so I'm going to talk about what is possible what works what not I will show you a lot of examples of only projects which we did young last two years here in Germany Switzerland and Austria so it's a hype but it's always good to think back when I studied physics at me in Munich the website entitled is not possible to win chess one year one week later doopy doopy blue has been winning against Kasparov and then changed the chess play and today you can load down this as a app for free so it's a kind of disruptive smart assistant what you see here and this existed all the times machine becoming more intelligent and this is what I'm going to talk about in the next half hour in 2011 we have been this was the date of birth of Jeopardy who saw this video on YouTube please raise your hand so for I I come camp conference a lot of people didn't see it so go there and have a look how over the Watson
computer here in the middle actually has been winning in answering complicated questions out of all type of sport history science against these world champions and the most important point is look at the face of a vet because he understands at this point in time he will never ever ever give win again this game so here this was a major quantum leap in artificial intelligence in terms of natural language processing and answering questions so we talked here this current word about their AI it does not mean that the systems are intelligent but from outside they seem to be intelligent so they can be very damp inside they are learning so a lot of the hype is wrong that they're becoming that learning and when the artificial neural networks becoming deeper and deeper you talk about deep learning and and that's what is going the hype you see here the Gartner hype cycle at the top of the trend you see where the hype is really hot and that's why it's all newspapers in the conference you
did you see deep learning machine learning so and and so ever everybody talks about it and what is then happening that the people tested and say oh it's too expensive it's too slow it does not work and so a kind of realism takes place and then one or two years later it's all over and then actually your company your organization either made it happen and you have an advantage or you are losing against disruptive models so now we'll have a look which of these upcoming technologies in this curve shows all future trends of on earth in all industries in all countries so which are coming in the next five to ten years and you see the red ones are linked to deep learning technologies and that's why you can see AI has a new electricity so you should really think about it it will not disappear and it will be everywhere in everything and so you it's very actually important and generates the general purpose technology and you see what is coming now in the
next five years is the converting conversational user interface augmented data discovery so the the the algorithms will become more intelligent and more active and I show you a project project D beta which we actually did in and generally you can see on YouTube so this is the next version of Jeopardy where the machine called the beta is a bit louder please she will be arguing for the resolution we should subsidize preschool ladies and gentlemen here we go Project debater greetings Harish I have heard you hold the world record in debate competition wins against humans but I suspect you've never debated a machine welcome to the future thank you that we should subsidize preschools we are going to talk about financial issues but not only about them so here you see it's a machine debating the pool and cons of a particular thesis you do usually put public money into police codes and now the verb champion is actually arguing against this topic one of the that's your kind of wizardry as well and you're very
good at it so ladies and gentlemen please welcome our debater arguing against the resolution we should subsidize preschool Irish Natarajan well thank you very much everybody it's a pleasure to be here for this historic event and it certainly was a pleasure to listen to project debater there was a lot of information in that speech and lots of facts and lots of figures the problem though is the reality of subsidizing preschools is one which does not deal with the underlying problems in society it is one which often makes those worse and in the end is very little more than a politically motivated giveaway to members of the middle class so now what was happening you saw the little boys jumping from the machine debater she were listening and now thinking how to answer now this speech about not to put public money in preschool a snippet from around two rebuttals and again each side will have four minutes to respond to the opponent's arguments first with her rebuttal Project debater arguing for the resolution we should
subsidize preschool thank you again for starters I sometimes listen to opponents in wonder what do they want would they prefer poor people on their doorsteps begging for money would they live well with poor people without heating and running water giving opportunities to the less fortunate should be a moral obligation of any human being and it is a key role for the state to be clear we should find the funding for preschools and not rely on luck or market forces this issue is too important to not have a safety net so what you see here that the machines are becoming more intelligent more proactive and even go deeper in the semantics of what does it mean so that you can extract arguments poor and for a certain title you can do your own project debater projects look in the internet what I know is showing you is what is today out in the market Enda with for now or use cases so you know superintelligence not not a real general artificial intelligence not available but never
artificial intelligence available and here the opportunities are highly underestimated so here the companies your organization you need to be quick to really follow the trend and the four are here see and right the Watson assistant see and recognized visual recognition service generate knowledge and Watson discovery serves and speak and listen speech to text so AI based services help you to automate simple intellectual tasks like the tractor on the field a physical work now here simple intellectual tasks and I'm showing you now examples on these four live on your use cases so the first is read and write the Watson assistants here you see a internet page you can check out I listed here from the inter insurance and they sell a sewer automatic robot based on what's an assistant here teeth insurance so you you can answer question here in German language if you are regularly at the dentist you get a discount bonus F and you buy it and the board of the company claims I'm actually selling a hundred percent more than before
