of Learning Structural descriptions and instances Scenarios and locations; eating in a fast food restaurant Perceptual and semantic representations – PowerPoint PPT presentation

Number of Views:6
Avg rating:3.0/5.0
Slides: 32
Provided by: Comp6152
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Varieties of Learning


1
Varieties of Learning
  • Structural descriptions and instances
  • Scenarios and locations eating in a fast food
    restaurant
  • Perceptual and semantic representations
  • What McDonalds looks like and why you go there
    sentences and their meanings
  • Wholes and Parts
  • What McDonalds looks like and what the golden
    arches looks like.

2
Part-Whole Recognition
  • Procedure
  • 48 study items presented
  • whistle-BALL
  • recognition tests for individual words
  • RACKET NET BALL
  • recognition test for pair
  • whistle-BALL pretty-WET cave-BLUE
  • Results
  • Recognition of single words independent of pair
  • Conclusion
  • More than one representation of study item may be
    independently encoded in memory

3
Varieties of Learning
  • Structural descriptions and instances
  • Scenarios and locations eating in a fast food
    restaurant
  • Perceptual and semantic representations
  • What McDonalds looks like and why you go there
    sentences and their meanings
  • Wholes and Parts
  • What McDonalds looks like and what the golden
    arches look like.

4
Categorization
5
Categorization
  • Definition
  • The Role of Perception in Categorization
  • The Role of Semantics in Categorization

6
Definitions
  • When an infant kicks to move a mobile we may call
    the mobile
  • The target of the kicking response
  • The stimulus of the kicking response
  • The cue for the kicking response
  • The moving mobile is the consequence of the act
    of kicking.

7
Generalization
  • When the same action has produced the same
    consequence in response to more than one target,
    the probability of a novel object activating a
    representation of the action and its consequence
    is a function of its similarity to all previous
    targets of the action.
  • An infant who has kicked to move two different
    mobiles will kick to move a third mobile that is
    a composite of the previous two.

8
Learning a Generalization
  • The more targets of an action, the more likely
    that a novel input will match the target set
    sufficiently to elicit the action.
  • Hence, when an action has produced the same
    consequence for more than one target, the result
    is a bootstrap effect between the number of
    targets associated with the action and the
    probability of a new target eliciting the action.
  • The greater the variety of targets to which an
    action has been successfully directed (i.e.,
    produced the same desired response), the greater
    the variety of targets that will elicit the
    action.
  • Once an infant learns that kicking can move two
    different mobiles, the infant will kick to any
    mobile.

9
Categorization
  • When representations of different items become
    associated with the same consequence of the same
    action, the items are said to be instances of the
    same category.
  • In explicit categorization, in addition to other
    possibly actions uniting the instances, the
    action of naming produces the same result, the
    verbal category label, for the instances.
  • All colors are called colors.

10
Categorization
  • Definition
  • Category instances are different targets of the
    same action.
  • The Role of Perception in Categorization
  • The Role of Semantics in Categorization

11
Categorization
  • Definition
  • The Role of Perception in Categorization
  • Natural Categories
  • Artificial Categories
  • The Role of Semantics in Categorization

12
Categorization
  • Definition
  • The Role of Perception in Categorization
  • Natural Categories
  • When instances automatically match features in
    each others representations as the result of
    perceptual organization
  • Hence, these categories are determined by how the
    perceptual system organizes the inputs.
  • For example, color and shape categories (Heider,
    1972)
  • Artificial Categories
  • The Role of Semantics in Categorization

13
Categorization
  • Definition
  • The Role of Perception in Categorization
  • Natural Categories
  • When instances automatically match features in
    each others representations as the result of
    perceptual organization
  • As a result of perceptual organization, some
    features are more salient than others.
  • Salient features make categorization easy so that
    categories may be formed by simple observation,
    even without instructions to do so (Fried and
    Holyoak, 1984).
  • Examples Color categories, chair, table,
    bird.
  • Artificial Categories
  • The Role of Semantics in Categorization

14
Natural Rectangle Categories
  • Higher than wider versus wider than higher.
  • May be sorted without feedback.

15
Salient Features
  • The salient features of a representation are
    those features that are weighted most heavily by
    the perceptual system in determining the
    similarity between two representations.

