 So, I think that was a great introduction on artificial intelligence, part of what makes it a complicated space is because people attribute a lot more complexity to it than is currently possible. So what I'm going to talk about a little bit is what I do. Sorry, how do I, how do I go to my slides, so you live? How do I go to my slides? No problem, no problem. Okay, great, thank you. Okay, so let me start a little bit by telling you a little bit about myself. I really see myself as a lifelong student and I think a lot of women who are here also think of themselves that way. I didn't start off in technology, I actually did a lot of literature and Dostoevsky and surrealism and political philosophy and I think that was all very good that I did that. And then I ended up meeting a bunch of very interesting people from Eastern Europe who looked at technology and math as poetry and they really, it was not the way I had been exposed growing up in Pakistan where you just like learned it and you got your equations and you topped your cohort. It was really look at this, look at how amazing this is. And that is how my interest started in technology and then I went on to Columbia and I actually joined the Foo Engineering School and did applied math, applied computational genomics and computer science. I started off as a programmer on Wall Street and at that time, you know, people wondered in my family why I was being a back office programmer because it was such a little respect for technology so very long time ago. Now of course, all those people are sitting in board rooms so we've come a long way and right now I'm doing my PhD which I'm finishing up on smart cities because cities present two things that are interesting to me. One they present systems of systems, lots of data but they also present ethics, privacy, all those concerns that I think are important and moral dilemmas that are important to have when you're talking about artificial intelligence. So a couple of months ago, I met a couple of my co-founder. His name is Faisal, he has a PhD in machine learning from University of Eindhoven and I met him in Pakistan and we said that, you know, we both had a vision that was to use artificial intelligence, use machine learning and data science to really build some amazing products and services and to build a world class consulting firm and incubator. And it was really about letting go and letting people set their imagination free and I keep talking about this, this idea of coming out of the confines of what is traditionally considered technology or traditionally considered analytical thinking or statistics. Great teachers, good teachers will inspire you because they see it as something so creative and which requires periodically leaps of faith which requires a passion for the art of inquiry but at the same time the discipline of scientific reasoning. And when you move especially to the corporate world or the private sector then of course you need to have empathy because without empathy you can't put yourself in the shoes of people from who you are building products. And so we, those are our principles that we really believe in at OAI and that's kind of what we stand by. We are a mix of academics and industry which again is really important in this field in particular, again brings you back to that lifelong learning thing. We have four professors, my co-founder is a professor as well and together we are constantly using very deep expertise. ADDO AI, it means to add in Latin, ADDO and ADDO AI means we just add artificial intelligence and in that way power data driven applications. What I'm going to do today is give you a real taste of what a company like my mind does and I'll start with three kinds of examples. The first one is from Smart Cities. So we always want to start with the problem, we don't want to start with the technology or the data these days if you have an interesting problem to solve or an aspirational product to build you will run into the need for technology and data. So no need to start from there. The problem that we face in Singapore and of course across the world is we want something that is sustainable with 70% of the human population going to be living in cities we are under a lot of infrastructure congestion and pressure. So how do we create a car light manner of urban transportation? Well let's look at what our digitally savvy consumers want. They want a service. They don't necessarily want to own cars. This is true in Singapore. This is true personally for me. We got rid of our own car a couple of months ago and it is true for many of our friends back in the U.S. And the reason is because they want to think more in terms of services than in terms of products. So if you think of Kevin Kelly who I highly recommend his book The Inevitable he talks about the fact that all businesses are essentially service businesses. So we like to own books. Now we can read them on Kindle. We can run them. We like to own music. Now we stream them on Spotify. We like to own cars. Now how can we get a service on an app? How can I have one app that gives me access to all possible kinds of services? So that just with one app I can say I want to go from here to Marina Bay Sands and please let me book bike sharing and then I want to take a bus service and then you know what I want to be healthy so my personal preferences I want to get a mile in every day. So give me the best route and I don't have much of a budget because I'm a student or something like that. So we needed an integrated mobility service where you have an ecosystem of all the different public and private companies. They would come together and they would integrate and we would be able to create a platform. This platform would be kind of like the iTunes of mobility services or the Amazon of mobility services where your buses, your e-scooters, your autonomous vehicles all of them could be found on one app and you would search for and get exactly what you needed. So what happens then is it's called mobility as a service. That is also the topic of my PhD. I studied a case study in Berlin and now we are working with a very large public transportation company here to build that prototype the first of its kind in Asia. What does that involve? Well, what involves is that for citizens it's a simple app. You want to get from A to B in the most comfortable and cost efficient manner but at the back end behind the scenes there's a lot of integration. You have to integrate. You have to make systems talk to each other. You have to have