Tuesday, June 9, 2015

Transcript - Spoil Sports of Prediction (concluded)

Class: Informatics, Computing, and the Future
Instructor: Dan Berleant
Transcriber: Brooke Yu
Date: Thursday, February 28, 2013

Professor:  Okay, so let me start by handing out the next homework, which of course, is on the web.  I'll give you hard copies so you'll have it.  You can look at it now.  We're going to just take a look at it to see what it looks like.  Alright, any questions on the next homework? 

Let me turn this off.  So, it's a fairly straightforwad.  You find some videos to critique and work on your project.  I recommend you think of it in terms of the spoil sports of the prediction game.  That could add something to your thinking on the topic. 

Any questions about the homework or anything else? 

Too much homework?  Not in this class, right? 

Okay.  Spring break is coming up pretty soon. 

So we talked about several spoil sports as they call them.  We talk about the observer effete, the uncertainty principle, the butterfly effect... that leaves three more. 

These are a little less sort of physics and more of philosophy. 

I call them external perturbations to the system.  We have something called existentialist angst, and the care horizon. 

We're way down here. 

Male Student:  So as far as the current project that was due Tuesday of this week, do you want us to apply this to it?  Do we submit it separately on our blogs? 

Professor:  Yeah, you'll think about your project in a certain way.  By the end of the semester you'll have enough homeworks to glue together and a lot of it will be done. 

Male Student:  I understand the point.  I just hope to get it all up to date this weekend. 

Professor:  Alright, so.... okay, so external perturbations.  So what do you think?  What's an external perturbation? 

I mean, if you've got... if you're playing pool again and the table is tilted, that's something external to the game of pool.  It's an external change that affects the way the system works. 

The balls all kind of tend to sort of swerve in one direction if you tilt the table, so it's external.  A perturbation is just a push or an influence. 

So in essence... you want to think about.  Well, when you're trying to predict the future, you want to know what happens next, but you need to know where things are now.  You know the configuration of the pool balls and how fast they're going.  That would help you predict what's going to happen in the next few seconds. 

Same the for meteorologists- they collect all of this data and feed it to the computer, and the computer figures out what is supposed to happen tomorrow.  It works pretty well, and that's how they predict the weather these days.  They figure out where things are now and just crunch it all into a big computer to see what will happen next. 

But you have to know what outside influences will affect the system.  What's an outside influence on the weather?  Any ideas? 

Okay, thunderstorms are a part of the system, and hopefully the system will be predictable then.  Oh, did you say solar storms? 

Oh, yeah, you're right.  When the sun has some glitch in its energy output, that is an external stimulus.  The simulations they do to predict the weather have to assume the sun will be a steady influence, and sometimes it has little glitches and it spews out streams of particles which hit the atmosphere and cause the magnetic field to go haywire. 

So solar storms.  Any other ideas for other examples of external influences to the weather? 

Well, here's another one.  I keep saying how the national weather service has a very extensive collection of sensors all that they can simulate on a computer.  But, you know, weather in china affects weather in the US too.  Weather conditions outside the borders of the US flow into the US.  The US national weather services don't have sensors blanketing the world. 

Actually, they may be using Canadian information, but I'm not sure about the extent of that.  Or Mexico won't have a set of sensors. 

So from that standpoint, the weather outside the US is an external influence that affects the weather in the US that the simulators can't predict. 

Okay.  So those influences affect our ability to predict.  That's why it's a spoil sport of predictions because it can make predictions useless. 

Let's take pool again.  You know the position and velocity of every ball, and you have a computer where you can plug in all the data.  Probably soon they'll have cameras that can collect this data.  That's probably possible now because we have some pretty powerful computers. 

Okay, so you know, a fast computer can figure out what's going to happen on a pool table, but it can only go several seconds into the future because issues cause problems- the observer effect... what are the others? 

These "Issues" are other spoil sports we talked about.  Any others? 

