 Welcome to the AI for Good panel at South by Southwest. We're very excited to have everyone here today to talk about the potential for AI to do good in the world. We have three amazing people on stage with me today. We have Don Nafis from Intel Labs. Laila Ibrahim from Coursera. And Pratua Bharti, who's an Intel student ambassador from the University of Florida. And I'm going to let them tell you a little bit about what they are doing and why they're excited about AI. So my name is Don Nafis. I work in the research labs at Intel. And I study how people learn about data, data literacy and that kind of thing. Hi, I'm Laila Ibrahim. I'm the Chief Operations Officer at Coursera. We're an online learning platform. And I actually also spent 18 years at Intel and some time in venture capital. So I'm really excited about AI because I think it has a way to... I believe it has potential to transform the way that we learn. Hi, I'm Pratua Bharti, an Intel ambassador for graduate students and for deep learning and artificial intelligence. And I'm also a PhD student in computer science and engineering for... I'm doing my PhD there and what my study is about... I'm studying more about data learning and modeling the differential system where we can learn more about the health data. So one of my projects was activity recognition using the accelerometer and the gyroscope sensors doing the activity recognition model. And the other work also I'm doing that collecting different kinds of images from mosquitoes to identify that if the mosquito species is harmful or non-harmful, like if that particular mosquito is Gika mosquito or the malaria mosquito or they are the non-harmful mosquitoes. So this is why I... when I see AI or deep learning, so I understand that I can transfer the knowledge from an expert in context and I can bring that knowledge of... I can bring that knowledge to the other people, ordinary people, to just use that model and understand that what's going on where they don't have any other knowledge also. So it's a good segue into... I was thinking when you were talking about potential, but so Don, when you think about AI and what you've done, what excites you about the AI and all of the conversations around how it can be used these days? You know, what I'm most excited about is actually the civic uses and particularly when AI can actually get into the hands of people who see the world a little bit differently. So an example of that is right now I'm working with a group of a citizen group who is concerned about air quality. They have a point source polluter that they're worried about and actually they have very good air quality instrumentation going on and they've literally spent the last year just getting their heads around like what is this data actually telling us, right? How can we represent that data? What can we actually do with it? They also did a pilot where they combined wearables and other kinds of health tracking to figure out, you know, geez, do our blood oxygens or heart rates spike when there's pollution in the air, right? That kind of time series sort of correlation is simple to many of the people in this room, but it's really not simple at all if you're an ordinary person just worried about the air. And so now they're asking questions of, well, wait a second, can we actually, you know, sort of use some sort of technique to say, oh, you know what, this spike when it's these two chemicals and it goes over this threshold, that looks a whole lot like the point source polluter, right? But when it happens a little bit differently, that's actually the guy down the road who's like painting his car or whatever, right? And so they're asking that question of, you know, the data, but, you know, that's a perfect application of AI and then the question then becomes how can we get people like that to actually take advantage of the technology and know what's in it to be able to, you know, really understand what comes out. So Lila, when you're coming at it from a different perspective, you guys are really engaged in online learning and helping people learn. So when you think about AI and the trends you're seeing now, what do you think is really exciting? Where is this going that's an exciting potential? I think it's interesting the past few years, we've been talking a lot about big data, but data is only useful if you can find some way to analyze it and highlight the trends or be able to provide a service on the top of what the data is telling you to do. So from my perspective, what I'm most excited about from AI is the personalization that can happen because of it. So for example, in learning, which I think is a big social and economic issue around the globe, regardless of what zip code you're from, you can now be, you can imagine a time when you could actually use AI to be able to assess where you're at, what skills you will need to move forward, and have basically a specific learning path associated with your online learning. And I think we haven't had that ability to do it before, but I think AI is a tool for us to use in a way that, to advance the way that people are. Pratul, what got you interested in doing this, right? So you're getting your graduate degree, and so what excited you, what made you want to do this? Oh, yes. Actually, when we talk about AI, so AI, we are studying for a long time, right? But AI or deep learning become more productive, I think from the last six, seven years, when the hint and talk more about the back propagation and more deep learning came into the picture. So what I interested me now as a student, because initially when we were doing some traditional machine learning, so we had to put a lot of features, like the context-based features, because if you want to know the, suppose natural language processing, then you have to also include people who know how the pronunciations are there. So more context-based people you need to bring. So basically, suppose when you are trying to figure out some different image-based model, then you have to have some that context-based people. But now once the deep learning is there, you don't, or even if you need, you don't need that much context of the data because now the model itself figures out how the pattern, because you don't need to teach them how to find the feature, not to code them. Eventually, by the different, different layers, automatic now machine learning or the deep learning finds a pattern or finds the beautiful features through the auto-encoder and the different techniques. So that's what incites