 Live from Austin, Texas, it's the Cube. Covering South by Southwest 2017, brought to you by Intel. Now, here's John Furrier. Okay, welcome back everyone. We are here live at South by Southwest of the Intel AI Lounge at Silicon Angles, the Cube, talking to some great guests. The theme for this week is AI for Social Good. I'm John Furrier with Silicon Angles. Our next guest is Frederico Gomez Suarez, Technical Advisor and Volunteer at Thorn, doing some really amazing things with technology for the betterment of society, specifically a use case. So, Frederico, welcome to the Cube, welcome to the AI Lounge here at Intel. Thank you very much for having me. So, talk about Thorn, and first of all, you work for Microsoft, but you're a volunteer, I'm a volunteer, we're going to talk about what Thorn is and what you guys do. It's really a great story. So, Thorn is a non-profit which focuses on driving technological innovation to fight child sexual exploitation. And it does it two ways. One of them is by doing research to find the new trends and the new ways that this is happening. But also by using the latest technology to find ways that we can actually find this problem. Thorn has something called the Innovation Lab, but we're always trying new technology, we're trying AI just to find new ways to find the problem. So, this is really a great use case of where technology is being used for the betterment of society and good because what you're doing is taking really cutting edge, big data, machine learning, AI techniques, and the rage right now is facial recognition. So, talk about where and how it works and what's the results and can you share some of the impact? Yeah, so, as part of my volunteer work, one of the projects I have been working is called the Child Finder Service. And the idea of this work is if we have an image, particularly an image of a child who have been missing, can we use facial recognition to determine whether another image is the same child? And this is actually a pretty challenging problem because the child may have gone missing many years back and now we want to match against another picture where the child may show much growth, much growth depending on the duration, right? And if you imagine the impact of actually having this technology, a person who is trying to look for a missing child, if they have to go through a lot of pictures, it's actually hard to determine whether two people are the same person or not. So we're helping in that case. We're helping so that you don't have to go through so many pictures so that we can highlight the ones that the machine thinks is actually the same person. Take us through how it works in just a use case, just as an illustration. Yeah, so when a child goes missing, the National Center for Missing Children, which we work with, they publish a poster and that poster has an image of a missing child. Now, once you have that image, you may want to say, well, are there places where the picture of that child may be showing up? One place that there's usually pictures of children being exploited are online ads. So let's say that there's online ads that you want to say, well, in any of these ads that they use for exploitation, could there be the same child in both of them? So that's actually a use case and just using face recognition technology, we can try to make that problem easier or faster than it would be if you were trying to do it manually. And you're doing a demo here in the Intel AI lounge. What's in the demo? What are you showing? So in the demo, I'm showing how difficult it really is to do face recognition by hand and how by just having some assistance from a machine, you can go from having to look at hundreds of images and spending potentially hours to doing it in seconds. So how do you get involved? This is a volunteer organization. Take us through your journey. How did you get involved and talk about how you guys are getting more people involved and how can someone get involved? Absolutely. So as for Microsoft, there is the hack for good community and they encourage us to go and donate our time, our skill to nonprofits. And two years ago I had this idea and I did a hackathon and after the hackathon I got connected with Thorne, I learned about what they do and that's how I pretty much got involved. I was really fortunate that Microsoft supported me to actually go spend time with a nonprofit. And when I started working with Thorne, I realized, hey, there's other tech companies also willing to help. So in this Child Finder Service project, I work with Intel, I work with other companies all coming together to find ways to solve this problem using the current technology available. And you know, if Thorne is always looking for volunteers, we're looking for what we call our tech defenders. If you go to our website, which is wearethorne.org, WAC, SXSW, you'll find a link where you can actually volunteer your skills as a technical defender for Thorne. So talk about, that's very cool by the way, for people should check out Thorne's website, Thorne. Yeah, yeah, yeah, it's wearethorne.org, WAC, SXSW. Okay, wearethorne.org, slash SXSW for South by Southwest. Yes. So talk about the technology, because obviously Intel makes chips, makes stuff go faster, you got more compute, you got more cores, you got now cloud technology, and you're seeing at Google Next where they were showcasing their Xeon processor, the AI trend now is becoming really, really, really big. I know Microsoft Azure, Amazon Web Services, they're all having these machine learning libraries, and the big trend is self-learning machines or deep learning, so this is a tech trend, but now when you apply it to this, it really can work. So what is some of the technology and what are some of the data sets that you use? How does it work under the covers? Yeah, so we actually started with an open source technology for face recognition, and this was, after we started with this technology, we realized that we had to make it better. So we had to build data sets ourselves. For the data sets, we have images of the posters that are published for the National Center. We have also started asking people to donate images over time of themselves, because we need images of people when they're children or when they're older, and that's how we'll be building data sets, and then having the data set, we need to go and