05.26_Knowledge3_updated

05.26_Knowledge3_updated - Knowledge Representation III...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
© S. J. Luck All rights reserved 1 © S. J. Luck All Rights Reserved Knowledge Representation III • Over the past 30 years, psychologists have developed models of cognition that are based on “neural networks” – Also called “parallel distributed processing” models – Also called “connectionist” models • Rather than using a computer metaphor for the mind, these models use a brain metaphor – Processing is conducted by a large number of “stupid” processing units (like neurons) – The units operate in parallel (in contrast, a typical computer does one step at a time) – Both procedures and knowledge are based on the pattern of connections between units Sara Smartz: 1x4 + 2x5 + -1x2 = 12 Fred Flunck: 1x2 + 2x0 + -1x5 = -3 A) Do you eat breakfast every morning? 0 1 2 3 4 5 B) Do you read the text before lecture? 0 1 2 3 4 5 C) Do you cram the night before a test? 0 1 2 3 4 5 Never Always How good of a student are you? Take our Scholarly Index test. To ±nd your Scholarly Index, add A to two times B and subtract C: SI = 1xA + 2xB + -1xC Weights -1 +2 +1 A B C R Input Units Weights Output Unit A unit is like a neuron The activation level of a unit is like a neuron’s electrical potential The output of a unit is like a neuron’s ±ring rate The weights correspond to the “strength” of a synapse A B C D Input Layer Hidden Layer E F Output Layer .5 .3 -.2 .7 .2 -.1 .6 .2 Input layer receives inputs (like sensory receptors) Hidden layer(s) form intermediate representations (like association cortex) Output layer provides outputs (like spinal motor neurons) Backward and lateral connections are also possible • Like the human brain, neural networks can learn from experience • Training a neural network: – Start with connection weights randomly assigned – Network is given an input and produces an
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 11/08/2010 for the course PSC PSC100 taught by Professor Luck during the Spring '10 term at UC Davis.

Page1 / 3

05.26_Knowledge3_updated - Knowledge Representation III...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online