Kenneth Cukier: Big data is better data

Prevodilac: Milana Stojadinov Lektor: Mile Živković

America's favorite pie is?

Koja je omiljena američka pita?

Audience: Apple. Kenneth Cukier: Apple. Of course it is.

Publika: Od jabuke. Kenet Kukir: Od jabuke. Naravno.

Kako to znamo?

How do we know it?

Because of data.

Zbog podataka.

You look at supermarket sales.

Posmatramo rasprodaju u supermarketima,

You look at supermarket sales of 30-centimeter pies

prodaju zamrznutih pita prečnika 30 cm,

that are frozen, and apple wins, no contest.

i jabuka pobeđuje.

Bez konkurencije.

Najveći deo prodaje je od jabuka.

The majority of the sales are apple.

But then supermarkets started selling

Zatim su supermarketi počeli da prodaju manje pite,

smaller, 11-centimeter pies,

pite prečnika 11 cm.

and suddenly, apple fell to fourth or fifth place.

Odjednom, jabuka pada na četvrto ili peto mesto.

Why? What happened?

Zašto? Šta se dogodilo?

Okay, think about it.

Dobro. Razmislite o tome.

When you buy a 30-centimeter pie,

Kada kupite pitu od 30cm,

the whole family has to agree,

cela porodica mora da se složi,

and apple is everyone's second favorite.

a pita od jabuka je svima drugi omiljeni izbor.

(Laughter)

(Smeh)

But when you buy an individual 11-centimeter pie,

Ali kad kupite zasebnu pitu od 11cm,

you can buy the one that you want.

možete da kupite onu koju vi hoćete.

You can get your first choice.

Možete da uzmete vaš prvi izbor.

You have more data.

Imate više podataka.

You can see something

Možete da vidite nešto

that you couldn't see

što niste mogli da vidite kada ste ih imali u manjim količinama.

when you only had smaller amounts of it.

Now, the point here is that more data

Dakle, poenta je da više podataka

doesn't just let us see more,

ne samo što nam omogućava da vidimo više,

more of the same thing we were looking at.

više o tome što posmatramo.

More data allows us to see new.

Više podataka nam omogućava da vidimo novo.

It allows us to see better.

Omogućava nam da vidimo bolje.

It allows us to see different.

Omogućava nam da vidimo različito.

In this case, it allows us to see

U ovom slučaju, omogućava nam da vidimo

what America's favorite pie is:

koja je omiljena američka pita:

not apple.

nije od jabuka.

Now, you probably all have heard the term big data.

Svi ste verovatno čuli izraz "veliki podaci".

In fact, you're probably sick of hearing the term

Verovatno vam je i loše na pomenu izraza

big data.

"veliki podaci".

It is true that there is a lot of hype around the term,

Tačno je da se podigla velika buka oko ovog izraza,

and that is very unfortunate,

što je loše.

because big data is an extremely important tool

Zato što su veliki podaci veoma važan alat

by which society is going to advance.

pomoću kog će društvo da napreduje.

In the past, we used to look at small data

U prošlosti smo posmatrali "male podatke"

and think about what it would mean

i razmišljali o tome šta bi značilo

to try to understand the world,

da pokušamo da razumemo svet,

and now we have a lot more of it,

a sada ih imamo mnogo više,

more than we ever could before.

više nego što smo ikada imali.

What we find is that when we have

Shvatili smo da kada imamo mnogo podataka,

a large body of data, we can fundamentally do things

u principu možemo uraditi stvari

that we couldn't do when we only had smaller amounts.

koje nismo mogli sa manje podataka.

Big data is important, and big data is new,

Veliki podaci su bitni, i to je nešto novo,

and when you think about it,

kada razmislimo o tome,

the only way this planet is going to deal

jedini način na koji će se planeta suočiti

with its global challenges —

sa svojim globalnim izazovima -

to feed people, supply them with medical care,

nahraniti ljude, obezbediti im medicinsku negu,

supply them with energy, electricity,

pružiti im energiju, struju,

and to make sure they're not burnt to a crisp

da se pobrine da ne izgore

because of global warming —

zbog globalnog zagrevanja -

is because of the effective use of data.

jeste zbog efikasne upotrebe podataka.