Butterfly effect.  When you're talking about pool- it's really a game that's from a physicists standpoint- it has perfectly round balls, and everything works in an idealized way.  It practically is mathematical.  If you have a ball here and another ball hits here, it'll bounce one way, but if it hits here it'll bounce significantly different.  So the point makes a big difference.  That tiny change in this one makes major changes in the way that one works- that's the butterfly effect in essence. 

So there's the observer effect, the uncertainty principle, the butterfly effect.  You don't know how fast things are going- well, with ool balls it's pretty accurate.  It might actually end up on the other side by quantum tunneling.  Did we get all four?  Yeah. 

Even for a human to know what's going to happen in a few seconds- you still can't figure out what the table's going to do more than a few bounces, right?  Even the world's best player can't. 

There are some other issues- these of the external perturbation issues.  For example, a draft hits the table and pushes the ball a little bit.  This one makes a bigger change, and at the end the outcome is completely different. 

Another possibility is if the table is slightly tilted.  After a couple bounces, that will make a difference.  Or maybe someone bumps the table.  A pool table is pretty heavy and you don't notice when it gets bumped, but it does make a little difference. 

So these influences can affect the evolution of the system- that's why they're called influences.  And they're external to the system. 

Another example- remember the water wheel.  Who remembers? 

Female Student:  You never know which way it's going to go. 

Professor:  Okay, how does it work? 

Female Student:  It's like buckets and they pour into each other. 

Professor:  Uh, there is water that is poured into them, so I guess the hose of one could get water into the next.  But water is pouring in and out and the weights of the buckets change and that causes the wheel to slow down and speed up and reverse direction in an unpredictable way because the tiny changes in water flow create the butterfly affect.  The wheel was named after Lorenz, the discoverer of the butterfly effect, who was studying the weather. 

So you have this water wheel and it does crazy things.  What if it's raining?  Okay?  Now you've got external water droplets that are not part of the system that are hitting the buckets randomly.  That will change things, and that makes it even more impossible and unpredictable because every rain drop is an external perturbation. 

[Teacher reading: [On board.]  

I don't know how much of a difference it would make in the water wheel- less than a minute would be my guess. 

That's just a guess, it might be a lot less than a minute. 

Male Student:  So it's contribution to the ecosystem, I guess- in a sense

Professor:  Yeah, you can think of the wheel as an ecosystem and the rain is an external influence.  You could simulate a real ecosystem and they'll model the rainflow. 

Male Student:  Yeah, especially the sulfur rain when it hits trees. 

Professor:  Yeah, so they try to say "well, what will happen with that external influence and they try to model It."  Okay anybody remember what the Lorenz water wheels were made of in most cases?  Everyday objects?  Yeah, bicycle wheels and ordinary buckets.  It's not high-tech. 

So, you know, every external nudge is like a butterfly wing flap, so the butterfly effect becomes a spoil sport.  

But where do those butterfly wing flaps come from?  They're external influences.  They're not a part of the system- rain drops, butterflies flapping their wings.  The US national weather service- when it measures air speed, it doesn't take into account the butterfly, although perhaps if one flew near a sensor it might register. 

Okay, so... a lot of you have a car.  How many drove to class today?  So what sort of external influences might affect the future of your car? 

Okay, well, price of gas will affect your ability to drive your car.  But anything that affects the actual car itself? 

Male Student:  The transmission going out? 

Professor:  Okay, here's the thing about transmissions.  I was reading that some people don't like to change their transmission fluid because they think it'll mess it up.  So the question is- I don't know.... I couldn't find any clear statement about it, but people think it's risky because they think it'll cause their car to break.  I did some research, the problem is not changing the fluid, but the flush.  They try to clean out the inside of the transmission by forcing fluid through it and that just dislodges dirt which gets caught in it.  So if you want to change your transmission fluid, you should probably drain it and replace it but don't try to flush it or blast out the dirt.  Let the dirt stay where it is. 

So those bits of dirt that might get dislodged could make a difference between one that lasts longer and one that breaks sooner.  So the flushing would be an external influence.  It makes a difference between one that keeps going and one that doesn't

Another thing that might affect the future of your car- your run over a hole and you mess up the alignment of the wheels and the tires wear faster.  How about an accident?  That will make your car break a lot sooner. 

