My colleagues and I are fascinated by the science of moving dots. So what are these dots? Well, it's all of us. And we're moving in our homes, in our offices, as we shop and travel throughout our cities and around the world. And wouldn't it be great if we could understand all this movement? If we could find patterns and meaning and insight in it. And luckily for us, we live in a time where we're incredibly good at capturing information about ourselves. So whether it's through sensors or videos, or apps, we can track our movement with incredibly fine detail.
Moje kolege i mene fascinira nauka o pokretnim tačkama. Kakve su to tačke? Pa, to smo svi mi. Krećemo se u našim domovima i kancelarijama dok kupujemo i putujemo po gradovima i svetu. Zar ne bi bilo sjajno da možemo da razumemo sve ove pokrete? Da imamo uvid u njihove šeme i značenja. A srećom po nas, živimo u vremenu u kom nam neverovatno dobro ide da pratimo informacije o sebi. Bilo da je to putem senzora, snimaka ili aplikacija, možemo vrlo detaljno pratiti naše pokrete.
So it turns out one of the places where we have the best data about movement is sports. So whether it's basketball or baseball, or football or the other football, we're instrumenting our stadiums and our players to track their movements every fraction of a second. So what we're doing is turning our athletes into -- you probably guessed it -- moving dots.
Ispada da je jedno od mesta gde imamo najviše podataka o pokretu sport. Bilo da je to košarka, bejzbol, ili američki fudbal ili onaj drugi fudbal, postavljamo instrumente na stadione i igrače da bismo pratili njihovo kretanje u svakom deliću sekunde. Ono što radimo je da pretvaramo naše sportiste u - verovatno već pretpostavljate - pokretne tačke.
So we've got mountains of moving dots and like most raw data, it's hard to deal with and not that interesting. But there are things that, for example, basketball coaches want to know. And the problem is they can't know them because they'd have to watch every second of every game, remember it and process it. And a person can't do that, but a machine can. The problem is a machine can't see the game with the eye of a coach. At least they couldn't until now. So what have we taught the machine to see?
Tako imamo gomilu pokretnih tački, i slično većini neobrađenih podataka, teško je raditi sa njima, a nije baš ni zanimljivo. Ali postoje stvari koje, na primer, košarkaški treneri žele da znaju. A problem je što ne mogu da ih znaju zato što bi morali da gledaju svaki sekund svake utakmice, sve to zapamte i obrade. A čovek to ne može, ali mašina može. Problem je što mašina ne vidi utakmicu očima trenera. Bar do sada nije mogla. Šta smo to naučili mašinu da vidi?
So, we started simply. We taught it things like passes, shots and rebounds. Things that most casual fans would know. And then we moved on to things slightly more complicated. Events like post-ups, and pick-and-rolls, and isolations. And if you don't know them, that's okay. Most casual players probably do. Now, we've gotten to a point where today, the machine understands complex events like down screens and wide pins. Basically things only professionals know. So we have taught a machine to see with the eyes of a coach.
Pa, počeli smo sa osnovama. Naučili smo je stvarima kao što su pasovi, šutevi i skok pod košem. Stvarima koje zna većina prosečnih obožavalaca. A potom smo prešli na malo komplikovanije stvari. Događaje poput postapova, pik end rola i presinga. Ako ne znate te pojmove, u redu je. Većina prosečnih igrača zna. E, sad, došli smo do tačke da danas mašina razume kompleksne događaje poput akcija i vajd pinova. Praktično, ono što znaju samo profesionalci. Zapravo, naučili smo mašinu da vidi očima trenera.
So how have we been able to do this? If I asked a coach to describe something like a pick-and-roll, they would give me a description, and if I encoded that as an algorithm, it would be terrible. The pick-and-roll happens to be this dance in basketball between four players, two on offense and two on defense. And here's kind of how it goes. So there's the guy on offense without the ball the ball and he goes next to the guy guarding the guy with the ball, and he kind of stays there and they both move and stuff happens, and ta-da, it's a pick-and-roll.