So what is new about big data? What is the big deal?

Šta je novo u vezi sa velikim podacima? U čemu je velika caka?

Well, to answer that question, let's think about

Da bismo odgovorili na to pitanje,

razmislimo kako su informacije izgledale,

what information looked like,

physically looked like in the past.

fizički izgledale u prošlosti.

In 1908, on the island of Crete,

1908. godine na Kritu,

archaeologists discovered a clay disc.

arheolozi su pronašli glineni disk.

They dated it from 2000 B.C., so it's 4,000 years old.

Smestili su ga oko 2000. g. pre Hrista, dakle star je 4000 godina.

Now, there's inscriptions on this disc,

Na tom disku postoji zapis,

but we actually don't know what it means.

ali ne znamo šta on znači.

It's a complete mystery, but the point is that

Potpuna je zagonetka, ali poenta je u tome

this is what information used to look like

da su tako informacije izgledale

4,000 years ago.

pre 4000 godina.

This is how society stored

Tako je društvo čuvalo

and transmitted information.

i prenosilo informacije.

Now, society hasn't advanced all that much.

Društvo nije baš toliko napredovalo.

We still store information on discs,

I dalje čuvamo informacije na diskovima,

but now we can store a lot more information,

ali danas možemo da čuvamo mnogo više,

more than ever before.

više nego ikada.

Searching it is easier. Copying it easier.

Pretraživanje je lakše. Kopiranje je lakše.

Sharing it is easier. Processing it is easier.

Deljenje je lakše. Obrada je lakša.

And what we can do is we can reuse this information

Možemo da koristimo te informacije iznova,

for uses that we never even imagined

na načine na koje nismo ni zamišljali

when we first collected the data.

kada smo počeli da sakupljamo podatke.

In this respect, the data has gone

U tom smislu,

podaci su prešli iz skladištenja u protok.

from a stock to a flow,

from something that is stationary and static

Od nečega što je stacionarno i statično

to something that is fluid and dynamic.

do nečega što je fluidno i dinamično.

There is, if you will, a liquidity to information.

Ako ćemo tako, informacija je kao tečnost.

The disc that was discovered off of Crete

Disk, koji je otkriven u blizini Krita,

that's 4,000 years old, is heavy,

pre 4000 godina, je težak.

it doesn't store a lot of information,

Ne sadrži puno informacija,

and that information is unchangeable.

i te informacije su nepromenljive.

By contrast, all of the files

Nasuprot tome, svi fajlovi

that Edward Snowden took

koje je Edvard Snouden uzeo

from the National Security Agency in the United States

od Državne bezbednosne agencije u SAD-u

fits on a memory stick

staju na memorijski uređaj

the size of a fingernail,

veličine nokta,

and it can be shared at the speed of light.

i mogu se razmenjivati brzinom svetlosti.

More data. More.

Još podataka. Više.

Jedan razlog zašto danas imamo toliko podataka

Now, one reason why we have so much data in the world today

is we are collecting things

je što sakupljamo stvari

that we've always collected information on,

o kojima smo uvek skupljali informacije,

but another reason why is we're taking things

ali drugi razlog je zato što uzimamo stvari

that have always been informational

koje su uvek bile informativne

but have never been rendered into a data format

ali nikad nisu prebačene u oblik podataka

and we are putting it into data.

i stavljamo ih u podatke.

Think, for example, the question of location.

Zamislite, npr. pitanje lokacije.

Take, for example, Martin Luther.

Uzmimo Martina Lutera za primer.

If we wanted to know in the 1500s

Da smo 1500. god. želeli da znamo

where Martin Luther was,

gde je Martin Luter,

we would have to follow him at all times,

morali bismo da ga pratimo u svakom trenutku,

maybe with a feathery quill and an inkwell,

možda sa perom i mastilom,

and record it,

i da to beležimo,

but now think about what it looks like today.

ali razmislite kako to izgleda danas.

