Big Data and the Rise of Augmented Intelligence: Sean Gourley at TEDxAuckland

so remember the day when I got the first computer and my dad brought it home it was an old apple 2e and the best thing about that computer was of course the computer games on it and one of my favorite games was a game called Galaga and some of you may remember this game you're a little spaceship and you have to dodge the aliens and they drop little bombs on you and you have to kind of do it faster and faster as levels go up and up and I got pretty good at this but I'd always lose the computer would always be able to think faster than me it's reactions were simply too good the computer would win and as I get a little bit older I realize that the reason for this was that the human brain can only think at about 100 milliseconds anything beyond this we can't physically do and you and the computers of course can go a lot lot faster so as humans we can't beat computers at reaction games but of course we we think strategically we're smarter than that so we try and take on computers at the game of chess and this is Kasparov 1997 and the whole world is watching the best human chess player take on the best chess playing machine IBM's deep blue and we know now that the computers can beat humans at chess Kasparov loses and for many people it was a story of artificial intelligence through sheer computational horsepower being able to out think the brightest minds of our generation but Kasparov wasn't so sure about this Kasparov started to realize of course I'm going to get bitten by machine but what if I teamed up with the machine what if I started to work together with the machine and he invented a new style of chess called freestyle chess and in freestyle chess you can team up two to two grandmasters can work together a grandmaster can work with a machine a grandmaster can even get on psychedelic drugs if they think it will improve their game and the first World Championships were held in 2005 48 teams showed up and they played each other over the course of a few months on the Internet and they're broadly speaking two categories of teams the headless team where you're getting advanced computers making all the moves because they were smarter than the humans and on the other side were the centaurs the man half machine where grandmaster's would team up with large computers in the hope that that would give them the edge and as the games went on one thing became clear the center started to dominate and this wasn't what we expected but it was demand plus the machine that were able to beat the machines and they dominated so much that in the final four they were all centaurs and they're all Russian grandmasters with industrial scale computers except for one team called Zacchaeus Zacchaeus remained a mystery they kept their identity hidden but they played good chess they played aggressive chess they played creative chess they played a little bit unorthodox chess but ultimately they played winning chess and people were abuzz around the internet as to who this team's our guess was and the rumor that was going was that it was Kasparov and Kasparov had teamed up with some Russian Russian grade military machine and that was the thing that was going to win but the truth of course is a lot more surprising there were two amateur players most definitely not grandmasters a soccer player a soccer coach indeed and a database administrator and they were using three consumer grade computers one of which said borrowed from the dead and they've beaten the best machines in the world and the best grandmasters in the world and what we learned from this is that if you know when to listen to the Machine and when to listen to the human and if you know when you've got two machines arguing which one you should go with Zacchaeus was driving the machines and it wasn't artificial intelligence for them it was augmented intelligence and augmented intelligence I believe is an exciting and fascinating tool that we as a human species can use to go forward in the next millennium and quite briefly augmented intelligence on the one side you've got a very complex world and we use mathematics to simplify it and then on the other side we've kind of got the human version 1.0 and we use visualization to enhance our natural cognitive ability and somewhere in the middle of that we get augmented intelligence and it's being used today to solve some of the most difficult problems in the world weather prediction is an incredibly difficult problem you have chaotic systems hot and cold turbulent airflow and you're trying to model not just today but three four days in advance and for a lot of you sampled you know weather prediction they haven't got very good at that but the actual reality is over the last 50 years they've gone from about twenty to twenty percent accuracy on a 36-hour forecast to now almost eighty percent this has been a huge triumph of predictive modeling and it's been thanks to largely to machines some of the biggest machines in the world but of course is also humans when you add a human on to these models they end up with a 25 percent improvement and the models predictive capability humans and machines coming together to solve very very difficult problems that neither one could solve by themselves and so when we study this we need to study the interface between humans and machines in particular we need to understand this machine as human brain and one of the things if you're an expert of those you that know experts they're pretty good at they have this kind of sense of intuition and this intuition is almost this magical quality but they always seem to know the right thing to do and we can actually we can actually sit down and measure this and a group just last year out of Japan started to pull apart the neural circuitry of intuition and they did this by taking two groups of chess players professional expert chess players and amateur chess players and they stuck them into an fMRI machine to