 Hello everyone, welcome to the special CUBE conversation. We're here in Palo Alto, California, the CUBE headquarters and studio. An exciting next guest, Chuck Davis, who's the CTO of Element Data, doing a decision cloud. But really the story is about the role of data in decision-making, biases, cognition, all the stuff that we love around big data. Welcome to the CUBE conversation. Thank you for having me, John. So we were talking before we came on camera about stories around how decision-making is always grounded in data. Obviously it's analog data, not a lot of digital data. With the emphasis on cloud now, we're seeing data at the center of the value proposition. In fact, recent event at AWS, Amazon Web Services, IBM Think, a couple weeks earlier, you got blockchain, AI, and data in the middle for IBM and Amazon. It's cloud scale and machine learning. Obviously this is a big part of the culture we're seeing in society globally, not just in North America and the U.S., but around the world. So people are trying to get their handle on all this stuff. So let's have that conversation. Absolutely. Your thoughts on data science, where it is today, how far has it come over the past 20 years? I think it's come tremendously far. There are certain things obviously that have happened, these inflection points in the industry that have taken place. One of the big inflection points obviously was the reduction in cost of storage, which then led to the event of big data. And so now you can collect more stuff and it was cheaper to store it. Obviously you can transmit it much easier. But now I think we're starting to really look at the origins of data itself and the implications of that. And I think that that is really where we're going to spend the next decade or so, is really understanding the things that actually make up the data. I mean, we generate the data ourselves. And so we've spent a lot of time in computer science trying to figure out ways of isolating signal and noise. And the things that make us uniquely human tend to be characterized as noise. And I think we're starting to finally come around and realize, especially with the advent of AI and ML, that that's not necessarily noise, that it might be the most important signal. You know, when I started SiliconANGLE in theCUBE nine years ago, one of the slogans really was extracting the signal from the noise. And but our tagline was, where computer science intersects social science. And that really wasn't that original. And Steve Jobs had the technology liberal arts kind of, you know, science with Apple and when he was turning around Apple, that was the big comeback there. But this really was, we're seeing it now. We're seeing a lot of infrastructure shifts happening, whether you look at blockchain and cryptocurrency and changing the nature of money and decentralization, whether you're seeing things on social media. Data now is really going to be the key equation. So you guys are doing something in your company. You've been doing this for 20 years. Take a minute to explain what you're working on. And you've been doing this for 20 years, but talk about the project you're executing now. Sure, so element data is concerned with creating a, at scale, essentially a decision cloud. And so what does that really mean? Well, what it means is that if you were to, we're really focused on the psychology of data. So not just data for data sake, but the psychology of data. What does that mean? It means the way that data is perceived and the way that human beings actually make decisions. And that tends to be based on a series of trade-offs that it's based on our cognitive experiences. It's the reason why I could present you with something, you can present me with something. And if we were to look at it from a strictly logical standpoint in terms of science, we should come to the same conclusion, but oftentimes we don't. And so since we give rise to data, how do we maintain the integrity of that signal, that emotional intelligence that's contained within the data? That's what we're trying to do. Talk about the role of bias because to me, I like to study the social science aspect of technology impact, whether it's new venture creation, entrepreneurship, or just a better society impact. And everything seems to be mission-driven these days, so it's very relevant. But as you get these, and certainly the US elections have recently polarized everybody, the role of bias is actually an interesting concept. I want to get your thoughts on how bias is changing people's either subjective or objective views of things and what is bias good or is it bad, and how should we handle biases because everything's now contextual and there's not a lot of context that you go to Facebook any day these days. You see all kinds of weirdness going on, but as people are connected and are sharing the same data and looking at the same signal, this is going to be biases built into everything. So what's your take on bias and from a data science and data discovery and cognition? I'm biased against bias. So it's interesting, you can't escape bias. A bias is part of the human condition. Cognitive bias is the neurological shortcut that we take to arrive at decisions or conclusions. And so as a result, you want to be able to classify that and understand it, but you really can't get around it. And so if you have some form of classification, which is one of the things that we're working on, how do you start to classify these different biases? Then you can actually start to recognize them. If you start to recognize them in data, then you can start to figure out ways to change perception or at least to surface that these biases are present. So we all engage in them. It's just a matter of how do you effectively make people aware that these biases are present and are present in their data sets. And in order to do that, you need a classification system. So the decision cloud concept that you guys are going down, I love that idea. And you take a graph approach, it's a social graph kind of concept, decision graphing if you will, is about collecting like a thousand point, million points of light if you will data points that you're collecting together to help people make better decision making. Is that right to get that right? Or how would you describe that decision cloud or decision graph concept? Right, so the graph is pretty straightforward. You know, there's a fair amount of complexity, of course, hidden in it, but nonetheless, essentially what you have is typically for human beings, we are, or anyone, we really are focused on making a decision. The decision usually options a certain amount of criteria and then also weights or importance that we ascribe to those. And so the decision cloud is about capturing those options, the criteria, and also those weights. And that's the central, that's the central node around an individual and or their role within an