you so you can check it out and by Saturday night at inter insurance funny they call it if the empathic physicians assistants the empathic insurance assistant so you might want to check out I like the teacher jumping but they call it if CN recorgnize here actually visual recognition and services come in place here startup called any line from Austria they are doubling every year based on visual deep learning capabilities like year the vin police you can recognize the matriculation plenty of this B&W or your passport or other cars and their help actually to automate the recognition of visual patterns we doing a project called as mercedes now for the last two and a half years with Daimler together its life out look at the car outside for all a Class E Class C Class S car if you buy a Mercedes you get it who who is driving a Mercedes here in the home so you can check it out or take a rental car and you will get asthma sailors and here you can
talk to it I can show it you I have it on my mobile and it will answer to you for instance how to save fuel or you Hoover in the car because it's a rental car you don't know all the buttons and lights Hoover and it's recognized four and twelve speedometer and then you want to know what is the speedometer about so you click on speedometer because as a yellow light you don't know what does middle yellow light and then you see a yellow light is the yellow control brake lamp and you learn the do I need to stop can I go on so the the people like this because it's you can talk you can write you can click on these ships and it explains you the car more than 300 opened different different topics are covered in technical chat pod and we have been in this project against the Google Amazon the Microsoft of the word and Microsoft and diners today running up the planning up to 20 projects on this AI platform
I show you what the word wide responsible of customer service of Daimler is telling on YouTube here in the discussion in a material sitting apparently are about as materials so what I have is a status our first virtual assist in the car and also outside the car so whatever you want to know about your car how can I charge my telephone what kind of Drive motors am I able to choose you you are able now to simply ask your virtual cognitive assistant we integrated the augmented reality for the first time in a digital product so what you can do is you can simply choose here you can language and you see for example that whatever you have in terms of manuals here what to choose what kind of button is for what type of things yeah whatever it is ask Mercedes will be able to answer all your questions I love them he answers also questions in terms of whether he's still available you would like to marry you so it's not only a check
but it's a true virtual assistant with all the other let me say Alexa whole electrical sander Cooper homes yeah we're able let me say they're also linking our service our product it's only the first and let me say at the beginning of our digital transformation love it thank you so much thank you so you heard the last sentence it's only the beginning of our digital transformation for a time but BMW folks won this is a matter of survival to have a smarter system once a car our autonomous you need to be able to talk to it without order driver so this is a very hard fight and in the outer in the innovation half of diamond look at Autobahn it's called there are always 50 young startups actually building new models and there is a model how you can be in the car after two years of time get a lot of money and a lot of this is based on our Watson platform using the services I'm showing you here so the third is
generate and navigate knowledge this is not so surely but actually even probably more powerful because you can read this AI at pre-trained deep learning our government insight PDF documents and what I show you is a new capability of our Watson discovery service called smart document understand and in the past it was always difficult to delete documents because you could not understand it's a title body text image or our table here now this needs new capability you can talk to the document and say give me the the the country in this table and goes and it can read into a single field of a table imagine you have a company like aliens or telecom with millions of contracts and you want to know show me the contracts which are running out this week and from a million documents you have a hundred contracts from Deutsche Telekom running out this week in a smart tool so this are highly scalable capabilities and this is really what you want to do as a start-up so now speak to
a speech-to-text Roland Berger actually said it's a big new trend speech-to-text I show you a project we are very proud of and it took place and the international space station here you have the wonderful you out of the space station 400 kilometer Heights flying on several times a day and project Simon the cool interactive mobile companion the little football you see here actually worked with Alexander Garcia this captain the German captain of an international space station who worked until last Christmas and we actually implemented all cool capabilities now in one device in Simon to show how you really can build a nice smart companion smart assistant and and I want to share with you this historic moment when on November 15 in the afternoon Simon vocab so think about it like the first arts for this you know who knows Star Wars so you know there are so did so so this is a kind of first episode it's four and so now the moment on November 15 so this current station in Italy we
have been winning this project against NASA and and others so now the moment I is for him to be aware of his surroundings so after which that you detach it please say Simon wake up Simon wake up hello I am your space companion there's much from our TV so these are the two project slivers from a deal allergen space agency and Abbas and maybe you missed what they said this they said this I means how to write history so in a very short period of time in less than two years this the team of some PhD some young guys actually developed a very cool device which is one now operated in international space station helping the astronauts to do biological and chemical instruments experiments so they're each step one two three four five and so I technically I show you how this works so here you have the different services so the eyes are actually here the cameras are several cameras so you can talk to Simon like to a doc and say Simon come
and then Simon recognized my face and it's fully independent here body movement its autonomous flying in space at zero gravity so it's following you and you can say help me to do the experiment here and and then it's coming so these are the eyes the camera the brain is in the thankful data center what's an assistant the ears are speech-to-text so there are multiple microphones listening to what they're hear what you talk and then thinking the answers coming back through the link and so this is now the voice of Simon my mission is to assist you wherever I can and to motivate you no matter which task needs to be done so in a newer version we can actually find out up to twelve different emotional states of the person talking to Simon and it can trigger the voice Simon is using so imagine now you are far flying out to Mars and you are a little bit in a depression mode Simon might talk to you like this I know we are very far