measure the blood flow through their brain and they just flashed up an image of a chess board but only for one second and then he did for one second because they wanted to bypass the conscious part of the brain and get into the subconscious part of the brain and what they found was in the experts there were two parts of the brain that lit up that didn't light up for the amateurs and the first part of this brain was the brick unius located up and towards the back and it was 2 times bigger response for the experts and this is really interesting because the Baku Gnaeus is involved in Vizio spatial reasoning basically pattern matching and what this result shows is that although the image was only shown for one second the experts were able to and what the pattern was subconsciously the amateurs they just saw noise the next thing they did they said two seconds what's the best move that you should make next and then the people had to choose the best move and when they did this again the experts had a part of the brain light up the amateurs didn't and this is the quartet nucleus part of the brain associated with learned response functions habits and long-term memory and the interesting thing here is it didn't just light up but it lit up in a very interesting way the stronger the signal was and the quartet nucleus the more likely the best move was correct the experts had this internal hot-and-cold that we can actually measure and so the neural circuitry for intuition you start on the one side you've got your visual cortex then the pro cuneus abstracts that image and then that matches it up in the quarter nucleus against the learned responses of the long-term memory and the expert chess players over many many many years and countless hundreds of games they build up this neural circuitry and they build it up and they perfect it and over ten thousand hours that have developed something pretty special and what it is that's special is it allows the computational part of the game to take place in their subconscious and the strategic part that can be left for the consciousness and so sometimes when it seems like an expert to step ahead of you it's because they are a step ahead of you but they've had to work bloody hard to get there which brings me to quit which is also bloody hard work the company started two years ago two and a half years ago now with the tagline augmenting our ability to perceive this complex world and we think that our 10,000 hours a little bit long to become an expert in something and so we want to try and build software that does a lot of that subconscious lifting of computational processing for you and I want to show a little bit of that to you today and I want understand the space industry and I want to take you into the world of global space industry and much like our play little games with the computer on Galaga this is the real deal now so it's kind of exciting now if you go into if you go into google and type in space industry into the news you'll get four thousand two hundred articles back which is great and Google says I've done it in point one seven of a second and you're like fantastic and you realize it'll take me two and a half weeks to read all this and it's a little bit of a problem because Google is only looking at keywords but you can actually train computers with natural language processing to find out all the entities and their relationships and you can do this for all the documents and so instead of being stuck with a big pile you can start to build something that's a little more interactive something is a little more human and I like to show you that to you now this is the kwid software running and you've got five thousand articles there and each article as an individual story but of course the articles don't exist in isolation they're connected to each other they connected when they share the same companies or the same people and of course you start to also have some articles that are more connected than other ones so you can kind of change the size and now a little bit of texture and structure starts to emerge so we can apply a physics engine to this now to get a little spatial resolution and so the similar stories will start to come together and the different stories will start to spread apart and all of a sudden we start to see a little bit of space and we see the stories there and we can put links on and looks always biological and it kind of is because it's an evolving conversation that we're monitoring and we can dive in now and actually zoom in and see that the different parts of this map except it's not a map whether a cities it's a map now we're technologies and concepts are related to each other and we can see NASA coming across we can see some of their deep space missions and we can see that coming across as we zoom and we see the European industry and we come down through the middle and we see things like the commercial space debris and we also see cyber warfare and now down on to the bottom of real GMD kill vehicles and the US missile defense shield trying to stop ICBMs coming across their borders and the computer is processed all this and we as a human can now interact with it and this is a big deal and we can also get a little more resolution on this so we can start to find the names of the clusters and we see there we've got a few interesting ones we've got the NASA budget and of course NASA you don't talk about NASA without talking about the budget because they're undergoing severe cuts and that's interesting because the US military is much larger and it's down there and they're also connected to things like Earth Observation and commercial satellite imagery but by far and away the largest part of the spoke level space industry is the thing we tend to forget which is satellites for communication GPS your broadband TV coming in and this is really we think about space we really think about a world of data and so when you've got this here it's as though you've got the map of an