You know that somewhere,

Znate da negde,

probably in a telecommunications carrier's database,

verovatno u bazi podataka telefonskog operatera,

there is a spreadsheet or at least a database entry

postoji tabela ili bar podatak u bazi

that records your information

koji beleži informaciju

of where you've been at all times.

o tome gde ste bili u svakom momentu.

If you have a cell phone,

Ako imate mobilni telefon,

and that cell phone has GPS, but even if it doesn't have GPS,

koji ima GPS, čak i ako nema GPS,

it can record your information.

on čuva informacije.

In this respect, location has been datafied.

U ovom smislu, lokacija je postala "podatkovana".

Now think, for example, of the issue of posture,

Razmislimo, npr. o pitanju držanja,

the way that you are all sitting right now,

načinu na koji upravo sedite,

the way that you sit,

načinu na koji vi sedite,

the way that you sit, the way that you sit.

načinu na koji vi sedite, i vi.

It's all different, and it's a function of your leg length

Svi se razlikuju, i zavise od dužine nogu

and your back and the contours of your back,

i leđa i od konture leđa,

and if I were to put sensors, maybe 100 sensors

i, ako bih postavio senzore, možda 100 senzora

into all of your chairs right now,

u sve vaše stolice,

I could create an index that's fairly unique to you,

našao bih indeks koji je jedinstven za svakoga,

sort of like a fingerprint, but it's not your finger.

kao otisak prsta, ali nije od prsta.

So what could we do with this?

Međutim, šta bismo mogli sa tim?

Researchers in Tokyo are using it

Istraživači u Tokiju ga koriste

as a potential anti-theft device in cars.

kao potencijalni alarmni uređaj u kolima.

The idea is that the carjacker sits behind the wheel,

Ideja je da ako za volan sedne lopov,

tries to stream off, but the car recognizes

pokuša da pobegne, ali automobil prepozna

that a non-approved driver is behind the wheel,

da za volanom nije odobreni vozač,

and maybe the engine just stops, unless you

možda zaustavi motor, osim ako vozač

type in a password into the dashboard

ne unese šifru u kontrolnu tablu

to say, "Hey, I have authorization to drive." Great.

da kaže: "Hej, imam dozvolu da vozim". Odlično!

What if every single car in Europe

Šta ako bi svaki automobil u Evropi

had this technology in it?

imao ovu tehnologiju?

What could we do then?

Šta bismo mogli tada?

Maybe, if we aggregated the data,

Možda, kada bismo nagomilali podatke,

maybe we could identify telltale signs

mogli bismo da uočimo znakove upozorenja

that best predict that a car accident

koji najbolje predviđaju

is going to take place in the next five seconds.

da će se dogoditi automobilska nesreća u narednih pet sekundi.

And then what we will have datafied

Tada bismo u obliku podataka beležili

is driver fatigue,

zamor vozača,

and the service would be when the car senses

i svrha bi bila da kada kola osete

that the person slumps into that position,

da je vozač upao u određeni položaj,

automatically knows, hey, set an internal alarm

automatski kaže: "Hej, pusti interni alarm."

that would vibrate the steering wheel, honk inside

kojim bi zavibrirao volan, zatrubio iznutra i rekao

to say, "Hey, wake up,

"Hej, budi se!

pay more attention to the road."

obrati više pažnje na put."

These are the sorts of things we can do

To su neke stvari koje možemo da uradimo

when we datafy more aspects of our lives.

kada prebacimo u podatke više aspekata naših života.

So what is the value of big data?

Koja je vrednost velikih podataka?

Well, think about it.

Pa, razmislite o tome.

You have more information.

Imate više informacija.

You can do things that you couldn't do before.

Možete da uradite ono što niste mogli ranije.

One of the most impressive areas

Jedna od najimpresivnijih oblasti

where this concept is taking place

u kojoj ovaj koncept igra ulogu

is in the area of machine learning.

jeste mašinsko učenje.

Machine learning is a branch of artificial intelligence,

Mašinsko učenje je grana veštačke inteligencije,

which itself is a branch of computer science.

koja je grana računarskih nauka.