me a lot. Even I can do those work which I have no idea about, which I have literally no knowledge, but still I can work on those problems and I can build a model where I can, just based on the data, mind the data and automatically figure out what's going on and how to find those patterns again for the unseen data. So that's part of the thing as a student incites me more. So I listened, you know, when we talked, you mentioned the big data and making sense of big data and civic causes and solving problems and helping the data. So if you think ahead in the potential, what industry or social challenges do you think will benefit most from AI? Lila? Well, I hinted at it before, but if we take education, for example, so Coursera has, since it started about five years ago, we have 25 million learners with 75% of our learners outside the United States. What we've seen is that, in general, what's happened is education has always been like, let me go and take a class and then maybe based on what you've taken we can recommend something next for you of something we think you might like, kind of like what Netflix does with movies. Now, imagine if instead we took that a step further. So I could say something like, I want to learn digital marketing and I could take an assessment to know where I'm at and we would be able to map the skills of all of the 1,500 plus courses on our platform and I would be able to know exactly which class to go into, at which point. The quizzes and the assessments in that learning, my learning path would be able to guide me. Like it would be able, like maybe I'm motivated by competition, maybe I'm motivated by rewards, maybe there's this trend of me missing these problems. And my goal at the end of this is to get a job. So how can we help almost this personal coaching through the learning? And something I mentioned earlier, so much of education has been limited by making it affordable and accessible and having it high quality. A lot of times that's based on where you live and instead I think with AI what we can do now is make the world's best content available to everyone in a much more personalized way. I think that's really powerful. It really unleashes the potential of humans. That's a good one. Yeah, that's one of my personal passions. Don't think we could get stuck there. So Don, what would you pick? Well, you know, I've had long-standing interest in health and I think there's some very real promise in health. I work with a group of people called the Quantified Self Community, which are folks who are very interested in the data they collect about themselves and actually get kind of creative and kind of a little hacky about the kinds of data they collect. You'd be surprised when people measure about themselves. And one of the reasons they do it is precisely because for many years, for large part of the modern period, medicine has been based on sort of notions of population-based studies. We do randomized controlled trials often for very good reasons. And I had an MD say to me who participates in this community. He says, you know, a randomized controlled trial is really great for the first 500 people. You sort of learn what the side effects are going to be for a drug, for example. But the problem is when that drug goes to market, you could be number 501. You could have a very different reaction. And so sort of keeping track of those things in a much more ongoing, live way actually has real potential for sort of catching those moments. Now, I also think there's a kind of a balance here because there's a lot of promise, particularly when we're talking about you got AI to do diagnosis, to sort of parse medical images, particularly when, you know, technicians are getting tired. They've looked at the same set of blobs all day long. You know, at the same time, you know, we've got to figure out how to balance their workloads so that they do also, you know, they're able to have some control when they say, oh, wait, my view is actually different from the machine's view. And we need to actually trace back like why there's that discrepancy. And if we can shift, you know, through those work practices are kind of in line and in tune with what the machine does, then I think we can have something really great. Pratul, what about you? Could you repeat the question? So if you think about the potential of AI, you know, you're about to get your, you know, your PhD, what do you want to do with it? What do you think is the most exciting opportunity out there? What industry or what problem? Okay, so what I think the way AI can put most of it in fact, I think in the medical field and also I think in the business, like the predict the how much sales they are going to do after six months. So yeah, these two things I think already they are going very great. But I think in the medical, as I say, that already a lot of people are putting AI to find their cancer from the image and also the X-ray analysis, they are doing a lot better than the technician. So I think they have a lot of potential but still like the different cancer, all the cancers are not the same. Different cancers have the different things. But not all, and also like, all the expertise, all the doctors are available at every center, but the people need every time anywhere, right? So if we have these kind of model and the cloud and where just we take the image and get all sort of information immediately, then I think we can cure more things very quickly because a lot of time we listen to that because of the doctor couldn't reach at the time that's why we couldn't save the patient. So what I think that with the help of all the big data and the cloud system and the machine learning we can reach to the people more quickly and more easily everywhere compared to the now. So this is the medical side I think we can do a lot about using AI or the machine learning thing. And the second thing I believe in the business as I told you because a lot of times when you have suppose some event is coming a lot of the business is thinking that how much backup we need to have and how much sales we'll be doing so that we are not running out of the thing we can sell or also the prediction so okay so last time we did this much business the how much business. So I think a lot of scope we have in these two things and also I can say that there is no thing left now at least where you can say AI doesn't have an impact. So I would say wherever there is a data there is a pattern, obviously AI can make an impact on that. But as of now I think these