expert and if you would ask an expert to make this map they might sketch something out like this on the back of an envelope they try and replicate what was in their head for you to see but of course we can do a little better than that because napkin drawings nice but 3ds better and you haven't went to work ten thousand hours for this either so we've got the whole space industry with all the connections you see the central part of the European industry connect to the US military up to the Commercial Crew and SpaceX and satellite broadband and this is at your fingertips and this is pretty damn cool and so we can apply zoom back out and apply a coarse-grained filter to it and it splits it out the algorithms decided there are two main clusters going on here and you can kind of see it's United States and the rest of the world and the computer said there there are two things going on here but what's really interesting is what's happening at the borders and you see one down by the US military border and you see another one up at the Commercial Crew border so let's zoom in and take a little look at what's going on there and we zoom in and we can see this first cluster here is about Export Control Agency and reform and Obama looking to lift some of their control barriers that prohibits the exportation of space technologies and the reason largely that this division exists is because the Cold War and the Cold War the US said I'm not going to share any of my space technology but now they're starting to lift that because they want access to the highly lucrative launch vehicles that take you up and commercialize the coming backwards and forwards we go up to the top now and we see another one we see the u.s. chief administrator of NASA Bolden going off to China and he's going off there to discuss human spaceflight and this is fascinating because NASA of course is broke and they need to get up to the space station so they're talking to China because China can do that and it's trying to becomes a bigger player on this landscape the landscape will change and we can monitor it as it moves along and we can do this not just for topics but we can also take companies and at the center here we've got Space Exploration Technologies Elon Musk's company that just delivered cargo to the ISS and is connected very strongly to Boeing but also up to air in space and we can throw this up into 3d now and this is where it gets pretty damn cool because at the top we've got SpaceX and red we've got its first degree connections a little bit lower down a yellow and on the bottom we've got the secondary connections and green and you can again go on and explore and immerse yourself but at this point you've got thousands of objects and tens of thousands of connections and there's no way the human brain can do this unaided and at this point this is really what augmenting intelligence is about that you as a relative novice can sit down log in and upload an entire specs ecosystem into your mind so you can make better strategic decisions with the rest of your cognitive processing power I think this is pretty damn cool but you can do this not just for space data of course you can do it for any type of data you can do it for the patents that Apple holds or you can do it for a merger and acquisition activity from large companies but you can also do it to monitor things like political movements and as you see something now like Occupy Wall Street and Occupy Wall Street as it emerges the global phenomena sort of had this chaotic feel to it it was almost as if there was no structure there but of course if you throw all the documents to meet you Occupy Wall Street up you get something a little bit different then you see up here now we've got Occupy Oakland being very different from Occupy New York representing the political schism between the two elements they're connected of course through police brutality which tends to amplify the messages and then the political discussion down the bottom is one homogeneous group because they haven't decided where they want to go it's connected through by inequality and the political response of taxation and we can see that not just as a snapshot but we can see this actually evolved through time and when you think about what this means you can see clusters emerge you can see ideas fragment you can see the political movement start to get its voice and you can see it be amplified by the brutality of the police on the streets and when you look at this you start to realize that if you can understand this with an advanced level of intelligence perhaps you can start to manipulate it and if you can start to manipulate it it also start to realize the power of the software for good or for bad we live in a very complex world there are 7 billion Minds now and those 7 billion Minds have created a world that not one of them can understand and yet we still have to make decisions we have to decide whether or not the sent troops to Iraq and we have to decide what to do about climate change and we have to decide how to deal with a global financial market that doesn't want to stay still and we're not going to be able to do it with the brain that we were born with we're going to need augmented intelligence and I'm pretty excited to be building a small part of that because I think it has big potential and thank you for sharing it with me

Channel: TEDx Talks

Views: 199,078

Rating: 4.9182663 out of 5

Keywords: tedx, Sean Gourley, Augmented Intelligence, Big Data, Dr Sean Gourley, Dr. Sean Gourley, TEDxAuckland, TEDxAuckland 2012, cyborg, TED Talks, TEDx NZ, New Zealand, physicist, TED fellow, NASA, Pentagon, Quid, Quid software, Business Intelligence, Chess, Neural Circuitry, Intuition, 3D visualization, singularity, strategic thinking, user interface, data visualization, network analysis, network theory, open source intelligence, complexity, Auckland, TEDx Auckland, 2012, NZ

Id: mKZCa_ejbfg

Length: 17min 2sec (1022 seconds)

Published: Wed Dec 05 2012