Kako smo ovo uspeli? Ako pitam trenera da opiše nešto poput pik end rola, on mi objasni šta je to, i ako bih to ubacio u algoritam, izgledalo bi užasno. Stvar je u tome što je pik end rol u košarci ples između četiri igrača, dva u napadu i dva u odbrani. I evo kako to otprilike ide. Jedan momak je u napadu bez lopte i on ide do momka koji čuva drugog momka sa loptom i otprilike ostaje tamo, obojica se pomeraju i nešto se dešava, i ta-da, to je pik end rol.
(Laughter)
(Smeh)
So that is also an example of a terrible algorithm. So, if the player who's the interferer -- he's called the screener -- goes close by, but he doesn't stop, it's probably not a pick-and-roll. Or if he does stop, but he doesn't stop close enough, it's probably not a pick-and-roll. Or, if he does go close by and he does stop but they do it under the basket, it's probably not a pick-and-roll. Or I could be wrong, they could all be pick-and-rolls. It really depends on the exact timing, the distances, the locations, and that's what makes it hard. So, luckily, with machine learning, we can go beyond our own ability to describe the things we know.
To je takođe, jedan primer užasnog algoritma. Tako, ako igrač koji vrši udvajanje - on se zove bloker - prilazi blizu, ali se ne zaustavlja, to verovatno nije pik end rol. Ili ako se zaustavi, ali se ne zaustavi dovoljno blizu, to verovatno opet nije pik end rol. Ili, ako priđe blizu i zaustavi se, ali to uradi pod košem, to verovatno nije pik end rol. Ili ja možda grešim, možda je sve to pik end rol. To zaista zavisi od preciznog tajminga, udaljenosti, lokacije, i to je ono što otežava stvari. Srećom, sa mašinskim učenjem, možemo da idemo dalje od naše sposobnosti da opišemo stvari koje znamo.
So how does this work? Well, it's by example. So we go to the machine and say, "Good morning, machine. Here are some pick-and-rolls, and here are some things that are not. Please find a way to tell the difference." And the key to all of this is to find features that enable it to separate. So if I was going to teach it the difference between an apple and orange, I might say, "Why don't you use color or shape?" And the problem that we're solving is, what are those things? What are the key features that let a computer navigate the world of moving dots? So figuring out all these relationships with relative and absolute location, distance, timing, velocities -- that's really the key to the science of moving dots, or as we like to call it, spatiotemporal pattern recognition, in academic vernacular. Because the first thing is, you have to make it sound hard -- because it is.
Kako ovo funkcioniše? Pa, po primeru. Odemo do mašine i kažemo, "Dobro jutro, mašino. Evo nekih pik end rolova, evo nekih stvari koje nisu. Molim te, nađi način da napraviš razliku." I ključ za sve ovo je pronaći svojstva koja joj omogućavaju da to raščlani. Tako, ako bih hteo da je naučim razlici između jabuke i pomorandže, mogao bih da kažem: "Što ne uzmeš boju ili obllk?" Problem koji rešavamo je, šta su to te stvari? Koje su ključne stavke koje kompjuteru omogućavaju da upravlja svetom pokretnih tački? Razumevši sve ove veze sa relativnom i apsolutnom lokacijom, udaljenost, tajming, brzina - to je zaista ključ nauke o pokretnim tačkama ili kako mi volimo da zovemo, spaciotemporalna šema prepoznavanja, akademskim žargonom govoreći. Jer kao prvo, mora da zvuči teško - jer to i jeste.
The key thing is, for NBA coaches, it's not that they want to know whether a pick-and-roll happened or not. It's that they want to know how it happened. And why is it so important to them? So here's a little insight. It turns out in modern basketball, this pick-and-roll is perhaps the most important play. And knowing how to run it, and knowing how to defend it, is basically a key to winning and losing most games. So it turns out that this dance has a great many variations and identifying the variations is really the thing that matters, and that's why we need this to be really, really good.
Ključna stvar, za NBA trenere, nije to da žele da znaju da li je došlo do pik end rola ili ne. Oni žele da znaju kako se odvijao. A zašto je to njima tako važno? Evo malog uvida. Izgleda da je u modernoj košarci ovaj pik end rol možda najvažniji deo igre. Znati kako treba da se izvede, i kako da se odbrani, je suštinski ključ pobede ili poraza u većini utakmica. Tako ispada da ovaj ples ima mnoge varijacije i identifikovanje tih varijacija je ono što je stvarno važno, i zato nam je potrebno da ovo bude baš, baš dobro.