train them, and that's where using hardware, particularly using GPUs to actually do training really is key for us. You know, the technology really under this is deep learning for us. We use, using existing deep learning models and improving them with our particular scenario, because there's special challenges in our case, not only with the age, but also a lot of the images that we process. Sometimes there's heavy makeup, sometimes there's things like that. Or resolution, right, depending on the photo. And you know, low resolution images particularly, they're a challenge, so we need to improve it, we need to keep training to actually get to the point where we feel we have a really robust system. I want to ask you a personal question, and this is something we were talking about on our intro segment, and something that I've been thinking a lot about. I haven't written about it yet, but I've been starting to tease it out on some of my thought leader interviews, is that in every major inflection point in the business of technology, there's always been a counter-culture movement, and it seems to be that, if you look at all the news, whether it's political, or tech company news, and all the stuff happening around the world, it seems to be a social, good culture developing. You're seeing a counter-culture where what was once valued, tech, or public proprietary algorithms is now changing to open source, community, societal benefits. There seems to be a lot of activity, and no one's kind of put their finger on it, and you're a great use case of that example. And I see the Hack4Good community in Microsoft is growing, and there's people, peers of mine, working on all these kind of interesting projects, helping on profits. And that's called Hack4Good? Yes. What's it called? Hack4Good in Microsoft. So that's a Microsoft hackathon with employees, who just say, hey, let's pick something good to do, and they apply their programming technical skills to... Yeah, and there's a lot of support, and we're encouraged to do it, to me it's inspiring to work in a company that really encouraged that. And you know what? I see the same when I look across the industry. I see people willing to spend their evenings, like I spend my evenings working on some of this, or weekends, but we're passionate about making a difference. And I know I'm not alone. I met a lot of people, and I know there's a lot more out there. Is there a community people can check out? Is there on the website? Is there open source communities? Is there certain software groups that are playing more than others? Actually, I don't know. I know in my space, I think a great place to start is joining Thorne's Digital Defenders. But I would say if someone is passionate about a cause, it could be anything. And say I want to help. There's non-profits out there for that. And when I work with non-profits, they're so passionate about it. And sometimes they just need help in little things and having so many tech community going and helping makes a huge difference. So I would invite people to just go, if you're passionate about it, just go for it. Find a non-profit, they'll be happy to work with you. Fred Rico, I want to ask you if you could share just some anecdotal impact that you guys have had. Can you share some successes, some advances, and just to highlight some of the things? Yeah, so Thorne just published their digital report and it was really encouraging. So Thorne has a couple of different tools that they build. One of them is called Spotlight. Through the use of this tool last year, about 2,000 children who were victims of trafficking were recovered from around 6,000 victims. And each victim is a person. And the fact that we're making the difference in those lives is extremely encouraging. And that's just one of the things that we were able to contribute. So that's one of the stories that we have. And to me, it's not only that, to me it's also the fact that I see people who are willing to actually get engaged, learn more about this problem, it's another huge win. Final question for you, Fred Rico. Describe the scene here at the AI Lounge at Intel. What's for folks watching who aren't at South by Southwest, what is the vibe here, what are they showing? Obviously AI is the theme. AI for social goods are broadcast here. Hashtag is Intel AI, if you're interested in sharing, we'd be appreciative if you could retweet and share the love. What's your thoughts on the vibe here? Describe the scene here. You know, when I look around, all the demos are amazing. Like each one of them, you're blown away by it. And it just shows you how, in a practical way, AI can be changing lives or doing amazing things. There's the drones, they're on the video, the drones. And I love those, they look amazing. And then there's also the demo around, you know, using an art style and getting a picture. And I'm going to get mine in a second. I think if you combine, you know, you'll see how AI really in practice is able to just contribute to people's lives. So, and the vibe is awesome. And I'm loving it here. Well, I want to say congratulations. You do amazing things. Thank you. It's a real testament to where the society's going. AI for social change. Microsoft has a hackathon for good. And this is not a one-off. I mean, Microsoft certainly has had that. Google's got the 20% work on your own project. Intel has it. Companies are getting involved. They counterculture is developing for societal benefits and all of these new things happening like autonomous vehicles, smart cities. These are paradigm shifting society changes around the world and will require human involvement. Congratulations and thanks for sharing. Thank you very much. And we have a hashtag just for our project which is hashtag defend happiness. Defend happiness? Yeah, which is all about stopping, you know, sexual exploitation and trafficking all around the world. Okay, hashtag defend happiness. Please put it out there and share it. Tweet this video and for the betterment of society. I'm John Furrier with Rodrigo here at the Intel AI lounge. More coverage from South by Southwest. Three days of coverage, full day cube today. Some interviews tomorrow. Intel has some amazing super demos. They're going to be showing here throughout the weekend. Stay tuned on the cube we'll covering it. We'll be right back with more after the short break.