16
Nose is more salient feature than ears so it
determines category
17
Categorization
  • Definition
  • The Role of Perception in Categorization
  • Natural Categories
  • Basic level categories are natural categories
    that exist in the perceptual system and the
    world.
  • Examples chair, table, bird.
  • Artificial Categories
  • The Role of Semantics in Categorization

18
Basic Level Categories
  • Are broadest category at which members have many
    perceptual features in common.
  • Have similar shapes.
  • Are among the first categories learned.

19
Picture Naming of Basic Categories
  • Pictures of objects are named by the basic
    category term 99 of the time.
  • However, experts are different.
  • People with expertise (detailed knowledge of
    differences among category members) use the
    subordinate term to name pictures of objects in
    their area of expertise.
  • Same as others for categories not in area of
    expertise.

20
Categorization
  • Definition
  • The Role of Perception in Categorization
  • Natural Categories
  • Artificial Categories
  • The instances of a category do not automatically
    match features in each others representations as
    the result of perceptual organization.
  • Hence they are not formed just by observation.
    Rules, descriptions, or feedback must be given to
    learner for category concept to be formed.
  • Nevertheless, perceptual features are used to
    recognize instances of artificial categories
  • The Role of Semantics in Categorization

21
An Artificial Category
  • Size greater than tilt versus tilt greater than
    size.
  • Requires feedback for learning.
  • Rule may not be capable of articulation by
    learner.

22
Categorization
  • Definition
  • The Role of Perception in Categorization
  • The Role of Semantics in Categorization
  • In addition to perceptual matching, an observer
    extracts and labels perceptual features for a
    technical definition containing defining features
  • A birds lay eggs and all birds and only birds
    have feathers

23
Categorization
  • Definition
  • The Role of Perception in Categorization
  • The Role of Semantics in Categorization
  • Verbal labeling of perceptual features makes
    category definitions possible.
  • Once sufficient language is learned, verbal
    definition plays a role in category learning

24
Gelman Markman (1986) study
  • 3-4 years of age
  • Bird (flamingo) gives baby mashed food Bat gives
    milk.
  • What does (black)bird do?
  • 68 say mashed food

25
Categorization
  • Definition
  • The Role of Perception in Categorization
  • The Role of Semantics in Categorization
  • Verbal labeling of perceptual features makes
    category definitions possible.
  • Once sufficient language is learned, verbal
    definition plays a role in category learning
  • Categories may also be defined by nonperceptual
    features.

26
Types of Categories
  • Perceptual.
  • Most objects (rocks, animals, etc.)
  • Functional.
  • Tools, furniture, weapon, occupational, etc.
  • Kinship.
  • Mother, uncle, etc.
  • Abstract.
  • Justice, etc.
  • Categories defined by enumeration.
  • 26 letters of the alphabet.

27
Categorization
  • The Role of Semantics in Categorization
  • Verbal labeling of perceptual features makes
    category definitions possible.
  • Once sufficient language is learned, verbal
    definition plays a role in category learning
  • Categories may also be defined by nonperceptual
    features.
  • Categories have internal structure determined by
    the similarity among instances.
  • Knowledge of a category includes
  • Knowledge of typical instances
  • Knowledge of atypical instances
  • Frequency of occurrence.

28
Typical and Atypical Members
  • There is an asymmetry in comparing the similarity
    of typical and atypical instances
  • ___ is virtually a ___. Try pink versus red.
  • Definition of typicality for multi-feature
    instance representations
  • Instance typicality is a function of the number
    of other instance representations it is similar
    to.
  • Similarity may be determined by both category
    relevant and category irrelevant features.
  • Typical instances also called focal and
    prototypical instances.

29
Categorization
  • The Role of Semantics in Categorization
  • Verbal labeling of perceptual features makes
    category definitions possible.
  • Once sufficient language is learned, verbal
    definition plays a role in category learning
  • Categories may also be defined by nonperceptual
    features.
  • Categories have internal structure determined by
    the similarity among instances.
  • Categories are hierarchically organized and share
    instances

30
Verbal categorization makes hierarchical
organization possible
  • Category labels may be treated as instances and
    associated with more general category labels.
  • Dogs, cats, etc. are all animals
  • All mature learning depends on prior knowledge.
  • Novel items are associated with existing category
    labels on the basis of similarity to familiar
    instances. These are the semantic associations.

31
Semantic Categorization and Learning
  • Semantic categorization is the basis of
    declarative memory
  • As a result, additional representations are
    activated when a perceptual representation is
    constructed.
  • Semantic categorization also facilitates
    procedural memory
  • Category nodes in structural descriptions make it
    possible to automatically construct an infinite
    number of perceptual representations
About PowerShow.com