The Stilwell Brain

published on July 9, 2020

♪"I think, therefore, I am"But am I?I think HaA single microscopic brain cellcannot think,is not conscious,but if you bring ina few more brain cells,

And a few more,and connect them all,at a certain point,the group itself will be able to thinkand experience emotionsand have opinionsand a personalityand know that it exists

How can such astonishing thingsbe made fromsuch simple ingredients?Well, answering that questionmeans learning not only whowe are,but, more importantly, how we areToday, using whatneuroscientists know so far,I am going to make my hometown

Function like a brain!( all cheering, applauding ) ♪A single brain cell is tiny,both in size and abilitiesBut when enough are together,they can do amazing things

Like be aware of themselvesWhen the collective powerof a group working togetheris greater than the sumof their individual parts,that is called "emergence"In a similar fashion,we as individualsare connected tothe people around us

Those connectionsform communities that,when functioning properly,can work togetherto accomplish amazing featsA great example is"wisdom of the crowds"Even if not a single personin a crowdknows the right answerto a question,collectively, they could allsomehow know the right answer

In 1987, economistJack Treynorconductedthe "Bean Jar" experimentHe asked 56 studentsto guess the numberof jellybeans in a jarNow, as you can probably guess,not a single one of themguessed the right answerBut amazingly, when he tookthe average of their guesses,

What he got was a number withinjust 3% of the real answerNow, some peopleguessed way too high,but others guessedway too low,so all together,their errors balanced out,and from a whole bunchof wrong guesses,the true answer emerged

What else can a crowd do?If I got a bunchof humans togetherand had each one of themact like a brain cell,turning on or off in responseto the actions of other people,could I make a neural networklike the one in our brain?

And if I had enough people,could intelligence, emotions,a mind, emerge? If I recruited every single person in the country of China and arranged them like neurons,

would the result not only be a simple brain, but something that can think and feel and be aware of its own existence?Well, this is the China Brainthought experiment,first proposed by Lawrence Davisand, later, Ned BlockIt's never been done beforeand, well, unfortunately,

I don't have accessto everyone in ChinaI made some calls,and like a lot of them are busyBut the first step is to seewhat a crowd in real lifecould even doThis hasn't been donesuccessfully before,but I want to blowa neural network

Up to the scale of a crowdAnd what better crowd to usethan one made of the peoplewhose emergent propertiesmade me who I am today?That's right, I am going hometo Stilwell, Kansas ♪ ♪

( birds chirping )Michael: For help designing the brain we would make out of people, I recruited Chris Eliasmith, director of the Center for Theoretical Neuroscience at the University of Waterloo

So Chris,we're headed south,going down tothe heart of Stilwell,- where I grew up- NiceWe're going to do somethinga little bit weird UmI want to create a brain- Right- OK? But with a crowd of people

It sounds like a challenge,for sureI looked into it,and I found that the roundwormhas a brainthat's made up of only300-some-odd neurons- That's right- We can get 300 people,and where better to getthese people to make a brain

Than my hometown of Stilwell?This was the communitythat, in many ways,made me who I amMichael: This is all downtown Stilwell Some of my earliest memories are from hereThis used to be,and maybe still is,a feed store,and they would have sno-conesduring the summer

It was the most awesome,delicious thing everBut as you can see, a lot of corn is grown in Kansas, but around here, the main thing that I saw being grown- was just sod- Oh, really?Yeah, There's a famous sod farmaround here

Whose slogan was"High on grass"- ( Chris laughs )- It was prettypretty edgy for the timeOK, so back to the brainthat we're gonna makeYou know, building brains isin my job descriptionI wrote a book called How to Build a BrainMichael: Chris is known for is neural network,

the Semantic Pointer Architecture Unified Network, or SPAUN, which is one of the world's most complex computer simulations of the brain It uses 66 million simulated neurons to perform functions like counting, reasoning, and image recognition

SPAUN is cutting-edge, but neural networks are nothing new The first was made by Dr Frank Rosenblatt of Cornell University in 1957 His network, called the Perceptron, was designed for image recognition,

and he hoped it would become capable of learning, just like a brain But the project was only partially successful, and after some controversy, fell by the wayside It was only when researchers in the 1980s came back upon Dr Rosenblatt's work,

and as computing power increased, that the field of artificial neural networks came back to the mainstream Today, it is alive and well SPAUN, and even neural networks used in self-driving cars, are expanding the possibilities of computer learning