The general idea is that instead of

Glavna ideja je da umesto

instructing a computer what do do,

da kažemo računaru šta da radi,

we are going to simply throw data at the problem

jednostavno ubacimo podatke u problem

and tell the computer to figure it out for itself.

i kažemo računaru da ga reši sam.

Pomoći će vam da ga razumete

And it will help you understand it

by seeing its origins.

gledajući u njegove korene.

In the 1950s, a computer scientist

U 1950-im, informatičar u IBM-u,

at IBM named Arthur Samuel liked to play checkers,

Artur Semjuel, voleo je da igra "Damu",

so he wrote a computer program

te je napisao kompjuterski program

so he could play against the computer.

kako bi igrao protiv računara.

He played. He won.

Igrao je. Pobedio je.

He played. He won.

Igrao je. Pobedio je.

He played. He won,

Igrao, pobedio.

because the computer only knew

Jer je računar znao dozvoljene poteze.

what a legal move was.

Arthur Samuel knew something else.

Artur Semjuel je znao nešto drugo.

Arthur Samuel knew strategy.

Artur Semjuel je poznavao strategiju.

So he wrote a small sub-program alongside it

Napisao je mali potprogram, pored ovog,

operating in the background, and all it did

koji je radio u pozadini,

was score the probability

i samo računao verovatnoću

that a given board configuration would likely lead

da data situacija na tabli pre vodi

to a winning board versus a losing board

ka pobedničkoj tabli nego ka gubitničkoj,

after every move.

nakon svakog poteza.

He plays the computer. He wins.

Igra protiv računara. Pobeđuje.

He plays the computer. He wins.

Igra protiv računara.

Pobeđuje.

He plays the computer. He wins.

Igra protiv računara. Pobeđuje.

And then Arthur Samuel leaves the computer

Zatim je Artur Semjuel pustio računar

to play itself.

da igra protiv sebe.

It plays itself. It collects more data.

Igrao je. Sakupljao je više podataka.

It collects more data. It increases the accuracy of its prediction.

Sakupljajući više podataka, povećavao je tačnost svog predviđanja.

And then Arthur Samuel goes back to the computer

Zatim se Artur Semjuel vratio do računara.

and he plays it, and he loses,

Igra, i gubi.

and he plays it, and he loses,

Igra, i gubi.

and he plays it, and he loses,

Igra, i gubi.

and Arthur Samuel has created a machine

I tako je Artur Semjuel stvorio mašinu

that surpasses his ability in a task that he taught it.

koja prevazilazi njegove mogućnosti u igri kojoj ju je naučio.

And this idea of machine learning

Ova ideja mašinskog učenja

is going everywhere.

se širi na sve strane.

How do you think we have self-driving cars?

Šta mislite, odakle nam samoupravljajuća vozila?

Are we any better off as a society

Da li napredujemo kao društvo

enshrining all the rules of the road into software?

ubacivanjem svih pravila vožnje u softver?

No. Memory is cheaper. No.

Ne. Memorija je jeftinija. Ne.

Algorithms are faster. No. Processors are better. No.

Algoritmi su brži. Ne. Procesori su brži. Ne.

All of those things matter, but that's not why.

Sve to je bitno, ali ne zbog toga.

It's because we changed the nature of the problem.

Nego zato što smo promenili koren problema.

We changed the nature of the problem from one

Promenili smo prirodu problema

in which we tried to overtly and explicitly

od one u kojoj smo direktno

objasnili računaru kako da vozi,

explain to the computer how to drive

to one in which we say,

do one u kojoj kažemo:

"Here's a lot of data around the vehicle.

"Evo ti mnogo podataka u vezi sa vozilom.

You figure it out.

Shvati sam.

You figure it out that that is a traffic light,

Shvati da je ovo svetlo na semaforu.

that that traffic light is red and not green,

Da je crveno, a ne zeleno.

that that means that you need to stop

Da to znači da moraš da staneš,

and not go forward."

a ne da nastaviš."

Machine learning is at the basis

Mašinsko učenje je u osnovi

of many of the things that we do online:

mnogih stvari na mreži.

search engines,

Pretraživači,

Amazon's personalization algorithm,

Amazonov personalizovani algoritam,

computer translation,

računarsko prevođenje,

voice recognition systems.

sistemi za prepoznavanje glasa.