So the car has sort of a natural process of getting older and wearing out, but external changes and influences can make a big difference. 

How about you?  Your future?  Can you think of an external influence that might.....not necessarily you, but.

Well, you pick up a cold or flu bug and that might cause you to miss classes and you blow a quiz or something. 

In the germs would be the external influence and what would have happened was you would have taken the quiz and been fine. 

I think it's, you know if we're thinking about how an external- an unpredictable external influence- it does affect how people's lives go.  Other people too, they're also unpredictable and can help you or mess you up. 

Here's another example [On board.]   

We're getting back to weather forecasting again, and Lorenz discovered the effect because the computer he was using to simulate a weather system was getting weird answers, and he discovered it was because the computer had roundoff error.  Computers have arithmetic roundoff error.  How many people knew that? 

You know, numbers have limited accuracy when stored on a computer, right?  However many digits it is- after 14 digits or so the computer won't know what it is.  The computer can't tell the difference between a 20 digit number that ends in 1 and a 20 digit number that ends in 2.  It'll chop off the last few numbers. 

You can try that in a spreadsheet.  So when you're simulating a system and you've got numbers in the computer doing calculations, the computer is giving wrong answers by a tiny bit because it's rounding off.  How many digits are in the answer if you divide 1 by 3? 

Male Student: it's somewhere around .....

Professor:  Right, what do you think?

Female Student:  It's a repeating decimal. 

Professor:  Well, the computer will chop off after a dozen numbers or so. 

Male Student:  In 2011 or so they had a man in Japan use an old computer to calculate pi and it was like a trillion numbers.  It took 20 gigs of space to keep the measurement of the ending of pi itself. 

Professor:  Actually, pi never ends.  It goes on forever.  I mean, there are people calculating additional digits of pi with algorithms, but it just keeps going. 

Let's try it.  I'll prove it.  Let's bring up a spreadsheet.  Alright, [Teacher reading: [On board.]  

Okay here's about 30 or 40 digits.  I don't know. 

Okay, then I hit return, and now here's what it thinks I typed.  How many digits- I don't know, let's count them. 

16 digits. 

That's not what I typed.  I typed more 3's.  Hit return and they disappear.  That's not microsoft's fault.  That's this stuff- hardware.  Can't do it. 

Well, I mean the computer has gigabytes of memory, but the way the computer is designed is that each number is only given a limited number of bytes.  In Java, a decimal number probably is 8 bytes, which is 8 memory locations.  Each memory location can only hold 256 different numbers, so 8 can hold 256^8, which is not enough to go past 15 digits, or whatever it is. 

If I type in some... oh, this is worse than I thought. 

Okay, I'm sorry.  It's even worse.  I think whoever programmed excel could have done better.  Here, I typed in not the number, but the expression 1/3 and an equal sign which will make it calculate it on the spot.  Somewhere in the program it sores the 1, the division sign, and the 3, then when I hit return, it still just give a limited number of digits. 

It's 3.333... divided by 10 is 0.333333....

Male Student:  Is it like a floating decimal? 

Professor:  Yeah, I can't remember what it's called in the computer.  There's a special notation where some of the bytes are used just for this number, and another for the exponent.  Computer arithmetic is very technical, but however technical it is it doesn't work very well.  It'll do 1/4 pretty well, because 1/4 only has, you know, 2 decimal places. 

But 1/3, totally bad.  And it turns out that there are applications here that make a difference- like weather forecasting can make a difference. 

Alright.  Where were we?  And so this roundoff that I just showed you in excel was how Lorenz discovered the butterfly effect- the roundoff errors were tiny but made a huge difference. 

I'll show you- when I say he found it by doing some experiments with weather situations, he was just using 3 equations.  I can show you some stuff on youtube to show you the results on youtube.  He had an ideal situation with heating on the bottom and how it rolls over in a cube of air. 

Let's look at it.  It's called the Lorze attractor.  Let's see if we can find a few of those on youtube. 

Here's a whole bunch of simulations.  If you look at it in one direction, it looks like a figure 8.  If you just follow the values- you simulate the future values of x, y, and z and it'll form in this shape. 

