 So I've been working on a mobile competing for about 15 years now and I believe that in the near future our smart devices will know almost everything about us. They'll make it so that we can continuously capture and analyze human behavior at a scale of fidelity that we've never been able to do before in human history. And these devices and the data they offer will allow us to revolutionize healthcare. Now let me take a step back and explain why I believe this is the case. Last year over 1 billion smartphones were sold and they are perhaps the most intimate devices we've ever created. Let me give you some stats about how young people use their smartphones. Over 83% say they sleep with their phones, over 90% say they check them first thing in the morning and over a third admit to using them in the bathroom. But it's not just the devices that are intimate, the data is also intimate. So today's smart devices already know who we know based on the call log and contact list as well as where we go based off the GPS. Tomorrow's smart devices will also know what our activities are, how well we're sleeping, as well as how close we feel to other people. Let me tell you more about sleep. How well you sleep is actually a really great predictor for your physical and mental health. My team at Carnegie Mellon has developed a sleep monitor using the smartphone sensors as well as machine learning techniques. Here you can see that the smartphone data drops off around midnight when the person goes to sleep and there's a spike of data around 8 a.m. when their alarm clock went off and they probably hit the snooze button a few times as well too. I also mentioned closeness is really important too. So having good friends is the equivalent of quitting smoking. So people who have really good social support are less likely to get sick, live longer, and have higher survival rates for diseases. Using communication data alone, my team has been able to infer how close you feel to other people. So on the next slide, you'll see a picture of my own call log data. On the right side, you'll see that my brother and I actually chat a lot. We chat all days of the week and all hours of the day. Whereas on the left side, you'll see that my best friend and I actually don't talk that often, but when we do talk, we have very long calls. And so we use features like these to build our machine learning algorithms to create a really accurate classifier. But our devices can do much more than just collect data. They can also provide meaningful interventions too to help us achieve really hard goals. So one of the big problems we're facing in developed countries for healthcare is that chronic diseases like obesity and diabetes are really challenging. And so what can our smart devices do to help with these kinds of challenges? If we want to make this happen, there's several kinds of challenges that we need to overcome. So one kind of big issue we have right now is that people aren't aware that or aren't ready to make these kinds of changes. This shows a trans-theoretical model behavior change. People on the left side are unaware of the impact of their behavior on their physical health. People in the middle procrastinate. Our smart devices can help nudge us from the left side over to the right. Another problem is that people often don't know what to do. Sometimes you'll get vague instructions like eat a low sodium diet or stop smoking. What if our devices could help personalize these kind of interventions to make it so that they're much more useful for us and are much more meaningful too? But there's also even more challenges with our physical and social environments. If it's really hard to get fresh food, if it's really hard to walk in environments like this. Or if we're just surrounded by unhealthy snacks, it's really difficult for us to be healthy. So our smart devices can also help with this too by helping us make more meaningful kinds of interventions and make it easier for us to make these behavioral changes. So our health is the sum of small decisions we make every single day. It's the choice of taking the stairs instead of the escalator, choosing to eat the apple instead of taking the cookie, or choosing to walk instead of taking the car. And so what if our devices could provide us meaningful interventions and want to nudge us so that we can be healthier too? And so we might use techniques from psychology like loss aversion or peer pressure and see how well they work for us. We might also use gamification too and our devices can learn over time which kinds of interventions are more likely to work and then start biasing those interventions more towards those directions. Our devices can also start monitoring our behaviors and see what kinds of activities we have to help us establish new kinds of healthy habits linked to those existing activities. So for example, if our smart devices see that we tend to watch TV in the evening, it might suggest doing some light exercise while we do this. And if it tests that we've also established this new healthy habit, it might slowly start notching things up, getting us to do more physical activities that we also find to be fun. Our smart devices can also do A-B testing of alternatives to see what kinds of things work best for us in what kinds of situations and can also aggregate this data across thousands or millions of people to see what kinds of things work for what kinds of people in general. Now, I sketch out a possible future for healthcare here and we're in the early stages of building this at Carnegie Mellon University. But there are many grand challenges before we can achieve this vision, perhaps the most noble of which is privacy. So how do we make it so that people feel like they're in control of their data? How do we make it so that people feel like they're getting tangible value and that there's assurances that their data will be used properly? So my closing question for all of you here is, how do we make it so that we have our smart devices that can offer us meaningful kinds of interventions for healthcare but it's also respectful of our privacy? Thank you.