 Live from Washington D.C., it's theCUBE, covering AWS Public Sector Summit 2017. Brought to you by Amazon Web Services and its partner ecosystem. Got it. Melvin Greer is with us now. He's the director of data science and analytics at Intel. Melvin, thank you for being with us here on theCUBE. Good to see you here at this point. Thank you, John and John. I appreciate you getting a chance to talk with you. It's great to be here at the AWS Public Sector Summit. We make it easy for you. Great, John and John. I never forget the names. I saw just that data science in general and analytics. I mean, tell us about, give us the broad definition of that, the elevator speech about what's being done and then we'll drill it down a little bit deeper about Intel and what you're doing with the term in terms of government work and health care work. Sure, well data science and analytics covers a number of key areas and it's really important to consider the granularity of each of these key areas primarily because there's so much confusion about what people think of as artificial intelligence. It's certainly got a number of facets associated with it. So we have core analytics, like descriptive, diagnostic, predictive and prescriptive. This describes what happened, what's going to happen next, why is it happening and what should I do about it? So those are core analytics. Yeah, analytics business. I'm so glad. So here in a different tech, we have machine learning and cognitive computing. These things are different than core analytics and that they are recognizing patterns and rely on the concepts of training algorithms and then inference, the use of these trained algorithms to infer new knowledge. And then we have things like deep learning and convolutional neural networks which will use convolutional layers to drive better and better granularity and understanding of data. They often typically don't rely on training and have a large focus area around deep learning and deep cognitive skills. And then all of those actually line up in this discussion around narrow artificial intelligence. And you've seen a lot of that already. Evan Chajan, you've seen where we teach a machine how to play poker or we teach a machine how to play Jeopardy or Go. These are narrow AI applications. When we think about general AI, however, this is much different. This is when we're actually outsourcing human cognition to a thinking machine at internet speed. This is amazing. I love this conversation because a couple of things in that thread you just brought up. The poker, which is great. It's not just Jeopardy. It's poker's unknown conditions. You don't know the personality of the other guy. You don't know their cards or dealings, that's a lot. That's like unstructured data and you have to think about that. But it really highlights the converge between supercomputing, paradigm and data. So that really kind of changes the game on data science because the old data warehouse model, storing information, pulling it back, has latency. And so we're seeing machine learning in these new apps really disrupting old data analytics models. So I want to get your thoughts on this because what is Intel doing? Because you guys have restructured things a bit differently, the AI messages out there. As this new revolution takes place with data. So how are you guys handling that? Intel formed in late 2016 its artificial intelligence product group. And the formation of this group is extremely consistent with our pivot to becoming a data company. So we're certainly not going to be abandoning any of that great performance and strong capabilities that we have in Silicon architectures. But as a data company, it means that now we're going to be using all of these assets in artificial intelligence, machine learning, cognitive computing and Intel, in fact, by using this is really a unique, in a unique position to focus on what we've termed and what you hear our CEO talk about as the virtuous cycle of growth. This cycle of growth includes cloud computing, data center and IoT. And our ability to harness the power of artificial intelligence and data science and analytics means that Intel is really capable of driving this discussion around cloud computing and powering the cloud and also driving the work that's required to make a smart and a connected world a reality. Our artificial intelligence products group expands our portfolio and it means that we're bringing all of these capabilities that I talked to you that make up data science and analytics, cognitive, machine learning, artificial intelligence, deep learning, convolutional neural networks to bear to solve some of the nation's most significant important problems. And it means that Intel, with its partners, are really focused on the utilization of our core capabilities to drive government missions. Well, give us an example then in terms of federal government and AI, how you're applying that to the operation of what's going on in this giant bureaucracy of a town that we have. So one of the things that I'm most excited about is that there's really no agency, almost every federal agency in the US is doing an investigation of artificial intelligence. It started off with this discussion around business intelligence and as you said, data warehousing and other things. But clearly, the government has come to realize that turning data into a strategic asset is important, very, very important. And