two things where a lot of scope are there in the very recent future. Having been in the tech industry now for 16, 17 years every few years there's this panic about talent like we don't have enough software and recently it's been AI right so everywhere I go it's oh we don't have enough AI people to keep this moving forward so I'm curious Lila do you see more people interested in taking it on of course Sarah and I want everyone's opinion do we have enough people in AI and if not what skills are it how would we encourage more people and how do we really get a diverse population because if we're honest in technology often times we say we want diversity but as we build the skills they have always been in a diverse population so I'm I'd really like your thoughts on do we have enough and if not how do we get there and make it diverse. Big question it is I think there are two parts to this the first one is just how do we get more people familiar with AI right now on course Sarah one out of every six learners comes to our platform and is taking a data science class that's to me that's like incredible like that there's that much interest in data science the second number that I find interesting is we've over 2.5 million learners have taken machine learning classes and that's a little bit more advanced so we're definitely seeing the interest but this is still like in raw numbers it seems like a big number but it's really not that large I think part of what we need to do is how do we demystify artificial intelligence machine learning and all the aspects of it and this starts with building a strong foundation of skills which the university have been providing and now with we've been bringing on more intermediate content now with companies like Intel and Google will be bringing more advanced courses online and I think one of the so part of it is like just getting the word out there and making it more practical and on the diversity inclusion aspect of it it's really interesting so I'm an electrical engineer by training that was like 25 years ago and so engineering has changed a lot since I graduated and there's always this how do you kind of reinvent yourself and so I think that there are a lot of people out there with poor skills already that just have an opportunity to up skill we did you know when I went through engineering and as I've talked to women engineers just as one slice of the inclusion pie there's always this we have this hypothesis that if someone can look and see someone like them who has built a career or had success then they think oh well maybe I can do that too but we decided to test this at Coursera so we actually had a machine learning class from University of Washington and we marketed it and we in one message we highlighted that professor we talked about professor Fox and professor Fox's accomplishments and in the other one we talked about professor Emily Fox and her accomplishments and what we found in our testing was that there was a 26% increase in women enrollments in the machine learning course just with the way that we marketed that course and there was no change for men we're doing a lot of experiments and that's one of the things I'm actually really excited about with Coursera is now that we have this platform with 25 million learners we can tell you when people fall off in a course we can see trends that we couldn't see before like in a classroom of 30 people if 10% misses a question you're like they just weren't paying attention but now we can be like okay there's 100,000 people taking this and 10,000 people missed it so professor your question's not right the material leaning into it isn't right and putting some of these triggers in place we think can really change it so I think part of this is just making the content more available and doing it in a way that provides a safe environment where people of many different backgrounds can learn the same information can I add to that? yeah I'd just like to add to that one of the things that we know from a research point of view historically as new technologies come along is that the way that particularly we get more women involved is when you know there's a particular problem that's actually worth solving and there's a problem that somebody cares about already and then if you can start to solve those with math or data or new technologies or coding that's a different path into STEM education and a really important one for lots of us myself included actually I'm an anthropologist I've been trained to and I've come full round once I've seen what they can do both for health and for the environment right so finding those areas where it's the emphasis is not and now you start to learn to code but actually here's what you need to know in this domain right now for this purpose and designing curriculum accordingly I think is going to be really promising the other thing I kind of wanted to add here was that particularly when we're talking about AI you know the early critique has been that a lot of these systems inadvertently most of the time end up embedding some bias in them right and one really good way to fix that is to make sure that women people of color people who are affected by these systems are in the room when they're getting developed so I'd actually say that there's a particular urgency in this technology to sort of solve that diversity problem I think that's a really interesting point because I think it's our responsibility of all of us just to be able to say how can we highlight applied examples of AI for good regardless of like where they came from and who's behind them and like really but also try to find special cases and shine a spotlight on them so that we can I think that's all of our responsibilities so that's a call to action for everyone out here and if you know of any really good examples too we'd love to hear about them because I think that's having looked at this and I agree with you on women it tends to be more problem solving so it's interesting as I talked to people about AI they're both the people that we need deep expertise who can build the algorithms but then you were hitting on it too you need a number of people who know how to use this or know how to interface with the data differently and so I think you can get a whole range of people my hypothesis is people who are deep building the algorithms and people who are more used to using them do you see students excited about it Pertual are you the exception or are you the rule these days in students you can speak for all students there's one