APIs and you have to negotiate these APIs because people are not open traditionally. So you have infrastructure. You need to know where your roads are. You need the maps and then on top of it you need to have transportation providers and this diagram is from a company called WIM Mass Global in Helsinki and I'm actually visiting them next week and they have launched a private company that actually creates mobility as a service. The idea is that you can aggregate all kinds of things in one. So where are we using the machine learning? Well, from the consumer perspective you begin to know the consumer more. You begin to know their habits, their needs, their preferences and more and more you can provide them the right thing at the right time that they need to get from origin to destination. From the city's perspective it's even more interesting because now you're not just a dead weight city that is essentially your buses run on one schedule and your autonomous vehicles on one and your bikes are always stationed in one place. Now you can be responsive. If I feel that things are more congested because I've done some forecasting then I've learned that in certain times of the day or under certain circumstances there's a demand for a certain kind of transportation that I will make more of that available. That is the definition of a smart city. A smart city is a citizen-centric city that essentially responds and in that way seems personalized to all of us. A friend of mine wrote a book called The City is Here for You. So we really want the city to be here for us and the way you elevate that to a smart nation is if something could be that convenient and seamless and the city begins to also offer you education and skills and investment in the right industry. So that's one example, a pretty interesting one that we're working on. So these are opportunities for entrepreneurs in cities. How about in other countries like Africa, Pakistan, India, Bangladesh? I'm very interested in those people as well. So here's the problem. If you've been to a village or any village in Africa or anywhere else you know that these people are very poor. They are therefore called the bottom billion. They often live between $1 and $4 a day and they're very vulnerable to what happens to their crops. Now 40% of Pakistan's economy labor is based on farmers and if it rains or there's an infestation or the soil is not good these people are in bad shape. They don't know what to do. Their entire household with their children, their wives, their parents, they're all disturbed. Their health suffers and we have never been able to give them insurance. Now they won't come and buy insurance. You won't see them, you can't go to a farmer and say you know what, we'll give you life insurance and health insurance for $100 a year. That's too much for them and they don't see the value of that. They'd much rather insure their smartphone than they would buy health insurance. But if you said that for a very specific calamity I will insure something, let's say a pest infestation or something like that and you'll have to pay me a little bit then they might be willing to do that because it's a three month crop cycle. And when you do that how are you able to tell? Well you have to use data. We call it Nica insurance because Nica means small and Punjabi and the reason we do that is because we use a variety of now new tools that we have to estimate the condition of the soil, the health of the crop and we advise them to make interventions. So satellite data is becoming a lot better and more high definition. With that you can do the kind of image processing that we were talking about earlier. You can use metriology data on weather. You can see the soil, you can see the moisture. You can give them, you can do predictions on health. You can understand on weather, you can understand they have 5 million cheap smart phones. You can maybe even get them and advise them and educate them through text messages. When you make interventions at the right time this is very good for the insurance company because the insurance company now doesn't have to make a payout. So the insurance company is willing to take the risk for a year better than 1 million Singaporeans who give $2 a year. They are able to do that and then on the other hand if there is a calamity then because the insurance company has diversified its own risk over millions of farmers it is able to make that payout. And so obviously you are using a lot of data analytics using a lot of machine learning solutions to be able to predict and understand the condition of the farmer and to do these analytics. And so that is a product that we are working on IMDA Pixel Labs a national design center has offered us a place where we are incubating this. So we do two things. We consult and we provide consulting services. We have a lot of machine learning engineers, artificial intelligence experts, we have PhDs and we have obviously software engineers and data sciences and then we also incubate our own products. Now this thing that I am telling you about is not new at all. Nothing that I am telling you is new. Somebody is likely doing it. But in Australia and other places they are using drones and the term for it is precision agriculture. Precision agriculture is when you use drones and they take pictures and not only are they able to real time tell you the state of a particular field but then the same ones can deliver the appropriate pesticide etc as well. The only difference is they are expensive right now. That is why we are counting on having a partnership with the satellite company. Now the satellite imagery is getting better but also drones are getting cheaper. So at the end of it you just see which one is easier for the business case. So there has to be a business case by the way in all of these. And not a research project. When you have a private company we have to make it for the product. For the other thing you have to have a win-win for all the mobility entrepreneurs and the platform provider also needs to take some kind of margin. Now what do artificial intelligence teams look like? They definitely have data. So you need your data scientists and then you need your computer software engineers and then they study these algorithms so you are good at science and you go to statistics and you go to computer science as well. It's basically the intersection of the two. But a lot of it is also life cycle of the data and understanding how to use it and then testing it, training the algorithms. Again, very good