Researchers recently have looked at

Istraživači su skoro posmatrali

the question of biopsies,

problem biopsije.

Biopsije raka.

cancerous biopsies,

and they've asked the computer to identify

Pitali su računar da ustanovi

by looking at the data and survival rates

posmatrajući podatke i stopu preživljavanja,

to determine whether cells are actually

da odluči da li su ćelije zapravo

cancerous or not,

kancerogene ili ne.

and sure enough, when you throw the data at it,

Zasigurno, kada ubacite podatke,

through a machine-learning algorithm,

pomoću algoritma mašinskog učenja,

the machine was able to identify

mašina je postala sposobna da prepozna

the 12 telltale signs that best predict

12 znakova koji najbolje predviđaju

that this biopsy of the breast cancer cells

da je biopsija raka ćelija dojke

are indeed cancerous.

zaista zahvaćena rakom.

The problem: The medical literature

Problem?

only knew nine of them.

Medicinska literatura je poznavala samo devet od njih.

Three of the traits were ones

Tri od tih simptoma su bili oni

that people didn't need to look for,

koje ljudi nisu trebali da traže,

but that the machine spotted.

ali ih je mašina uočila.

Now, there are dark sides to big data as well.

Ali, postoji loša strana velikih podataka.

It will improve our lives, but there are problems

Unaprediće naše živote,

that we need to be conscious of,

ali postoje problemi kojih moramo biti svesni.

and the first one is the idea

Prvi od njih je ideja

da možemo biti kažnjeni za predviđanja,

that we may be punished for predictions,

that the police may use big data for their purposes,

da policija može koristiti velike podatke u svoje svrhe,

a little bit like "Minority Report."

nešto poput fima "Suvišni izveštaj".

Ovaj izraz zovemo sposobnost predviđanja

Now, it's a term called predictive policing,

or algorithmic criminology,

ili algoritamska kriminologija,

and the idea is that if we take a lot of data,

i ideja je da ako uzmemo mnogo podataka

for example where past crimes have been,

npr. mesta prošlih zločina,

we know where to send the patrols.

znamo gde da pošaljemo patrole.

That makes sense, but the problem, of course,

To ima smisla, ali problem je, naravno,

is that it's not simply going to stop on location data,

u tome što se neće završiti samo na podacima o lokaciji.

it's going to go down to the level of the individual.

Ići će do ličnog nivoa.

Why don't we use data about the person's

Zašto ne koristimo podatke

high school transcript?

o nečijim ocenama iz srednje škole?

Maybe we should use the fact that

Možda da iskoristimo činjenice

they're unemployed or not, their credit score,

o zaposlenosti, o kreditnom stanju,

their web-surfing behavior,

o ponašanju na internetu,

whether they're up late at night.

da li su budni noću.

Their Fitbit, when it's able to identify biochemistries,

Ako njihov Fitbit može da prepozna njihove biohemijske parametre,

will show that they have aggressive thoughts.

pokazaće kada imaju agresivne misli.

Možemo imati algoritme

We may have algorithms that are likely to predict

koji bi mogli predviđati šta ćemo uraditi,

what we are about to do,

and we may be held accountable

i mogu nas smatrati odgovornim

before we've actually acted.

pre nego što delamo.

Privacy was the central challenge

Privatnost je bila centralni izazov

in a small data era.

u eri malih podataka.

In the big data age,

U danima velikih podataka,

the challenge will be safeguarding free will,

izazov će biti zaštita slobodne volje,

moral choice, human volition,

moralnih izbora, ljudske volje,

human agency.

ljudske odlučnosti.

There is another problem:

Postoji još jedan problem.

Big data is going to steal our jobs.

Veliki podaci će nam ukrasti poslove.

Veliki podaci i algoritmi će izazvati

Big data and algorithms are going to challenge

white collar, professional knowledge work

kancelarijske, visoko obrazovane radnike

in the 21st century

dvadeset prvog veka

in the same way that factory automation

slično kao što su automatizacija

and the assembly line

i pokretna traka

challenged blue collar labor in the 20th century.

izazvale radničku klasu u 20. veku.