So, here's an example. There are two offensive and two defensive players, getting ready to do the pick-and-roll dance. So the guy with ball can either take, or he can reject. His teammate can either roll or pop. The guy guarding the ball can either go over or under. His teammate can either show or play up to touch, or play soft and together they can either switch or blitz and I didn't know most of these things when I started and it would be lovely if everybody moved according to those arrows. It would make our lives a lot easier, but it turns out movement is very messy. People wiggle a lot and getting these variations identified with very high accuracy, both in precision and recall, is tough because that's what it takes to get a professional coach to believe in you. And despite all the difficulties with the right spatiotemporal features we have been able to do that.
Evo jednog primera. Imamo dva igrača u napadu i dva igrača u odbrani, spremni su da izvedu pik end rol ples. Igrač sa loptom može ili prihvatiti, ili odbiti. Saigrač se može ili saviti ili otvoriti. Momak koji čuva loptu može ići iznad ili ispod. Njegov saigrač može ili da se otkrije ili da igra do kontakta, ili bez kontakta i zajedno mogu ili da se zamene ili napadnu a nisam znao većinu ovih stvari kada sam počinjao i bilo bi divno kada bi se svi pomerali u skladu sa ovim strelicama. To bi umnogome olakšalo naše živote, ali izgleda da su naši pokreti zbrkani. Ljudi se mnogo meškolje i dobijanje ovih identifikovanih varijacija sa veoma velikom tačnošću, i u preciznosti i povlačenju, je teško jer zbog toga je potrebno da imaš profesionalnog trenera koji veruje u tebe. I uprkos svim poteškoćama sa tačnim spaciotemporalnim karakteristikama, mi smo to uradili.
Coaches trust our ability of our machine to identify these variations. We're at the point where almost every single contender for an NBA championship this year is using our software, which is built on a machine that understands the moving dots of basketball. So not only that, we have given advice that has changed strategies that have helped teams win very important games, and it's very exciting because you have coaches who've been in the league for 30 years that are willing to take advice from a machine. And it's very exciting, it's much more than the pick-and-roll. Our computer started out with simple things and learned more and more complex things and now it knows so many things. Frankly, I don't understand much of what it does, and while it's not that special to be smarter than me, we were wondering, can a machine know more than a coach? Can it know more than person could know? And it turns out the answer is yes.
Treneri veruju mogućnostima naših mašina da identifikuju ove varijacije. Došli smo do tačke gde skoro svaki kandidat za NBA šampionat ove godine koristi naš softver, koji je ugrađen u mašinu koja razume pokretne tačke u košarci. I ne samo to, davali smo savete koji menjaju strategije koje pomažu timovima da dobiju veoma važne utakmice, a to je vrlo uzbudljivo jer imate trenere koji su u ligi i po 30 godina i koji su spremni da prihvate savet od mašine. I to je vrlo uzbudljivo, to je mnogo više od pik end rola. Naš kompjuter je počeo sa prostim stvarima i učio sve komplikovanije stvari tako da sada zna dosta toga. Iskreno, ni ja ne razumem dosta toga što on radi, i dok nije toliko teško biti pametniji od mene, pitali smo se, može li neka mašina da zna više od trenera? Može li da zna više od nego čovek? I izgleda da je odgovor da.
The coaches want players to take good shots. So if I'm standing near the basket and there's nobody near me, it's a good shot. If I'm standing far away surrounded by defenders, that's generally a bad shot. But we never knew how good "good" was, or how bad "bad" was quantitatively. Until now.
Treneri žele da im igrači imaju dobar šut. Ako ja stojim blizu koša i nema nikoga u blizini, to je dobra pozicija. Ako stojim daleko okružen odbranom, to je generalno loša pozicija. Međutim, nikad kvantitativno nismo znali koliko je stvarno dobra ili loša. Do sada.
So what we can do, again, using spatiotemporal features, we looked at every shot. We can see: Where is the shot? What's the angle to the basket? Where are the defenders standing? What are their distances? What are their angles? For multiple defenders, we can look at how the player's moving and predict the shot type. We can look at all their velocities and we can build a model that predicts what is the likelihood that this shot would go in under these circumstances? So why is this important? We can take something that was shooting, which was one thing before, and turn it into two things: the quality of the shot and the quality of the shooter. So here's a bubble chart, because what's TED without a bubble chart?