talk earlier where we were talking about the fact that a lot of Hoopla is made in the press but 80 percent of the algorithms out there are supervised. They are not unsupervised or they are not learning. Most of them are supervised which means you need to train them on data which is another reason why they are very good at simulating human behavior and therefore putting humans at risk of automation. But the other thing is for very big companies you need data servers and so Baidu, Tencent, Facebook, they have server farms and you need that and you may even need a chip a chip whose only job is and Google has created one as well to process AI computation so that it's faster. So you see a lot of interesting hardware configurations to the chip level even to allow this to happen. Now our clients are not at that point we are not at that point but we know that very big companies are. So for startups you really need to think about what are you bringing to the table, are you putting together a problem and are you identifying a solution in an interesting way because just the raw power of the talent and the computational power that these big companies have makes it very difficult to compete with them unless you have an innovative idea which is where all that creative thinking empathy comes in handy. I'm going to point out healthcare because it's interesting we're trying to get into healthcare now and nudging people to adopt a healthier lifestyle something like this, we're doing this for one of our clients it's not rocket science to imagine that this makes sense we already have apps that do this but you take a facial image from a phone camera, you take your activity in biometric analysis, you integrate that with your preconditions and your health data and there you go you're able to give them some score you can teach them you can build a chatbot around that the interesting thing would be if the chatbot was specifically for this domain so one of the things is coming to chatbots is they don't cementically understand you very well so stateful chatbots, so stateless is every time you talk to a chatbot it thinks you're a new person stateful is it remembers all the conversations it has so at a minimum it should be stateful but even in stateful is it really understanding what you're saying and for that you need some kind of domain semantic ontologies so we had someone who spoke at our artificial intelligence meetup, we hold a popular artificial intelligence meetup in Singapore and he dialed in he's the CEO of Twiggle it's a company, one of the fast rising e-commerce startups from Israel, Ali Baba just invested in them and he showed us how when you search for a black dress on Amazon you get some responses not all of them are good if you search using Twiggle you get 30% more responses aligned with your search because his search algorithm is better because they understand he's able to understand what the person wants and he's created a product ontology for this, that's really what's making him stand out and I think that's really important to look for those opportunities and to really become domain specific in some ways whether it's logistics or banking or others so that's one thing but on the other hand when it comes to healthcare, when it comes to children you also need to be cautionary we always have a doctor who advises us is this legitimate what we're doing because you tell us to do something we'll do it but we need to be careful when we're nudging behavior that we're doing the right thing and so ethics do come into this they should come into this and sometimes you may have a great algorithm and it's working but you have to consider is that from a domain perspective is that the right thing to do in healthcare because there's such a demand for it here I think that's important to keep in mind Babylon Health raised just $60 million for having an AI doctor so when you call National Health Service NHS they often spend a lot of time on the line and they don't really know the answer well but now you can directly talk to this chat bot you don't need to wage, you can give the chat bot your symptoms chat bot will compare it to all the other instances that it knows of and then give you some recommendations when they compare the recommendations of this AI doctor I think to 12 real doctors I believe you can look that up online and they found that it was equivalent at worst so I often give this talk I think you guys are really motivated which is awesome but a lot of people think it's a buzz word, they think it's interesting they don't want to they don't want to really go into it because they think it's too much, it's hard it's really not my field I really hope you will because it's really interesting and it's really fun to do it and there's no reason why you can't and there are lots of people who are doing it in the developing world who are your potential collaborators or your competitors or people you'll manage or who will be managing you so you will run into artificial intelligence there are many many courses you don't have to complete all of them I am ashamed to say that I have started many but I almost never take the exams which is not good but at least I listen to the you know I listen to many of them they are very interesting this one you have to apply for actually I was thinking of doing that after I finished my PhD but this is a self driving car engineer nano degree here's what I find interesting about it it's a nine month course you can do it on audacity you can do it on audacity faster than you can find it in U.S. or anywhere else you know and so why would you wait for academia to to speed up but better than that why do you think it's relevant to people like us because it has been done in collaboration with some of the leading mobility experts who are thinking about self driving cars so they've said yes these are the problems Mercedes is saying this, DD is saying this Auto is saying this and more than that they are saying you know what in the course we guarantee you a job so there is a lot of incentive for people who may not be here in this room you can imagine how driven they are I often give this example that if you can take the course on a little smart phone and you know you've been kind of doing a side job as a poor mechanic in a village in Bangladesh or somewhere you're pretty good at cars I think you get the physics of it and then you're learning all this stuff well then you can really do something interesting with it that boy if you were smart enough would be your collaborator if you were