of you up here it'll be fine yeah so what you're asking is that AI help me or something when you're on a university do students understand the power of AI are they excited about it do they want to go into it is it confusing what would help okay so when I took a decision to study about AI and the machine learning so what was my thinking is first of all like there's huge scope in AI so it's good to study that field so that obviously you'll have not very difficult time to find a job so that was one of the reason and the second thing was that it involves a lot of mathematics where you can do a lot of that kind of field if you are an expert or if you are good enough in mathematics and the probability and finding a graphical model so that you have an edge to study these fields and interestingly I have that kind of edge so that kind of helped me to decide this thing and the third thing which I decided why because AI it's very intuitive and you can see the result very quickly because some of the study you do there's a lot of theoretical things you know it will when you do this happen like this but not always you can build a demo kind of where you can easily see the result so that's why the most exciting part of machine learning you just train the model and once you feed the data you can see the output immediately so that's also the one of the thing students are very excited about so what I'm doing people can see and appreciate the thing so that's why you use a lot of credit about it right okay so I can show the people this is what I am studying so they can understand that they can appreciate about because even when I go to my place and I saw my mother so they don't understand about when I saw them so even when you really watch and walk so my computer can understand that you are walking so then see I said oh how you can do that without having a camera so then I said okay so there is a sensor in this watch so it can detect you without seeing you so then see I appreciate oh even with that can be done so these kind of like different different things excite a lot I think me and other students where you can really see the results because a lot of people are passionate about the results so can I see the output of this okay everything is good but what about the output so that's why like it excited me a lot and these are the things I think I decided to pursue my career in this I haven't thought about that but you're right you get some immediate turn on the data right once you you can see how how it changes and you can see how you can retrain the algorithms fairly quickly too so so before we open it up to questions what do you think is going to be one really exciting thing that's happened in AI what's something that you really think in the next five years and that's a long time in technology I mean that's that's eons right but what's what do you think we can sit here and say this is going to be fundamentally different because of AI Don you're nodding gosh I'm nodding and nodding actually so I think I think the honest answer is I can't I just can't say at this moment but we have ways of thinking about it right so if we think about where we were in the early days of the internet right so we had all this technical research that sort of showed yes you could do connectivity between computers right and so and when it just started becoming a consumer thing you know where we in a position to even say you know five years ago probably not but actually I think the experience of people learning how to use the internet and finding really creative things to do with it than we never thought before I think that was kind of a 20 year project in a lot of ways and so we might I think the major changes are actually going to be sort of a lot further down the line potentially it's I think the but as I listen to all of you and I want to I'll do you guys a chance too but there's this piece of personalization and making sense of big data that I think will go faster my hypothesis is that we think it will because once people can make sense of the data they'll think about their data in a different way and what data they can use to solve the problems and right now you know I remember the first time I used the internet that dates to but I was like this will go nowhere because you can't find what you want right I remember this thought it was like this is useless right and it wasn't until you got the search engines that made it easier and I think AI has the potential to do the same thing with data that you hit on instead of it just being amazing amounts of data it can help you make sense of it so couldn't agree more so you're right I probably can't envision it but all of the data that's out there but I think it might be the small problems right instead of thinking about the big problems to me the power would be that a lot of people are solving the little things that they care about and it will see more of that nascent and native growth in problem solving and people feeling more empowered to solve their own problems and if I could just add to that a little bit I mean I the area where I expect a lot of innovation actually is in sort of as designers are designing their systems and AI is a component right it is actually not a solved problem how you communicate right what the thing is doing right and how you give people like some sense of what it is and maybe some other people like fuller a fuller understanding a deeply technical way right how to do that in these really really complicated systems like we're I think we are actually just at the beginning of that why what do you think any predictions no one's listening two things two things parallels come to mind the first is you dated yourself so all day mine myself as well 20 years ago I went to Japan to work on DVD standards and everyone thought I was crazy because they're like who's going to want to watch movies on a computer like people could not even imagine it and DVDs are like gone now right so and then so part of me is like I you can't even it's so hard to predict the future like so I think what's exciting is what we can't even imagine one of the my big lessons out of spending four years in venture capital was when you can democratize technology as a tool and you can make it widely available to everyone right and unleash the power and the potential of entrepreneurs to ask questions to come up with creative solutions problems you didn't even know existed and so I think that's to me it's almost like an open like I I just want to remain open and see what can really come