Lecture 16/17 (Pattern Recognition)

Question 1

Pattern Recognition

Answer 1

The ability to identify regular consistencies in items we have encountered before

Question 2

Template

Answer 2

An outline or layout of something

Question 3

Template Matching

Answer 3

Identify and match external stimulus to template

Question 4

Feature

Answer 4

A part of a stimulus that makes a significant contribution to its overall appearance or form

Question 5

Image Demon

Answer 5

The stimulus in the environment provides an image on the retina and in the mind

Question 6

Feature Demon

Answer 6

Detects the feature in the stimulus

Question 7

Cognitive Demon

Answer 7

A cognitive demon for every letter which “shouts
louder” for every feature that is present in the
image

Question 8

Decision Demon

Answer 8

Decides the identity of the stimulus based on
which cognitive demon shouts the loudest

Question 9

FEATURE DETECTION THEORY FOR LETTERS

Answer 9

Letter in the environment is captured on the
retina
* Detects features on the retinal image
* Detected features are compared with features
from previously encountered letters
* Identity of letters based on comparison of
detected features on retinal image with
previously encountered letters

Question 10

DATA-DRIVEN PROCESS

Answer 10

A stimulus is identified based on the observed
features of the stimulus
* Also known as bottom-up processing

Question 11

CONCEPTUALLY-DRIVEN PROCESSES

Answer 11

A stimulus is recognized and identified based
on its semantic characteristics (number)
* Also known as top-down processing

Question 12

Geons

Answer 12

“Geometrical Ions”
Objects with a basic volumetric shape

Question 13

Recognition by Components

Answer 13

Three-Dimensional objects can be broken
down into basic three-dimensional shapes
(geons)
Geons can be assembled to form an unlimited
number of object

Question 14

Simple Cells

Answer 14

Cells in the visual cortex that receive signals when
they detect a line of specific orientation

Question 15

Complex Cells

Answer 15

Cells in the visual cortex that receive signals when
they detect a line of specific orientation moving in
a specific direction

Question 16

Hypercomplex Cells

Answer 16

Cells in the visual cortex that receive signals when
they detect two lines, each line of a specific
orientation meeting at a specific angle moving in
a specific direction is detected in the environment

Question 17

Hebb’s Rule

Answer 17

Neurons that fire together wire together

Question 18

Cell Assembly

Answer 18

A group of cells that come to be grouped together
for a common purpose (some type of cognitive
process/activity)
* A consequence of Hebb’s Law

Question 19

Input Layer

Answer 19

The layer of nodes/artificial neurons that receive information about the stimulus

Question 20

Output Layer

Answer 20

The layer of nodes/artificial neurons that
produces information about identity of the
stimulus

Question 21

Unsupervised Learning

Answer 21

Researcher does not observe and direct the
execution of the task.
- No feedback provided
- Inputs are changed based only on features of
detected stimulus

Question 22

Feedback

Answer 22

Information about the neural network’s
performance to a task
* In this context feedback about whether the
network’s response is correct or incorrect

Question 23

Supervised Learning

Answer 23

Researcher directs the execution of the task.
- Researcher provides feedback

Question 24

Hidden Layer

Answer 24

• Allows links from input layer to the hidden layer
• Allows links from hidden layer to the output layer
• Links modelled by inputs and weights
• Feedback can be treated as Input provided to network from output layer
• Network can modify the values of weights based on feedback