foolish that person would either you know put you out of a job or before the machine does at least or actually not be there when you need him so I think you really we really need to think about who are the people who are acquiring these skills and what do we bring to the table I have to think about that every day in Pakistan they're really smart they're really creative they don't think what do we need to bring to Aisha who's sitting happily in Singapore I think to myself what am I bringing to them and I know how many of you are Singaporeans here but as a Singaporean that continues to be on top of our list of how do we differentiate ourselves how do we show our value to the rest of the world who is driven, motivated and learning so I think this is very good it's always good to have this kind of healthy pressure on one's health I think so education is the key in Singapore Singapore has given $500 to each citizens upgrade his or her skills I'm sure you can get more money if you took any of these courses they're very generous they want to encourage it in fact just yesterday they announced $150 million for AI projects there's something called AI.sg and the whole idea is let's make sure that Singaporean students immigrants companies they all have the facilities that they need and they're going to invest in companies they're also going to focus on financial services and healthcare and smart cities so any of this is of interest to you guys you should definitely look at that as well I would end with this I always end with ethics because I think that's very important it's always important and so everybody is going around talking about this crowd flower asked 200 data scientists what do you think you guys do this all the time what do you think are the issues and so they came up with this obviously the automation displacement of human beings by machines continues to be something all of us are concerned about because people need time to adjust to the older people so we have to support them in this process but also because we design things we embed our own values politics and biases in the things we design so I think it's really important that we're conscious of what we're doing self-awareness is also very important and then they had some other concerns as well in warfare intelligence I was just speaking at a panel on cyber security and artificial intelligence I was talking about artificial intelligence my co-panelist was talking about cyber security and yet a lot of valid concerns about that as well so I think there's a lot to talk about there's a lot to think about that it's a really interesting opportunity for women like yourselves to no matter what stage you are at in your career no matter whether you're in tech or data or not it doesn't matter there's a real joy in learning you should go for it you try to do it the companies that you work with will use it so it can only be to your benefit to acquire this knowledge there are lots of courses out there but you should do it with a joy of inquiry you should do it with a true desire investigative desire because more than anything else this really is about investigating and coming up with some ways to tackle a problem so that way it brings us right back to kindergarten and the things we like to do which is play around with stuff and I think if you go with that attitude and put aside some time you really enjoy it so I look forward to your questions thank you for giving me the time to speak to you today there will be a Q&A session can you share right now oh okay there you go based on the smart cities I think this is a question how do you get bias from say for example the different transportation from segments right I think part of it will be government, public private sector I think to have API built for them you have to get bias from these various segments of people and the second thing will be who would then bear the cost of API construction that's such a good question and the number one issue that you have in such a collaborative platform so you have to convince them they have to see something in it for themselves usually for partners they see a couple of things they see access to more consumers and they would have otherwise they also see data analytics as being useful to them to understand trends and each one usually bears the cost of his or her own API but it's a real negotiation what you should give them what you should take from them data who makes the payment, who owns the customer but I always say that in Smart Cities six, seven years ago when I was just starting out there was this idea that you could have this kind of super net or super government system but you can't Smart City 2.0 is all about companies coming together and plug and play through APIs and you see that in any product or service that you build so they're becoming conscious of this we're working with some Chinese companies and they're just really on board with everything else just consistently impressed both with their vision mega vision and their openness to try and learn and then take to capture the market further but that's a really good question is it like the integration that you're talking about of the different transportation so like let's say Uber like having cars Uber or Grab so what does that app so will it redirect I mean the booking will actually happen through Uber or will it be like through your system so actually when you're when you're using an API you're actually calling their system it doesn't hop onto their app absolutely that's really important when you're building a platform you don't want to reinvent the wheel that would be terrible because they got millions and millions of dollars and they've done a great job it's like Apple wanting to do all the apps itself or make all the movies that it sells on iTunes a good point I'm going to be reaching out to you anyway after this because I'm really interested to see how we can apply this and be in the frontier in terms of people development because I see a huge application I'm going to say a bit naive I'm not a tech person but I've been an early adopter of technology from e-learning days even before from computer based and now I'm doing mobile or whatever so where have you seen it successfully used within corporations like banking and whatever and I know that's a huge question but any sort of case that has come to mind where you see AI being used effectively in terms of and I sort of in my head I've got this vision of sort of machine learning collaborating with people learning so as you're picking up