out of making AI more available to the world for tool any ideas yeah five years is a long time actually so I really don't know five years but I can tell you like maybe two three years what it's going to be so three years back I was in China and obviously I had a lot of trouble understanding their thing so what I was doing like I was trying to locate the English word in their conversation and I was trying to make sense what they are trying to say so what was happening there so I thought that why not we have some automatic translator and nowadays we have that so you speak and they translate you in whatever language you want but still you think because many people use two three different language right mix English mix other language so that's where you find the machine different part because they expect always the English or same language input and then they translate the same output so what I'm thinking that that should be a global model you feed any kind of language and that will be output of whatever language you want in whatever comfortable thing it would be a huge coming removing the communication barrier the whatever part of world you go and you just keep on and put your headphone and you getting the things in whatever language you want so that would be a huge thing like whatever part of world you go and you don't need to worry about whether I will understand the things or not so I think that would be the I think it's already there it will be maybe two three years down the line we can find that but if you say so what's problem I'm facing right now if just for a student perspective because we are starting to collect the data and we are not from industry and the deep model requires lots of lots of data to train even a shallow model so why can't we find a technique which requires very less data but still build that strong powerful model which works as good as any like the Google net or the Alex net we have the popular deep learning model so I think in the three four five years down the line somehow we can find the method so even you have the very sparse data you have very limited data but still you can build a powerful model which can help the people and the society so I think that would be a breakthrough in deep learning I guess I want to I could go on but I want to see if there are any questions or ideas or thoughts from the audience about potential what's getting at what we need to do yes thank you by the way that's from there ok so one I'm just really glad that you guys are here and I'm glad that I'm here because I feel like this is what South by is supposed to be about I was lucky enough to get sneak in here and get a pass at everything and I'm actually one of your learners I'm on for Sarah and I'm actually really interested in fact that you said help because I'm actually thinking about going back to school and one of the things I was thinking about was data science and also kind of coupled in with public health so it's funny that you said help and I was wondering if you've seen anything or know anything as far as like how to how AI has been integrated as far as like how maybe public health or anything like broader maybe in that same you know topic basically because I'd love to know that I don't have a math right and I don't know what to say I'm not sure about that now but if I could figure out maybe kind of do that and go back to all that so anyway well if I could make a shameless plug for my book I totally will I just put out a book with Gina Neff at Oxford called self tracking and it doesn't talk about AI specifically but it does talk about what happens at that interface between sort of consumer health clinical health right how data kind of moves around with some of the sort of the institutional social questions that kind of come up which might it's certainly not going to ground you in data science anytime soon but at least give you a flavor of the landscape for how that stuff works right now it's interesting I actually think that like my generation when we went to college we had to we were force fit into like a specific way of learning like I had to do electrical engineering I couldn't go do health too again just it didn't mix together and nowadays I went as I go and speak at universities I see how all of these different fields are coming together and it's going to continue evolving and actually part of what we're trying to do at Coursera it's not disrupt the higher education it's just disrupt the labor market if you think about it healthcare like AI and healthcare this is going to look so different in five years ten years so you can go and be working in the field and how do you stay up to speed and change things that's why we're really excited like partnering with both the universities and leading companies how can we stay abreast of what's changing because I think your challenge now as wanting to go into that field is it's going to continue evolving so developing that your own self in terms of a lifelong learner is a really critical and I thank you it's an interesting work you'll be doing and I know we talk about the deep algorithm you know needing a math brain I don't think that's entirely true I've watched a lot of people who don't think they have a math brain but when they want to solve a problem they learn to do that and so it really is if you're interested in healthcare and with all of the resources online you can try a lot of things I agree universities I'm watching people get certificates they learn differently they apply them industries starting to hire differently so if that's what you're passionate about you can narrow in on that and probably get some experience and start to do it and make sure you like it before you get a full new degree so people are doing less of that but so I think it's an exciting one and there's a lot out there other questions okay there's someone over there hi so let me ask you all when I hear about AI it's typically in the hard sciences and I'm wondering what your thoughts are about AI and the soft sciences my specific field is filmmaking I'm a filmmaker in the arts and I'm wondering about thoughts about how AI you know I'm passionate about the arts and also education I want my students to get jobs you know have careers and I'm wondering if you have experience or thoughts about what role AI can play in arts filmmaking and the others in general so I'm actually going to use two examples and they're both real-time and I'm going to ask my husband's forgiveness he's not an artist but he's an archaeologist and people don't think about