Think about a lab technician

Setimo se laboratorijskog tehničara,

who is looking through a microscope

koji pod mikroskopom posmatra

at a cancer biopsy

biopsiju raka

and determining whether it's cancerous or not.

da bi zaključio da li je zahvaćena rakom.

The person went to university.

Ova osoba je završila fakultet.

The person buys property.

Ona kupuje imovinu.

He or she votes.

On ili ona glasa.

He or she is a stakeholder in society.

On ili ona je član društva.

And that person's job,

Posao ove osobe,

as well as an entire fleet

i celog niza stručnjaka

of professionals like that person,

kao što je ova osoba,

is going to find that their jobs are radically changed

shvatiće da se njihov posao znatno menja

or actually completely eliminated.

ili da će potpuno nestati.

Now, we like to think

Volimo da mislimo

that technology creates jobs over a period of time

da će vremenom tehnologija praviti poslove

after a short, temporary period of dislocation,

iza kratkog, privremenog doba dislokacije,

and that is true for the frame of reference

što je i tačno za taj referentni okvir

with which we all live, the Industrial Revolution,

u kom svi živimo, industrijsku revoluciju,

because that's precisely what happened.

jer tako se tačno i dogodilo.

But we forget something in that analysis:

Međutim, u toj analizi zaboravljamo

There are some categories of jobs

da postoje kategorije poslova

that simply get eliminated and never come back.

koje će jednostavno nestati i neće se vratiti.

The Industrial Revolution wasn't very good

Industrijska revolucija nije bila dobra

if you were a horse.

ako ste bili konj.

So we're going to need to be careful

Dakle, moramo biti pažljivi,

and take big data and adjust it for our needs,

i moramo velike podatke prilagoditi našim potrebama,

our very human needs.

našim ljudskim potrebama.

We have to be the master of this technology,

Moramo biti gospodari tehnologije,

not its servant.

a ne njene sluge.

We are just at the outset of the big data era,

Na samom smo početku doba velikih podataka,

and honestly, we are not very good

i iskreno, za sada ne rukujemo dobro

at handling all the data that we can now collect.

podacima koje sada možemo da prikupimo.

It's not just a problem for the National Security Agency.

To nije problem samo Državne bezbednosne agencije.

Businesses collect lots of data, and they misuse it too,

Firme sakupljaju dosta podataka, i takođe ih ne koriste dobro,

and we need to get better at this, and this will take time.

moramo ovladati time, a za to je potrebno vreme.

It's a little bit like the challenge that was faced

Podseća na situaciju

by primitive man and fire.

kada se primitivni čovek suočio sa vatrom.

This is a tool, but this is a tool that,

To je alat, ali alat koji će nas opeći

unless we're careful, will burn us.

ako ne budemo pažljivi.

Big data is going to transform how we live,

Veliki podaci će promeniti naš način života,

how we work and how we think.

način rada i razmišljanja.

It is going to help us manage our careers

Pomoći će nam da organizujemo svoje karijere

and lead lives of satisfaction and hope

i da živimo zadovoljno i sa nadom,

and happiness and health,

u sreći i zdravlju.

but in the past, we've often looked at information technology

Ranije smo često od informacionih tehnologija

and our eyes have only seen the T,

gledali samo u T,

the technology, the hardware,

u tehnologiju, u hardver,

because that's what was physical.

zato što je to ono što je opipljivo.

We now need to recast our gaze at the I,

Sada moramo da bacimo oko na I,

the information,

na informacije,

which is less apparent,

na ono manje uočljivo,

but in some ways a lot more important.

ali na određeni način mnogo bitnije.

Humanity can finally learn from the information

Čovečanstvo konačno uči iz informacija

that it can collect,

koje može da prikupi,

as part of our timeless quest

kao deo našeg vanvremenskog zadatka

to understand the world and our place in it,

da shvatimo svet i naše mesto u njemu

and that's why big data is a big deal.

i zato veliki podaci jesu velika stvar.

(Applause)

(Aplauz)