If I want to makea brain out of people,where do I start?That's a good questionI think the first thingwe want to do is figure outwhat we want our brain to doI would recommendsomething like visionVision Let's makethis brain see

Michael: Before we can design the intricacies of the brain we're making, let's look at how visual processing works Let's say we look at a cat Light information from every point on the cat lands on the retina

This information gets sent to our visual cortex The visual cortex is structured in layers– V1 through V6 Each of these layers are made up of neurons activated by specific features, like lines, angles, and shapes

The features that are detected are sent to the infratemporal cortex which puts all the pieces of the image together, and we get our Eureka! moment where we recognize the object we're looking at, what it means what feelings we have towards it

I love catsBut what should we haveour brain recognize?We don't wantreally high resolution imagesor images that dependon too much detail,- Ok- so things like lettersand digitsLet's say we use digits Ok?

– Ok- I want to be the onewho draws a digit,and then you will beon the output sideYou should be ableto determine what I've drawn;not because I showed itto someone and theytelephoned it back to you,but becausethey processed itintelligentlyThat's what we needto figure out,

How we're gonnashow an input to our peopleSo we should takesome small number of themand put them at the front,as the retina,and really just show them eacha little bit of the imageSo if, for instance,we're able to put like25 people in that kindof front row,the "input" layer,then whatever image we show

Should be made upof 25 pixels- Exactly Right- Twenty-five piecesI'm gonna draw25 people1, 2, 3, 4, 5,6, 7, 8, 9, 10,11, 12, 13, 14, 15,

16, 17, 18, 19, 20,21, 22, 23, 24, 25See? I can countThese are our retina cells,and each one isan individual personthat's literally standing,like, in a field

What do they then do next?They merely need to indicatewhether or nottheir cell is on or off- All right- So, they should start firingThey should start spikinglike a neuronWhat could they doto indicate

That they're firing or not?They could jump up and down,they could wave a flagOK, I like thatWhen Chris and I use wordslike "firing" and "spiking,"we're talking abouthow brain cells, neurons,

Talk to one anotherby sending an electric messagefrom one cellto anotherIt's calledan "action potential,"and it travels down the axonof the cellWhen the ionic flowinto a brain cell

Reaches a certain threshold,the cell will fire an electronicmessage down its axonSo a neuroncan either be on or offIt's either firingor it's not firingWhat we need to findis a way for a personto be either on or off–

Raising a flag or their handshould do the trick To illustrate our visual input, I will be drawing a number from 0 to 9 onto a grid divided into 25 squares, or pixels Now each person, or neuron, will receive one pixel If a neuron receives a pixel with writing on it,

it will fire The V1 layer identifies pixels in the retinal layer that form particular lines in the number, and the V2 layer identifies particular combinations of lines from V1 that form angles- What does V3 do?- V3 is more sensitive to color

We're only working withblack and white in this caseSo you're saying we won't evenneed to have a V3 in our brain?We're skipping V3 altogetherAll right, sorry, V3- So we're gonna go straight to V4?- YeahMichael: V4 neurons will fire

when their assigned combinations of angles have been detected At this point the basic shape of a number is beginning to take formAnd so actually the next oneis called IT- Ooh!- And that stands for infratemporal cortex

Michael: Now don't worry We haven't forgotten V5 and V6 They exist, they're responsible for higher-level image processing in our brain But for our demonstration, we don't need them We do, however, need the infratemporal cortex,

which is the final layer needed for visual processing in the brain we're designing Our IT will consist of ten neurons representing the numbers 0 through 9 They will be looking at neurons in V4, and will only fire when their corresponding neurons fire

For example, if one or multiple V4 neurons representing the shapes of an 8 fire, the IT neuron representing the 8 will also fire Voila! We just recognized a numberSo what happens afterthe infratemporal layer?So after that,I think that's where I'll be

It's gonna be me making adecision about what digitI think was actually shown atthe endSo I think we're gonna needa couple hundred people,so one question is,where do you putthat many people?I would love to usemy high school football fieldThe question is,is it gonna work?- I am hopeful right now- We got our work cut out forus

Michael: Knowing how we want to structure our neural network, it was now time for us to head to my high school football field where our "brain" will take form I spent a lot of time here impressing the world with my body's prowess–

at the clarinetOK, so Chris,I brought you herebecause we need to talk aboutthe actual logistics of gettingall these people togetherSo here's the planas I see itEveryone is going to bewearing a shirt