so there are a number of key domain spaces in the federal government where Intel has made a significant impact. One is in health and life sciences. So when you think about health and life sciences and biometrics, genomics, using advanced analytics for phenotype and genotype analysis, this is where Intel's strengths are in performance and the ability to deliver. We created a collaborative cancer cloud that allows researchers to use Intel hardware and software to accelerate the learnings from all of these health and life sciences advances that they want, sharing data without compromising that data. We're focused significantly on cyber intelligence, but we're applying threat and vulnerability analytics to understanding how to identify real cyber problems and big cyber vulnerabilities. We are now able to use Intel products to encrypt from the BIOS all the way up through the application stack. And what it means is, is that our government clients who typically are hypersensitive around security get a chance to have data follow their respective process and meet their mission in a safe and secure way. I'm not going to drill down on that for a second because this is kind of a really, really sweet area for innovation. Data is now the new development environment. We say bacon is, oil is the new bacon. It's the gold nuggets that I was talking with. You know what I'm saying? It's the new bacon. A new bacon. Data is the new bacon. Everyone loves bacon. Everyone loves data. There's a thirst for the data. And there's also compliance issues. And I asked you, the role of the CDO, the chief data officer is now emerging in companies. And so we're seeing that also at the federal level. I want to get your thoughts on that. But to quote the professor from Carnegie Mellon who I interviewed last week said, the problem with a lot of data problems is that it's like looking for your needle in the haystack. But there's so much data now you have a haystack of needles. So his premise is, you can't find everything. You got to use machine learning and AI to help with that. So this is also going to be an issue for this chief data officer, a new role. So is there a chief data officer role? Is there a need for that? Is there a CTO? Who handles the data? It's a tough one because there's a lot of tech involved but also there's policies. Yeah, so the federal government has actually mandated that each agency assign a federal chief data officer at the agency level. And this person is working very closely with the chief information officer and the agency leaders to ensure that they have the ability to take advantage of this large set of data that they collect. Intel's been working with most of the folks in the federal data cabinet who are the CDOs who are working to solve this problem around data and analysis of data. We're excited about the fact that we have chief data officers as an entry point to help discuss this hyper convergence that you described in technology. Where we have large data sets. We have faster hardware. Of course Intel's helping to provide much of that. And then better mathematics and algorithms. When we converge these three things together, it's the soup that makes it possible for us to continue to drive artificial intelligence. But that notwithstanding, federal data officers have a really hard job and we've been engaging them at many levels. We just had our artificial intelligence day in government where we had folks from many federal agencies that are on that cabinet and they shared with us directly how important it is to get Intel's input on both hardware, hardware performance, but also on software. When we think about artificial intelligence and the chief data officer or the data scientist, this is likely a different individual than the person who's buying our silicon architectures. This is a person who's focused primarily on an agency mission and is looking for Intel to provide hardware and software capabilities that drive that mission. I got to ask you from an Intel perspective, you guys are doing a lot of innovating since you have a great R&D group, but also the silicon you mentioned is important. And software's eating the world, but data's eating software. So what's next? What's eating data? We believe it's memory and silicon. So one of the trends in big data is real-time analytics is moving closer and closer to memory and then now silicon. Look at some of the security paradigms of the data involved is seeing silicon implementations, root security, malware, firmware, kind of innovations. This is an interesting trend because if software gets onto the silicon, to the level that there's better security, you have fingerprinting, all kinds of technologies, how is that going to impact the analytics world? So if you believe that they want faster, lower latency data, it's going to end up in the silicon. John, you described exactly why Intel's focused on the virtuous cycle of growth because there's more cloud-enabled data moves itself from the cloud through our 5G networks and out to the edge in IoT devices, whether they be autonomous vehicles or drones. This is exactly why we have this continuum that allows data to move seamlessly between these three areas and operationalizes the core missions of government as well as provides a unique experience that most people can't even imagine. You likely saw the NBA Finals, you talked about Kevin Durant, and you saw there the Intel 