data and he's actually a computational archaeologist so they've just finished a project where they didn't actually dig in the ground which takes all the fun away for kids by the way, right, no one but they did use GIS and then they do a lot of data manipulation but they had to teach the data how to learn from itself because there were so many tiny little fish bones and so even in something where people don't think about it you can start to see the data and how you can get deep learning and machine learning anytime there's massive data and it started to tell them more about where they could dig so they didn't disrupt the ground and specifically to arts we just did a project from Intel using the algorithms one of our deep technologists partnered with a ballet company in Portland and they put wristbands on the dancers and they programmed the algorithms so that they would generate their own music as they danced and so we now have I think it's U.T. Austin actually I'm pretty sure we just donated about 10 wristbands so they could start to use it in their music classes and they also gave it to musicians and it was for kids to get them interested and so they put the wristbands on four musicians and it amplified their music as they played so with a violin player picked up their actual playing it made it sound like they had an entire orchestra behind them and so then they got to have a sort of playoff with the full orchestra because a lot of kids got interested so even there you're starting to see multimedia and musicians figure out how you can use that data behind the scene to make their realities come true in a faster way and so that's why I do think you're starting to see more people think of themselves as a technologist than we have before because it allows them to get their dreams they've envisioned just the way they want it to look or sound technology starts to do it and when you get that much data you need something that helps you do it you can't be analyzing it if you want it real time so you wouldn't be able to do this real time without the algorithms because it would just take you too much human so there is actually a lot out there and I think you'll see more and more of it so if you want to talk more I'm happy to show those examples they are cool so this is my passion I think you will see people who can use the technology as opposed to seeing technology as something separate this will become more and more a part of how people are thinking about healthcare or music and that will lead us into places I don't think we do know what it would be like in five years because it's moving so fast and so that would be my short answer other questions other hi oh okay hi I'm in the healthcare industry I'm a little bit skeptic I guess I've been in it for a really long time I've been in it for a really long time and you talk about demystifying it and applications to patients and their healthcare but really when you think about it the people who hold the keys to the industry are the FDA and big pharma and you know a bunch of healthcare systems and hospital systems and how do you start to demystify it for them and make it applicable in a way that makes them a little adopted don't you oh dear oh can I plead the fifth I don't know I do know that it is you're identifying the exact same challenge that we saw in the research that you know new data types do not interoperate well with systems that are built on old data types where people know how to read those and they travel very quickly and they're already embedded in the system and that's kind of that clearly where there's efficiencies to be gained no doubt folks in those institutions can start to take a broader look but where those actually come from is frankly a good question it's interesting with my VC hat on there are a lot of startups right now that are trying to figure out how can they disrupt the system and really I think that's it's going to be interesting over the next few years but you see a couple of folks kind of popping up you know even you can take something even more established like Chan Zuckerberg initiative and the work that they're trying to do around health care that Cory Bargman is leading of having both the technology and the life sciences or Google doing some interesting work with Calico which is one of Coursera's founders who is an AI expert so I think we're kind of early in the stage but there's enough happening that I think we're going to make there's some things going to happen if I could just maybe I mean the thing is health is not always the thing that happens in a medical center right and so if we think about what's happening on the consumer side there is so much that we can learn about sort of how the real world happens on a daily basis outside of the lab outside of clinical settings so if you think about that as a matter for public health research that might affect health care delivery in an indirect way but it's not necessarily about delivering that health care there actually might be some promise on that side other questions thoughts good and I'm so sorry what was the name of the book you said she's very happy I'm so happy that's like the best question ever thank you she'll send you a book I will send you a self tracking by MIT Press mm-hmm just out right dog just out so I want to thank everyone today for your time here listening and asking good questions I would encourage all of you to find out more about AI deep learning, data science what you can do with it talk to people about it, it isn't scary but it is confusing at times but you can get a lot of information for Sarah shameless plug for for Sarah there's an amazing amount of online learning that you can do to explore without even paying any money so you don't have to get invested you can start to talk to younger people about it as well you know I have a fourth grader now wow she's past second grade and you know she has the same I don't like math but when she started to figure out what you could do with the data in math she liked it so talk to people, learn about it start thinking about how you can use it and encourage the system to start to solve problems and put pressure into it if you want to hear more about it from an intel perspective shameless plug Diane Bryant who is our senior VP in the data center has a keynote tomorrow at 9.30 in the morning she's a brilliant speaker I don't think there are many people here on stage who sound as amazing as Diane so please come and if you'd like to know more about any talk to any of us let us know how we can help