from home what can I do to cheer so that's the emotional markup language with Watson tone analyzer and so this is the kind of stereo trip how you can build smart assistant and scalable tooling for business processes for your companies for your startups and you can operate it in a in a very professional way so for all this you need software and I show you how you can do this and that all this is a very one-click of your you can in one single clip you type in you are in the middle of the AI or catalog of IBM in our Frank for GDR compliant data center and you see all the service side I showed you you can use them one second later and for startups it's for most of people it's for at the beginning fully for free you start you know harder to be installed no suffice to be installed no nothing you just start all right and we see I can see this on a weekly basis I see in in
the big startup centers in Jeremy like burning and Munich I see more and more startups coming on the cloud you can use it and if you need what the computer database here you have the different layers so it's a fully virtualized calculation center from heart to the AI stack here and in particular who is a data scientists and really using himself machine learning deep learning technology and so some are here so then for you guys the Watson studio is the AI development environment we offer and the very cool stuff now since the end of last year it's called Watson studio here you can fully integrate even predictive models from the form like SPSS I log all engines from the correlation were fully integrated with the new artificial networks then you can actually run deploy it and you can manage and on the manage you most of you have seen in the talks why is a lot of high managed don't want to use AI because they don't trust no transparency no trust is available and
Miss Watson open scale we actually opened a capability a pre pre trained pre-existing capability to monitor your AI system for a lot of people ai is a kind of black box look at the handle splat they or other articles si or black box to be difficult dangerous but with a I open scale you get a fully automated monitoring in development and in a home term in a runtime of your statistical significance of your statistical errors you know it's all about probability statistic in neural networks you get this out of and in a kind of where you can say it checks it can check the fairness of your algorithm so you can see it's a balance between women and men young and all is it a good decision why and and you get actually insight into the systemic errors of the algorithm the very nice development we are part of the board of the German government to do develop ethical design criteria and I open scale is here in there and we actually bring it to
the European Commission so that AI in Europe and AI in Germany made in Germany should have this label trust and transparency AI out of Germany so that you can really prove that you know what you do and it's not a black box now this can be a cool USP for your startup or your company so to summarize I IBM as a AI partner can offer you these four really very solid characteristic the first we have a very broad set of AI services what I showed you before you can send me an invitation on LinkedIn and and I the regular post directly links to what is new so you are very well informed the second is what I just talked about trust and transparency into AI a open scale the Watson studio in version 2 now and here just let me mention one thing probably how many data centers do we have from Microsoft and Germany remember me they had a calculation set up a structure Telecom terminated right so there's none so if you want
in a lot of industries like finance or health but it's legally binding that you need the data in the country you need the calculation Center GDP a compliant calculation center in the country and we have this in Frankfurt several thousand paying clients up and running now for three years so this is a very solid good base for startups to scale and last but not least big topic on the next conference might be hydrate cloud or multi cloud because in some areas that people want to get their private cloud behind their firewall big companies want this to product and we have this IBM cloud private and what's an assistant on IBM connived ICP we have offerings first offerings fully on your private cloud in your Center this is not maybe not the thing to start for a start-up because it's not really you need really to invest but if this is critical we have here the first solutions out and support fully hybrid and multi cloud solutions so to summarize my advice is to you moves
fast quick and a little baby steps but really go go go because there's nothing about a planning you need to check it out and you will see France in my team I we built a pretty fairly large large genome with several hundred people the best stuff we get are the young people just doing doing doing an event and you then you are surprised what is working so this is really different type of work than when when I was a PhD so yeah intelligent agent has high quality you control the quality it's self-service on the weekend in the night you get it's a very place in people it's augmenting we heard in other talks as well it stays loyal discreet and private the familiar touch point it can actually charge your brand of your company with this new experience the sexiness it's a family familiar touch point increases productivity of course large-scale implementing we always have a business case and it's really heavy serious savings by the way via Deutsche Telekom we do some stuff too
and personal responsibility so this is a kind of how I think about typical scalable solutions in 2019 and I want to summarize what I just talked about so AI is a is a multi- and it will not disappear simple in like intellectual tasks are automated and and support us to do better job and boost productivity and competitiveness AI based goals and business model are emergent quickly you earn better money you get you get better job you have a bright future because you amping up by the way we describe all this in this book with with one chapter is from from the different larger players and one chapter here from us success factors are that you need to collect data you need to really invest hire people and pick a partner so that you have software out-of-the-box available you should not start to install all yourself and pick the right you key use case is we want to find out what the right use cases you can come to our design and thinking centres in Munich
we have a large Watson and AI Center in Hamburg and in Berlin we do this with you in our centers with our teams model-driven and pick a pick a partner which protects your IP pick a partner which protects your data you should not do these people who are later on putting the hand on your data putting your hand on your IP go with this secure GDP are compliant cloud and again go step by step create information and create your local AI ecosystem like here on the conference you can send me invite on LinkedIn or go on the country log thank you for your attention