organization. As a result of that, that's a big deal because if you can understand how people have formed their decisions, then you can start to walk that graph and you can figure out how people or an organization got to where they are. It's the why. You know, the web and technology has been focused on who, what, when, and where, but we still have a very difficult time answering the why and the reason why we have trouble with the why is we really didn't have an ontology or a taxonomy for human decision making. And that has huge implications for the AI and the ML space. It's the reason why if you were to look at any one of the digital assistants and ask a question and ask for help, specifically help me make a decision, there's not a corpus of decision data that those agents can rely on to help to surface a decision. Personalized medicine is something I see a lot now and it's a big trend towards personalized medicine where the users can be more proactive, less responding to say conditions, but you're seeing personalization, it's not a new concept on digital, whether it's personalized recommendation engines and or other personalization techniques, but we see that changing now with certainly as users become in more control of their data. Is that where you can bring, you guys can bring that new kind of personality behind the data because bias will drive my selection criteria of making a decision or might hinder it. So these are, this is new ground in data science. How does the role of the person get involved in the decision of me? How do you guys handle that dynamic? Because your views might be different than mine. You make different decisions based on different criteria than maybe me. So yeah, it's different for per person. So you have almost an individualized aspect of it. How do you guys handle that in software? Right, well, I mean, what you're really talking about is allowing for human expression subjectivity to be part of that algorithmic mix. And so that's typically missing. So a good example would be in a medical context using a decision tool to make a treatment choice. Maybe around a particular drug regimen, a surgical option or perhaps a hospice option. And it really, for me, it might come down to quality of life, but quality of life is a subjective measure. And so as a result, it might not even be part of the recommendation engine. So our technology allows for that type of subjective input to be present and for you to be able to place a level of importance to it. And in our world, we refer to that as irrationality as opposed to the rational measures. But irrationality is not bad, it's human. It's a signal, it's not noise. How do you know what signal is? As you look at, I mean, looking at all kinds of data, there's a lot of factors, timing, things change over time, context changes. How do you guys look at that? Because this is super important. Something that might be relevant today, irrelevant tomorrow, not understood today, understood tomorrow. So there's timings and context around a lot of things that could be surfaced. Is that part of how you guys work with the Decision Cloud? It really is. I mean, it's really interesting that you mentioned that because that temporal nature, that dimensionality of time, that's one of the things that human beings aren't good at. That computers happen to be very good at. And so from a machine learning perspective, in terms of being able to train for that temporal sensitivity, that's how we address that. So what we're trying to do is we're trying to balance and leverage the strengths of machine learning along with the intrinsic understanding of psychology and coming up with a effective ontology that is represented within a decision set and merge the two together. Talk about computer science and social science coming together, our main tagline. Because this is really, you're seeing a lot of societal impact. Certainly the JOBS Act in Washington certainly enabled non-profits to actually invest in mission-driven ventures. You're seeing a spawn of entrepreneurship go on around projects that we never got funded before. So you're seeing a lot of people doing some amazing things. So how does data come out of global scale? I mean, how different cultures come into play? I mean, you need a lot of computing power. What's the computer science intersection as computer science changes the world? It used to be tech geeks would talk speeds and feeds. And now you have a human element where it's emotional. I want the app to provide value for me. I don't really care about the speeds and feeds of a product. So certainly that is colliding. What's your view of that intersection of this computer science and social science? Well, I think that it's going to become more and more prevalent as the tools get better, right? We're getting a better understanding. Like NLP, I can remember 10 years ago, it was largely a bag of words, right? And so, and well, for some people it's probably still a bag of words, but it's gotten so much better. And so there's so much more that you can do, but a lot of it is still slicing down into basically metadata. And moving beyond that, I think that as we start to look at that intersection of psychology and sociology, that becomes really, really important in terms of how the disciplines come together because they didn't. I mean, in computer science, right? I mean, the computer scientists never talked to the linguist, right? Now, if you're credible in this space, at least on the AI side, you're working with linguists, right? You're understanding origin. And I think that the same is now coming true of psychology, right? And in terms of AI, when we look at it from a psychology standpoint, we're looking at it from a standpoint of needs, right? Human needs, which is kind of the function of psychology, as opposed to when we're looking at decision theory, right? That's basically understanding how the decision is made. The game theory aspect of it is how those decisions affect other people and the impact that they'll have in the interaction and their reaction to it. So it's really the intersection, I think, of those specific disciplines that are going to be the most exciting area of technology going forward. And they're not mutually exclusive either. There's an interplay between those decision theory, gamification, and human interaction. I mean, the successful companies of the future and of today understand that and are fully incorporating that and are attempting to embrace it. And I think that this is exciting because it's kind of like the new frontier. I mean, and anyone that tells you that they understand the human mind, boot them off the show, you know, that- It's complex. Yeah, it's incredibly complex. We don't profess to, but what we can do is we could come up with the taxonomy and a nontology for some form of classification to begin that journey. And that's what we've done at Element Data. Talk about the