That is a different color,based on what layer they're inWe're also gonna giveeveryone one of those, um,like, "I'm runningin a marathon" kind of- The big bibs?- Bibs Thank you YesI knew that too You see,I'm always checking, becauseChris is one of those nerds

Who doesn't knowwhat us sportostalk aboutAnyway, the bibswill give every personan individual number,so if something goes wrong,we can targetthat one brain celland say, "Are you damaged?What do you need?"Take out the problem

OK, we're hereon this 40 yard lineI will be hereThis is gonna bethe input layerThe retinal cells will beall in front of me,all 25 of themYou are gonna beway down in the end zoneon the output sideAnd I'm gonna beusing the scoreboard-Ok- So when youmake your prediction,

Based on what you thinkthe brain has figured out,we'll put that onthe Visitor's sideand I'll reveal the Home numberas what I really wrote- Sounds good- So literally from here to that end zoneis the amount of spacewe're going to needfor these hundreds of people

To also havethe right eye linesJust have to make sure thatcommunication lines are open,meaning that it's easy to seewhoever you have topay attention toMichael: Our human brain will have a couple hundred people spread out in five layers across half a football field

Every single participant will be assigned to only react to certain neurons in the layer ahead of them And it's complicated, so their positions on the field had to be carefully chosen so that every neuron has a clear line of sight

to the neurons they are connected toIn a waysomething will be bornon this field tomorrow( laughs )- We shall find out- We'll find out!All right

( crowd chattering )Michael: So what does it take to turn Stilwell into a brain? Well, seven tents, 550 chairs, twenty gallons of coffee, three hundred flags, t-shirt and hats, our drone operator Jeff,

this cute little Gator, two hundred cinnamon rolls, and of course, our medic, Brian Now all that's left is to pull this all off A community is something that is bigger that the sum of all of its parts,

And so is a brainNow, today, I'm feelingpretty excited aboutthe neural networkwe're gonna build out of people,because there's a zero percentchance of rain,but a 100% chance of brainOK, we better get started

The gates are open, and our neurons are filing in First, they're all given color-coded t-shirts associated with the layer of the brain they will representJust want go in the center?Michael: And then they will take to the field to get in position

Michael:You all in the orange shirtsare the retinaYour job is to say,"Is there writing on my square?"or "Not" If your squarehas any writingor black marks on it,raise your flag–oh, and stand up Now I'm sure all you mega-brainiacs out there remember every detail about how this brain

is going to recognize numbers But just in case you don't, here's a refresher I will draw a number on a 25-pixel grid, break the squares up, and hand them out to every person in the retina layer The retinal neurons will only fire if they have writing on their pixel

The people in the yellow shirts,you guys are V1 Each V1 neuron will be watching three retinal neurons in front of them and fire only if all three of their assigned retinal neurons fire,

revealing lines that make up the numberYou guys are V2You're a bit more advancedYou're looking for combinationsof featuresthat make, for instance, angles The V2 neurons will be watching the V1 layer,

and fire only if their assigned V1 neurons fire, revealing angles V4 neurons will be looking at V2 neurons Their firing reveals combinations of angles that begin to form the numberFinally, the purple shirts

You all areinfratemporal cortexExtremely important role,not more importantthan the others, though Part of the IT's function is to inhibit incorrect results it receives from the V4 layer For example, if V4 neurons are indicating

both a 6 and an 8, an 8 will outrank a 6 because an 8 has more features Chris will determine the number by interpreting the results from the IT layer Got it? Good Because it's happening nowMichael (over loudspeaker):All right!It is time for meto draw my first numeral

Stand byThere it isNow it's time to distributethese pixelsto the photoreceptorsin the retina25, 2419

13All rightI have distributedthe input to the retina layerIs everyone ready?( all cheering, applauding )

Three, two,one, think! And they're off The retina layer has fired, passing off signals to V1 V2 sees V1 firing, and also fires, cuing V4 and IT There's a lot of flags on the play, folks

Look at all this processingGood work, stay up,I'm now walking back to Chriswhere he will tell mewhat you guys have processedAll right, Chris,it actually happenedway faster than I thought

It took me foreverto get over hereThey were already done"thinking"It was really interestingto watchWe had a little bit of noisein the system, for sure,because we actually kind ofhave two answers at the endSo I'm going to be doingsomething that brains do,

Which is kind ofmake a guess sometimesbased on the best evidenceMichael:All right, ChrisWhat numeraldo you think I drew?Chris:I think you drew a 3Let's get that upon the Visitor's scoreboardThe numeral I truly drew