360 demonstration where you're able to see how, through different camera angles, the entire play is unfolding. That is a prime example of how we use back-end cloud hyper-connected hardware with networks and edge devices where we're pushing analytics closer and closer to the edge. By the way, that's a real-life media entertainment example of an IoT situation where it's at the edge of the network, AKA stadium. I mean, we geek out on that as well as the Amazon has the MLB thing. Amy Jessie knows I love that because it's like we're both baseball fans. We're excited about it too. We think that along with autonomous vehicles, we think that this whole concept of experiences rather than capabilities and technologies. Most people don't know that that example of basketball takes massive amounts of compute. I mean, to make that work at that level, this is the CG environment we're seeing with gaming culture. The people are expecting an interface that looks more like Call of Duty or Minecraft than they are like a Windows desktop machine that's what we're used to. I think that's great. That's why we say we're building the future again. You touched on something you said a little bit ago that a data officer in the federal government has got a tough job. Yes. Big job. What's the difference between private and public sector? Somebody's handling the same kinds of responsibilities but has different compliance pressures, different enforcement pressures and those kinds of things. So somebody in the public space, what are they facing that maybe somebody on the other side of the fence is not? Our data officers have a tough job, whether it's about cleansing data, being able to ingest it. What we talk about when you describe this needle and a haystack of needles is the need and ability to create a hyper-relevancy to data. Because hyper-relevancy is what makes it possible for personalized medicine and precision medicine. That's what makes it possible for us to do a hyperscale personalized retail. This is what makes it possible to drive new innovation. It's this hyper-relevancy. And so whether you're working in a highly regulated environment like energy or financial services or whether you're working in a federal government with the Department of Defense and intelligence agencies or deep space exploration like at NASA, you're still solving many data problems that are in common. Of course, there are some differences, right? When you work for the federal government, you're a steward of citizens data that adds a different level of responsibility. There's a legal framework that guides how that data's handled. Well, trust is important. They're trying as opposed to just a regulatory and legal one. But when it comes to artificial intelligence, all of us as practitioners are really focusing on the legal, ethical and societal implications associated with the implementation of these advanced technologies. Quick question to end the segment. I know we're running all the time but I want to get this last point in. And this is something that we've talked on theCUBE a lot of me and Dave Vellante debating because data is very organic innovation. You don't know what you're going to do until you get into it. A little alchemy if you will. But trust and security and policy is a top-down, slow-down mentality. So often in the past, it's been restricting growth. So the balance here that you're getting at is how do you provide the speed and agility of real-time experiences while maintaining all the trust and secure requirements that have slowed things down? You mentioned a very important topic there, John. And in my last book, 21st Century Leadership, I actually described this concept as ambidextrous leadership. This concept of being able to do operational excellence extremely well and focus on delivery of core mission. And at the same time, be in a position to drive innovation and look forward enough to think about how not where you are today but where you will be going in the future, this ambidexterity is really a critical factor when we talk about all leadership today, not just leaders in government or just people who work folks on artificial intelligence. That's multi-dimensional, multi-discipline too, right? I mean. That's right, that's right. That's the DevOps ethos, that's the cloud. Move fast. I mean, Mark Zuckerberg had the best quote with Facebook. Move fast and break stuff. Up until the time he had about a billion users and then changed to move fast and be secure and reliable. Don't break anything at that point. Well, he got the, well, he understood that. You can't just break stuff. At some point, you got to move fast and be reliable. A secure experience. One of five books, by the way, I wanted to tell you. That's right. And I'm working on my sixth and seventh now, but yeah, I'll forget it. And also the managing director of the Greer Institute of Leadership and Management. So you've written now almost seven books. You're running this leadership. You're working with Intel. What do you do in your spare time, though? My wife is a chef, and so I get a chance to enjoy all of the great food she cooks. And I have two young sons, and they keep me very, very busy believing me. I think you're busy enough. Thanks for being on theCUBE. I very much appreciate it. Good to have you with us here at the AWS Public Secretary Summit. Back with more coverage live here on theCUBE from Washington, D.C., right after this.