It's just a simple simulation where he showed that if you rounded off- the question was which way it would go like the Lorenz water wheel.  A tiny change after a few cycles will alter the movement drastically. 

So let's... I'm tempted to show you this one.  Let's look at one of the others first just to see.  I don't know which is better or worse

It's the divergence issue.  That's what the butterfly effect.   

So it's showing the path of the x, y, and z point as the equations evolve. 

That's a different view of it

So what's going on here is it's showing different paths and they start out almost the same- different in their 13th decimal place or something.  So they go different directions sometimes. 

Alright, so you know, I've actually in undergraduate courses, I've given students the equations and I've told them to do this- if I did that, you could write a small computer program that would produce that, and I've had people do that just as a regular homework.  You have to learn a little more programming. 

Let's take a look at another one. 

This one is 8 minutes.  Let's try this one. 

Hmm. this one is kind of boring. 

Alright.  We get the idea.  Let's try another one. 

I don't like that one either.  Well, how about this one.  This will take ten minutes.  Let's try it. 

Everyone is interested in the weather

[Can't hear/can't understand.] 

[Muffled speaking]

You've got pets and you're going to leave them in the car in this kind of heat, make sure they have a lot of chew toys and roll up the windows so the heat doesn't get into the car. 

Professor:  The was a pretty good video, actually.  That was really good.  

I just need to keep a record of that one.  Well, I'll remember it. 

So that was really about the butterfly effect, so any external influence can be those butterfly wings. 

Alright, I'll just keep going.  This one is more kind of philosophical- not so physics-y.  Let's suppose you could control everything we've talked about and predict the future to a significant degree.  Then you could run into the problem- why care? 

There's a word called angst which means [On board.]  

Everybody has those feelings, sometimes. 

So does the future matter?  What do you think?  Anyone want to defend the future?  I charged the future with not mattering. 


Attorney for the defense? 

Defend the future.  I charge the future with not mattering. 

Now is your chance to defend it and say it does matter. 

You guys are making me say outrageous things to get a reaction. 

Male Student:  I could care less.  I mean, if a meteor came and killed us all, oh well. 

Professor:  I mean, I can't argue with that.  There are other ways of looking at it.  We still need someone to witness for the defense.  Anyone want to defend the future?  Do we all agree it doesn't matter? 

Professor:  It's just crazy, right? 

Male Student:  It's hard to really say anything neutral about it.  It's either one or the other petty much. 

Professor:  Okay.  Alright, well, I'm inclined to go on if only someone can just defend the future a little bit.  How about you?  You look like an optimistic soul. 

Female Student:  What's going to happen is going to happen.  There's nothing we can really do. 

Professor:  Will it matter? 

Female Student:  Well, maybe if something you did affect the future, but other than that, not really. 

Professor:  Okay, I can tell which way this jury's going to go.  I had jury duty yesterday.  Of course they didn't choose me to be on the jury, but I was still the for 2 hours.  Has anyone here ever been on a jury?  Your parents? 

Male Student:  My parents are lawyers. 

Professor:  Okay, well, they probably couldn't be jurors, but they might deal with them a lot.  Okay, here we're getting real pessimistic.  Does the existence of humanity matter? 

Male Student:  If we weren't here the world would be a lot better off. 

Professor:  One of the best known software developers- I think he helped develop Java his name is Bill Joy and he is a software expert.  He wrote a book called "why the future doesn't need us."  He made the point that the existence of humanity doesn't matter. 

How about the kind of existence?  Like, if life is difficult vs. prosperity.  Maybe a life of prosperity for everyone is a better future than a life that is difficult for everybody. 

Male Student:  That's kind of impossible though.  Not everyone can be prosperous. 

Professor:  Uh, that might be true, well, there are countries that sort of make sure that nobody-

Male Student:  Well, we buy all of our stuff from China now, so people in America lost their jobs, and the people in China are making 50 cents an hour to make this stuff. 