wisdom of crowds and how that weaves into it. I know there's some personal stories that you had around your wife with medical school that's doc and will document it. I think you guys are talking about that. But people tend to care what other people think. Certainly I noticed that on social media and people try to think what, understand, try to think that they know what I'm thinking, maybe not. So there's a lot of that going on around group dynamics and around collective intelligence. Wisdom of the crowds is a big part of the gamification, which does affect decisions and then ultimately how people feel. Everyone likes to be part of a group, should be accepted. But also, there's more data now coming out of this new dynamic. How is that data being weaved into decision-making? John's probably a whole show in and of itself. But when you talk about the wisdom of the crowd, there are a couple of things to keep in mind. First, if you look at the germination if you will of a particular concept. I mean, people will tend to coalesce around it and it tends to be around the topic, people's familiarity with it and a certain perception that they have. And if you're far outside of that perception, that's where you start to actually generate this excitement or I should say this level of engagement. So for instance, if you were to say something controversial, not necessarily expected, you're going to generate more interest. I see it all, I mean, kids do it all the time on social media, right? They do something dramatic, they say something, they know it's stupid, but they are able to generate a fair amount of interest. And hence they have a crowd that follows them. In that case, it's kind of the school of fish type of crowd theory. I think that fundamentally what you'll see though is this rise of data kind of moving in a different direction. I think that if you are able to expose the biases that people have, then they are, if they're aware of them, then they act differently. We talk about civil discourse a lot. Certainly during the election process around how can we have civil discourse amongst ourselves to have a good conversation, to surface data because there wasn't a lot of that going on. But when you get on digital, this notion of weaponizing content and creating memes, we talked on theCUBE, we've said this before, control the meme, you control the narrative, control the narrative, you control the conversation, control the conversation, you control the belief system, and then you own the populations. Kind of like that's the hacker formula. Mind hacking, it's been called. This is actually a new data opportunity to get that out. It's been arbitrage through the naivety of the newness of the web or the new social graphs. So we see some people certainly hacking that. How do you turn that into a positive data source? Because if what you're saying about biases, that should be surfaced quickly. So people kind of know what the collective group is thinking. How do we turn that mind hacking, gamification into a positive data set? Right, I mean it's interesting because people refer to it as weaponization and I refer to it as PsiOps, right? I mean it's something that has been done before in a different context and now we're starting to see it in the data context and the results are chilling because this isn't a leaflet dropping from the sky that you read, it's really about understanding who you are at an attribute level and understanding who you are from a perception level and really dealing with the psychology, either to incite you or to suppress you. So I think it's deeply concerning but I also think that there are really good opportunities for us to do things that are very positive. One of which for instance, that immediately comes to mind is the ability to allow people to understand their decision process. If you have a decision cloud, you can actually look at and see your journey, your path along a decision path and that's not something that's readily available in the world. The role of community too, we've been observing and we're digging into, I'm sure you have at some level too, the role of communities, you look at open source software, it's been a great example of successful consensus within communities and as a way to balance potential over-amplification or over-reaction or biases that could be checked or balanced together as an interesting new approach. Do you guys see any of that in the decision cloud where this new data source is coming in around communities and ecosystems? Well, I think that, what's interesting about communities is they tend to be self-forming, right? I mean, you can try to force people together but they tend to be self-forming. So if people share a particular concept or belief then there's a certain amount of attraction. I think that what's interesting is the ability to try to measure that, right? And to try to figure out how you can then expand that community with different beliefs and different viewpoints so that you get something that is not so homogenous but is more representative. And so that's something that we hope, I mean, we can't necessarily predict it but we hope that that's something that decision cloud would be able to influence. Well, it's been great to have you on the queue. I want to definitely follow up on some of those deeper conversations. But I got to ask you a personal question. You've been, how long have you been at this? How did you get here? I mean, you've been scratching this itch for how many years? I mean, how did you get to this point because has it been a lot of research you've been doing? Has it been other ventures? Tell your story. What's motivating you to get to this point? Yeah, basically the better part of 25 years, I mean, my background is both in computer science and behavioral biometrics. So I've always been interested in behavior and classification of behavior and trying to figure out from the standpoint or the discipline of computer science, how do you effectively really integrate the two? And one of the biggest riddles, if you will, which actually the code name of our product internally is conundrum is how do you solve the conundrum of decision-making? And we haven't solved it. I think we have a pretty good understanding of it. But by the same token, that seems to be the last big frontier, the last big open space. And that was something that I've pretty much worked my entire career, I think to get to this point of being able to have a phenomenal team to be able to solve this problem. Yeah, well, we'll check out element data. Great stuff. It's a systems problem now. You said it's not one thing. Right. A lot of interplay and a lot of dependencies and a lot of interaction, a lot of data. A lot of data, a lot of data. Thanks so much for coming on and spending the time. Chuck Davis is the CT of element data. Check him out. I'm John Furrier here in Palo Alto for a CUBE Conversation. Thanks for watching.