Opet, ono što možemo uraditi, koristeći spaciotemporalne podatke, je da pogledamo svaki šut. Možemo videti: Odakle ide? Pod kojim uglom je od koša? Gde stoji odbrana? Na kojoj su udaljenosti? Pod kojim su oni uglom? Za više odbrambenih igrača, možemo videti kako se igrač kreće i predvideti vrstu šuta. Možemo videti svačiju brzinu i možemo napraviti model koji predviđa koja je verovatnoća da će ovaj šut ući pod ovim okolnostima? Zašto je ovo važno? Možemo uzeti neki šut, ranije gledan kao celina, i podeliti ga na dve stvari: kvalitet šuta i kvalitet igrača. I ovde imamo grafikon sa mehurićima, jer šta bi bio TED bez toga?
(Laughter)
(Smeh)
Those are NBA players. The size is the size of the player and the color is the position. On the x-axis, we have the shot probability. People on the left take difficult shots, on the right, they take easy shots. On the [y-axis] is their shooting ability. People who are good are at the top, bad at the bottom. So for example, if there was a player who generally made 47 percent of their shots, that's all you knew before. But today, I can tell you that player takes shots that an average NBA player would make 49 percent of the time, and they are two percent worse. And the reason that's important is that there are lots of 47s out there. And so it's really important to know if the 47 that you're considering giving 100 million dollars to is a good shooter who takes bad shots or a bad shooter who takes good shots. Machine understanding doesn't just change how we look at players, it changes how we look at the game.
Ovo su NBA igrači. Veličina je veličina igrača a boja je njegova pozicija. Na x-osi, imamo verovatnoću pogotka. Ljudi s leva šutiraju iz teške pozicije, sa desna, iz lakih pozicija. Na y-osi je njihov procenat pogotka. Ljudi koji su dobri su pri vrhu, loši pri dnu. Tako na primer, vidite igrača koji generalno ima 47 procenat šuta, i to je sve što ste znali. Ali danas, mogu da vam kažem da igrač šutira onako kako bi prosečan NBA igrač šutirao 49 posto vremena, i gore je za dva procenta. A razlog što je ovo važno je što ovde ima dosta onih sa 47%. I tako je vrlo važno znati da li je onaj sa 47 kome razmišljate da platite 100 miliona dolara dobar šuter iz teških pozciija, ili loš šuter koji pogađa lake koševe. Mašina ne utiče na naš pogled na igrače, već utiče na naš pogled na igru.
So there was this very exciting game a couple of years ago, in the NBA finals. Miami was down by three, there was 20 seconds left. They were about to lose the championship. A gentleman named LeBron James came up and he took a three to tie. He missed. His teammate Chris Bosh got a rebound, passed it to another teammate named Ray Allen. He sank a three. It went into overtime. They won the game. They won the championship. It was one of the most exciting games in basketball. And our ability to know the shot probability for every player at every second, and the likelihood of them getting a rebound at every second can illuminate this moment in a way that we never could before. Now unfortunately, I can't show you that video. But for you, we recreated that moment at our weekly basketball game about 3 weeks ago.
Tako je pre par godina, u NBA finalu, bila jedna vrlo zanimljiva utakmica. Majami je gubio sa tri razlike, 20 sekundi pre kraja. Bili su na pragu da izgube titulu. Gospodin po imenu LeBron Džejms je ušao i pucao trojku za izjednačenje. Promašio je. Njegov saigrač Kris Boš skače i brani, pruža loptu saigraču, Reju Alenu. On ubacuje za tri. Idu produžeci. Pobedili su. Osvojili su šampionat. To je bila jedna od najuzbudljivijih utakmica. I to što možemo da znamo verovatnoću pogotka svakog igrača u svakoj sekundi, i verovatnoću skoka pod košem u svakoj sekundi može da rasvetli ovaj trenutak na način na koji nikad ranije nismo mogli. Nažalost, sada vam ne mogu pokazati taj snimak. Ali, za vas smo ponovili tu situaciju na našoj nedeljnoj utakmici pre oko tri sedmice.