Was a 3!( all cheering, applauding )Chris:Nice workThere was some noisein the systemI think we can perfect thisa little bit,because it wasn'ta confident 3,

But Chris did stillget it rightAnd by "Chris,"I mean all of youMichael: We gave each of our IT neurons a clear tube and plastic balls to make their job easier Every time they see one of the neurons they're watching fire, they put a ball in their tube

If a neuron they're inhibited by has more balls than they do, they stop firingMichael:Our model had a mistakeYou should have beeninhibited by 254,meaning that if 254 is firing,

And puts a ball in the tube,- you just sit down- OKBut we didn't have 254written down into your codeI'm gonna do that right now With that kink worked out, it was time to try again

Take that oneThank you24Michael:Three, two, onego!Michael:Ooh There's not much happeningin V2 and V4Or IT, for that matter

Looks like our brain diedMichael: Something definitely went wrong The processing stopped in V2So, I think our brain brokeNot a single person in V4or the infratemporal cortex

Has been activatedChris: Seems like somethingvery strange happenedwhen the lightwent through the lensand got to the retinaOK, but what you really mean tosay is that I handed the pixelsout to the wrong neurons- It's lookin' that way- Wow, who would have thoughtthat the worst-workingpart of the machine

Would be the actual peoplewho get to be people- and not brain cells?- I know, right?OK, I thinkI'm gonna do an 8 this timeAnd I'm gonna put itkind of up in this corner,or along that sideThis is pretty weird,it's not centered,

It's not filling upthe whole spaceLet's see if our braincan recognize it Each pixel of our image needs to go to a particular retinal neuron To make sure I didn't mess it up again, we put numbers on the back of each one

Is everyone ready?- ( all cheer )- OKThree, two, one, go!Michael:Oh, yeahWhoa!That was fastMichael: Blink, and you'll miss it, so let's take a break,

because this is the Mind Field Play of the Game The 8 I drew contained 13 pixels, and bam! the 13 retinal cells connected to those locations are firing Now, that's what I call a sensation V1 reads the formation perfectly They don't even know it, but each one firing

means a horizontal, vertical, or diagonal line has been caught Now look at neuron 40's speed It's sensitive only to a horizontal line low and to the right, which my 8 had If retinal cells 23, 24, and 25 all fire together,

such a line has been sensed, and watch this–boom! Champion reflexes there, folks If the V1 neurons a V2 player is watching, fire, they stand, and that means that the lines V1 caught made some corner angle Standing V4 neurons are shapes made by those corners,

but it all comes down to IT A bunch of them fire Lots of numbers contain the shapes I drew, but there can only be one MVP, and, dang, look at this teamwork! 3 is inhibited by 8,

0 is inhibited by 8 If 8 is getting this much activity, sit down and let that neuron score! This, my friends, is what we in the cognition sports biz calla gr8 playChris (over loudspeaker:The purple people neurons

Make me think thatyou wrote an 8An 8? Let's put an 8up on the scoreboardAnd as it turns out,the numeral that I did writewas an 8!( all cheering, applauding )- Nice job!- That was good

Michael: This was our first definitive success Now it's time to really put the system to the testSo I'm gonna draw a 1,but then I'm gonna adda line down hereand I'm gonna doa dot right thereAnd we're gonna see ifthat noise trips up our brain

Three, two, one, go!My guess is a 1The number that I drew was,in fact, a 1( cheers, applause )7and I'm gonna puta line through itGo!

Oh, yeahMan, these people are goodChris:The brain thinksit's just saw a 7I wrote a 7- ( cheers, applause )- Nice workNice work

Michael: With two successful results in a row, for the last test I want to see what will happen if I really mess with our brainI'm not even going to drawa number;instead, I'm going tofill in every single cellThis means thatevery single neuronin the retina will fire,

Will stand up and waveLet's see what happensMy prediction, of course,it's gonna look like an 8,because an 8 is a numeral thatfills in a lot of the cellsAre you all ready?- ( all cheer )- Good- Chris, are you ready?- Ready!