the patterns they talk to each other and helping each other continuously creating a culture of learning and adapting every day with technology being the amplifying so have you seen that happening or just to sort of see that as a big opportunity I do see that as an opportunity I haven't really seen it happen what I've seen is people doing something called people analytics where they are mining a lot of the data to predict and to understand and to learn the circumstances under which sales teams work very well together depending on which floor they're on and their movements in the office itself all of these factors they're really coming down the granular level to say how should we design an office and I think there's an office where they keep moving the coffee machine because they're doing some A-B testing to understand and they figured out when you put it in a certain place and that team kind of gets together and they analyze their emails and they found that when they get together for coffee they go back and those teams that got together literally the watering hole is called were better at sharing ideas and learning so they also compared two branches of a bank and they saw one in which the team was split on two levels and one in which they were all sitting on one level I think a loan to book ratio was greater so the one on the same level was able to do better as a team so the other thing is when you're sharing individually you may not be doing as well but the team does better and that is something interesting for an organization these things that lend insights because of the data and how that goes back into their learning it's not automated these are the kind of things where the chief learning officer or others would kind of glean these insights and then begin to experiment with how to improve the culture of learning and innovation but I think there's an opportunity for that for sure Thank you Do you provide these services AI services for the plants? Do you use any trade mark paintings or any technologies that are very specific to your company or do you use open source? I think that's a very good question we don't use any technology that is our own we don't have any to be honest we are just developing our first product now which is the NICCA insurance product that I talked about but the proof is that we use open source technologies and some of them may be paid as well because it depends on the client's preferences but we have not developed our own so in that sense we are not a product company we are a pure financial services company services company and we incubate products either our own or others and then we spin them out because I firmly believe you can only do one thing at a time so that's a good question So please services are only for new product development or is this also a plan for big data management Absolutely a lot of it is actually for existing products that people want to improve so a big bank has lots of data telco companies have lots of data in factories are working with one potential client factories wants to understand how can I predict when to repair them and how do I know how can I kind of align that with the demand all of these kind of things are very important and this is all existing there is no innovation here but the innovation is on cost optimization or operational efficiency what happens is that usually as you may know that in most companies it is hard to sell a project unless you can really show a return on investment that return on investment usually comes from showing growth and expansion to new revenue, new markets so we find that our clients struggle internally to sell some of these projects sometimes but that inertia is typical of large organizations Thank you so much share more example regarding how we can apply the AI in transport machine the most obvious one is self-driving cars a lot of self-driving cars are obviously using AI the other thing is you can make integrated mobility services such as the one that we are doing and all of that requires stitching together data from sensors and contextual data that you are learning to learn about how the city behaves during the day or otherwise you can also begin to understand where you should have new construction or where you should place your infrastructure like your MRT stations or other things like that the transportation industry in general is a very large sector from the actual thing that you build to the integrated services that you have to the larger transportation infrastructure that you need for instance, Jurong Innovation District is coming up it's going to be a carlight district in Singapore where should they put their bus stations where should they put their autonomous vehicles how many different kinds of bike sharing companies should they use how many bikes should they have all of these kinds of things require you to understand predict data that exists and then to tell something that you can do so I think that is really important now the classic obviously machine learning that we have seen is speech recognition so when you're in a car or in something you're speaking it's also image recognition which is used for self-driving cars those are things that are quite well covered in the media but what we do is kind of more infrastructure behind the scenes seeing how does the whole thing work I'll give you an example my co-founder just won an award at the World Bank and he was looking at women's behavior when new kinds of transportation modes were introduced so we do a it was called mapping for policy he just received an award at the spring meetings from the World Bank President and the idea was in developing countries women go to work but they don't feel safe and so we need to understand when and why they don't feel safe and so there are many feeder routes that go in and so we need to begin to understand what is it, is it from the lighting to the congestion to the people to the actual route that's being taken what makes them vulnerable to gender violence and so he analyzed 79 million trips in 2010 to actually make a model of this you can see you can come in transportation from so many different angles from user behavior to the persona of the person who's building the infrastructure to the original equipment manufacturer which is the car service to the entirely disruptive player like an Uber or other or something like that okay actually it's also time for time can you take one more question sure they decided to let you offer thank you so much very real pleasure to be here thank you very much