CS 182Sections 101 - 104Created by Eva MokModified by JGM 2/2/05

Q: What did the hippocampus say during itsretirement speech? A: “Thanks for the memories”

Q: What happens when a neurotransmitter falls in love with a receptor?

A: You get a binding relationship.

Q: What did the Hollywood film director say after he finished making a movie about myelin?

A: “That’s a wrap!”

http://faculty.washington.edu/chudler/jokes.html

Announcements

• a2 is out, due next Monday 11:59pm

– play with tlearn

– you can either run it on inst machines or download it and run on your pc (though this may give you some headaches…)

• Quiz on Thursday

Where we stand

• Last Week

– Basic idea of learning, Hebb’s rule

– Psycholinguistics experiments

• This Week

– Spreading Activation, triangle nodes

– Connectionist representations

• Coming up

– Backprop (review your Calculus!)

Quiz!

• What are does the Stroop effect show? What was the point of the eye-tracking experiment?

• Why is Hebb’s rule not the complete story for the learning that goes on in the brain?

• What’s a McCullough-Pitts neuron? How does it work?

• What does the “They all rose” experiment show? How can you explain the results computationally?

Declarative Non-Declarative

Episodic Semantic Procedural

Memory

Two ways of looking at memory:

facts about a situation

general facts skills

Stroop effect

• takes longer to say what color a word is printed in if it names a different color

• suggests interaction of form and meaning (as opposed to an encapsulated ‘language module’)

‘Word superiority effect’

• it’s easier to remember letters if they are seen in the context of a word

• militates against ‘bottom-up’ model, where word recognition is built up from letters

• suggestion: there are top-down and bottom-up processes which interact

Eye-tracking Experiment

• Three hypothesis for eye-tracking results:

– Cohort theory

– Neighborhood activation model

– TRACE (McClelland & Elman)

Memory

Short Term Memory Long Term Memory

Two ways of looking at memory:

electrical changes

structural changes

LTP

A X PT

G Q N L

W R V S

• Hebb’s Rule: neurons that fire together wire together

• Long Term Potentiation (LTP) is the biological basis of Hebb’s Rule

• Calcium channels is the key mechanism

LTP and Hebb’s Rule

strengthen

weaken

Why is Hebb’s rule incomplete?

• here’s a contrived example:

• should you “punish” all the connections?

tastebud tastes rotten eats food gets sick

drinks water

The McCullough-Pitts Neuron

yj: output from unit j

Wij: weight on connection from j to i

xi: weighted sum of input to unit i

xi f

yj wij

yi

xi = ∑j wij yj

yi = f(xi)

ti : target

Let’s try an example: the OR function

• Assume you have a threshold function centered at the origin

• What should you set w01, w02 and w0b to be so that you can get the right answers for y0?

i1 i2 y0

0 0 0

0 1 1

1 0 1

1 1 1

x0 f

i1 w01

y0i2

b=1

w02

w0b

Many answers would work

y = f (w01i1 + w02i2 + w0bb)

recall the threshold function

the separation happens when w01i1 + w02i2 + w0bb = 0

move things around and you get

i2 = - (w01/w02)i1 - (w0bb/w02)

i2

i1

“They all rose”

triangle nodes:

when two of the neurons fire, the third also fires

model of spreading activation

How we can model the triangle node with McCullough-Pitts Neurons?

B C

A

A B C

Anonymous Feedback: Lectures

feel free to comment on each instructor seperately

How is the pace?

Do you find the material interesting? too dense? too slow?

What will be most helpful to you in getting the most out of lectures?

Any particularly confusing topic?

Anonymous Feedback: Sections

Have sections been useful?

Any feedback on our styles of presentation?

How will sections be most helpful to you?

Any other comments?