(Laughter)
(Smeh)
And we recreated the tracking that led to the insights. So, here is us. This is Chinatown in Los Angeles, a park we play at every week, and that's us recreating the Ray Allen moment and all the tracking that's associated with it. So, here's the shot. I'm going to show you that moment and all the insights of that moment. The only difference is, instead of the professional players, it's us, and instead of a professional announcer, it's me. So, bear with me.
I ponovo smo oživeli putanju koja je dovela do tog uvida. I, evo nas. Ovo je Kineska četvrt u Los Anđelesu, park gde igramo svake nedelje, i evo ga ponovo trenutak Reja Alena i svi pokreti u vezi sa tim. I evo šuta. Pokazaću vam taj trenutak i sve uvide u taj trenutak. Jedina razlika je što smo umesto profesionalnih igrača ovde mi, a umesto profesionalnog komentatora, tu sam ja. Pa ćete morati da me istrpite.
Miami. Down three. Twenty seconds left. Jeff brings up the ball. Josh catches, puts up a three!
Majami. Minus tri. Još dvadeset sekundi. Džef donosi loptu. Džoš je hvata, ubacuje trojku!
[Calculating shot probability]
[Računa se verovatnoća pogotka]
[Shot quality]
[Kvalitet šuta]
[Rebound probability]
[Mogućnost odbrane]
Won't go!
Neće ući!
[Rebound probability]
[Verovatnoća odbrane]
Rebound, Noel. Back to Daria.
Brani Noel. Vraća do Darije.
[Shot quality]
[Kvalitet šuta]
Her three-pointer -- bang! Tie game with five seconds left. The crowd goes wild.
Za tri poena - bam! Izjednačenje pet sekundi pre kraja. Publika je u transu.
(Laughter)
(Smeh)
That's roughly how it happened.
Otprilike tako nekako.
(Applause)
(Aplauz)
Roughly.
Otprilike.
(Applause) That moment had about a nine percent chance of happening in the NBA and we know that and a great many other things. I'm not going to tell you how many times it took us to make that happen.
(Aplauz) Taj momenat je imao šansu od devet procenata da se desi u NBA i znamo to kao i mnogo drugih stvari. Neću vam reći iz koliko pokušaja nam je ovo uspelo.
(Laughter)
(Smeh)
Okay, I will! It was four.
Okej, ipak hoću! Četiri puta.
(Laughter)
(Smeh)
Way to go, Daria.
Svaka čast, Darija.
But the important thing about that video and the insights we have for every second of every NBA game -- it's not that. It's the fact you don't have to be a professional team to track movement. You do not have to be a professional player to get insights about movement.
Ali ono što je važno u vezi sa ovim snimkom i uvidima koje imamo za svaki sekund svake NBA utakmice - nije to. To je činjenica da ne morate biti profesionalni tim da bi pratili kretanje. Ne morate biti profesionalni igrač da biste imali uvid u pokrete.
In fact, it doesn't even have to be about sports because we're moving everywhere. We're moving in our homes, in our offices, as we shop and we travel throughout our cities and around our world. What will we know? What will we learn? Perhaps, instead of identifying pick-and-rolls, a machine can identify the moment and let me know when my daughter takes her first steps. Which could literally be happening any second now.
U stvari, ne mora uopšte da se radi o sportu jer se mi krećemo svuda. Krećemo se u našim domovima, u kancelarijama, dok kupujemo i putujemo po gradu ili po svetu. Šta ćemo znati? Šta ćemo naučiti? Možda, umesto identifikovanja pick-and-rolla, mašina može da identifikuje trenutak i da me obavesti kada moja ćerka prohoda. Što bukvalno može da se desi svakog trenutka.
Perhaps we can learn to better use our buildings, better plan our cities. I believe that with the development of the science of moving dots, we will move better, we will move smarter, we will move forward.
Možda možemo bolje da koristimo zgrade, da bolje planiramo gradove. Verujem da ćemo se sa razvojem nauke pokretnih tački, bolje kretati, pametnije kretati, kretati napred.
Thank you very much.
Hvala vam mnogo.
(Applause)
(Aplauz)