Professor:  Yeah, but you know, some people could argue that that's a distribution of wealth issue, not total wealth.  There's hardly anyone who really goes hungry in the US.  I don't think everyone here is prosperous, but at least the food problem in the US- well, there are people who are hungry, but there is a safety net to prevent that. 

It's kind of a downer of a conversation, but that was my point that that's that spoil sport is you sort of have to answer the question about why care before you can justify predicting it, I guess, or bothering it. 

So maybe I shouldn't ask the downer questions.  You know?  Chill out. 

Eat dessert first.  A very famous verse from the bible [On board.]  

What that means is you can't predict the future.  You might be run over by a bus, so you might as well have some fun today. 

But actually, I was looking up before class- someone said that an ancient king or whatever... I don't know... in ancient times someone came up with that, and it appears in the bible as well. 

Here's another one- surely you've heard this one. 

Male Student:  I thought Bob Marely said that. 

Professor:  Well a lot of people did.  Does anyone know about McFerrin?  I guess he's an old singer.  He was pretty well known in his day.  The phrase appeared on post cards first though trying to promote a kind of spiritual mystic named Meher Baba. 

He had some teachings which were more complicated than this but included this, but for emphasis, they made postcards with just this quote. 

Let's check this out on wikipedia.  So here's Bobby McFerrin,. released in 1988.  That wasn't that long ago.  It was originally a quote by this guy [On board.]  

Okay, here.  BoB Marley.  Unfortunately not correctly.  Here's Meher Baba he died a long time ago.  Actually, well, yeah. 

There have been people throughout history who have gone that rout.  But that's what the phrase suggests- the same as these.  And these are just, you know, results of sort of the philosophy that the future doesn't matter. 

Getting back sort of down to earth, a lot of times people make decisions that focus on the short term.  Like businesses and people- they focus on the next stock prices in the next quarter, even though they might do better focusing on a longer term goal.  Political decisions- politicians don't often think long term.  They think about the next election. 

It has to be a result of short term thinking.  

So politicians are people too.  Well, everyday people focus on the short term.  What about animals?  Do they think or act in short term? 

Short term or long term? 

Well, if you ever had a dog or a cat or something, surely you don't think the dog thinks much about tomorrow. 

He sort of reacts to what's happening right now.  Little kids- same way. 

"Hey, it's time to go shopping or to Grandma's."  Little kids will react instantly to what happens in the moment.  Most animals are pretty short term.  Some animals act ahead. 

Birds will fly south before winter, but that is instinctual, not thought about. 

You've heard the Aesop's fable the ant and the grasshopper.  Anyone not familiar with that? 

Okay Aesop was an Egyptian and he came up with a bunch of stories with lessons, and we still read them today.  The one about the ant and the grasshopper goes there was an ant and the grasshopper and the ant worked hard all day storing food and the grasshopper didn't save anything and the grasshopper made fun of the ant for working and the ant would just say "we need to prepare for the winter."

Then winter came and the grasshopper froze to death and the ant had food for the winter.  So the moral of the story is you should prepare for the future. 

You should all look you Aesop some time.  Your kids would love those fables if you ever have any. 

Okay, we're kind of running out of time.  There's a philosopher named Kierkegaard who really sort of dug into this tendency people have to quietly struggle with the meaninglessness of life.  I have some more here, but I want to get into the last... okay, if you have a tendency to think the future is meaningless, it's your choice, but you can also choose to think more optimistically, and you'll probably be happier.  It's not just a cheap self-help book.  It was written by the former president of the APA.  He says you'll be happier if you're optimistic. 
It's not just whether you are or not.  You can control whether you are. 

So I'm going to make a suggestion that it's a good idea to try to maximize the sum of all the positive feelings of the world.  If we could increase that sum- add it all together.  If the world was run so it would increase the amount of positive feelings in the world, what more could you ask for out of the world or the human condition? 

Of course I know that no country is run this way, but I think it would be nice if they where run that way. 

Alright,  Any commens?  Questions? 

Alright, I'll just say a couple of words about.... I've got so much to say, but so little time to say it.  I think that's good enough.  You get the idea. 

No comments:

Post a Comment