Michael:All right, I'm readyThree, two, one, go!Chris:What?!( laughing )- ( all cheering )- Michael: Look at that- Michael: Holy cow!- ( Chris laughs )

It looks to me like youessentially opened up the eyeand shone a laserright into itRight, yeahSo what I did is,I didn't even draw a numeralI just scribbledall over the whole thing,I filled in every single cell- So what does the brain think that I drew?- The 8I had no idea the inhibitionwould work that well

We were able to singleall of that messdown into just one guess,and it really wasthe smartest guessRight, it was the one withthe most features in itMichael:Congratulations tothe entire infratemporal cortex,V4, V2, V1, retina

You guys have been amazingGreat work( cheers, applause )Today, I was a neuronMy favorite partabout the brainwas that it actually workedMy favorite partwas just the whole experience,

Doing science,meeting cool peopleIt's just a really goodsimulation of howthe brain works,and it was just really coolto take some informationfrom thatMichael:And as always,thanks forbeing a brain( all cheer )

Michael: Our small-town brain did work as predictedMade up of only a couple hundred neurons,each one with no ideawhat number I was drawing,it was nonetheless ableto process the imageand determinethe correct answer

We created a living,breathing modelof a part of the human brainAnd our demonstrationwas a new wayto illustrate and sharehow the human brainprocesses visual informationWe were able to watch itprocessthink

In real time,and that's amazingIts success shows what we canachieve by working together,and we only useda small fractionof the number of neurons foundin an actual human brainSo imagine how powerfulthe connections ofnot a few hundred,

But a hundred billion human neurons could beNow, interestingly,a hundred billion peopleis about how many humanshave ever existedin the history of EarthThere are only a few billionalive right now,so I guess that meansget procreatingplease

I want to make a biggersuperhuman mindNo I'd like to thankevery neuronfrom my hometown of Stilwell,and the entire communitythat supported usBecause without themall working together,none of thiswould have been possible

And as alwaysthanks for watching ♪- ( no audible dialogue )- ♪

Related Videos

hey guys techrax here coming at you guys with another giveaway this is gonna be an iPhone 5s two of these brand new phones not the one in this video guys this i...
hey guy there's a popular game in the app store called flappy bird right now and it's supposedly really annoying believe it or not I haven't played ...
hey guys tech Rex here so right here with me I have an Apple iPhone 5s this is the gold one I also have with me some liquid nitrogen yes this is the real deal a...
hey guys techrax here right here I have the iPhone 5s with me this is the champagne or the gold color I still can't really figure out if it's champagne ...
hey guys tear cracks here so this video really exciting video I have the new Samsung Galaxy s5 for you guys this is the long-awaited cellular devices releases i...
everyone techrax here in this video I've got the latest Samsung Galaxy s5 right here as well as the Apple iPhone 5s and we're going to be doing a simple...
everyone techrax here in this video guys have a really exciting device this is the Samsung Galaxy s5 charcoal black and I'm really liking I after seeing the...
everyone techrax here here with me off the Samsung Galaxy s5 this is a perfectly fine s5 it is cracked from the drop test that I had with also one minor neck as...
hey guys tetrax here so in this video I'm going to try and burn the newly released Samsung Galaxy s5 this is the shimmering white 16 gigabyte model and if y...
everyone techrax here so I got my burn Samsung Galaxy s5 and I wanted to see whether the heartbeat sensor would still work the heart monitor on your galaxy s5 a...
hey guys Tech Rex here so I'm really excited to bring you guys a giveaway for my channel but this time I'm actually teaming up with a buddy of mine your...
hey guys techrax here so in this video I'm going to be hopefully instructing you guys how to make your very own a tech sandwich slash burger slash meal so y...
hey guys techrax here so I've got a galaxy s5 here this is the copper gold hopefully you guys can see pretty well it is sunset so it's getting a little ...
hey guys techrax here so right here with me I have a professional deep fryer in here is already some canola or corn or whatever oil I don't know vegetable o...
hey guys techrax here so just trying to make this video short and quick i'm having recently i got five hundred thousand subscribers and yeah most of you guy...
hey guys texture so in this video I have a drop test on the latest LG g3 device now this is actually the gold-coloured LG g3 this has not been released in the U...
hey guys techrax here so in this video we'll teach you guys how to make your iPhone indestructible this is essentially a case that's been around for yea...
hey guys techrax here soon in this video we had a train run over the iPhone 5s so we actually did this in two different instances initially we had a Space Gray ...
hey guys tech cracks here so in this video we're going to be dropped testing the newly released Amazon fire phone this is exclusive for AT&T and I belie...
hey guys techrax here so right here with me I have the Amazon fire phone this is the one I dropped and you know I thought what better what else do I do with thi...
Be the first to comment “The Stilwell Brain”

There are no comments yet.