(bright music) >> Hi, everyone. Welcome to theCUBE's coverage of WiDS 2022. I'm Lisa Martin, very happy to be covering this conference. I've got Vidya Setlur here with me, the director of Tableau Research. Vidya, welcome to the program. >> Thanks, Lisa. It's great to be here. >> So this is one of my favorite events. You're a keynote this year. You're going to be talking about what makes intelligent visual analytics tools really intelligent. Talk to me a little bit about some of the key takeaways that the audience is going to glean from your conversation. >> Yeah, definitely. I think we've reached a point where everybody understands that data is important, trying to understand that data is equally important. And we're also getting to that point where technology and AI is really picking up. Algorithms are getting better, computers are getting faster. And so there's a lot of dialogue and conversation around how AI can help with visual analysis to make our jobs easier, help us glean insights. So I thought it was a really timely point where we can really actually talk about it, and distilling into the specifics of how these tools can actually be intelligent beyond just the general buzz of AI. >> And that's a great point that you bring up. There's been a lot of buzz around AI for a long time. The organizations talk about it, software vendors talk about it being integrated into their technologies, but how can AI really help to make visual analytics interpretable in a way that makes sense for the data enthusiast and the business? >> Yeah, so to me, I think my point of view, which tends to be the general agreement among the research community, is AI is getting better. And there are certain types of algorithms, especially these repetitive tasks. We see this with even Instagram, right? You put a picture on Instagram, there are filters that can maybe make the image look better, some fun backgrounds. And those, generally speaking, are AI algorithms at work. So there are these simple, either fun ways or tasks that reduce friction where AI can play a role, and they tend to be really good with these repetitive tasks, right? If I had to upload a picture and constantly edit the background manually, that's a pain. So AI algorithms are really good at figuring out where people tend to do a particular task often, and that's a good place for these algorithms to come into play. But that being said, I think fundamentally speaking, there are going to be tasks where AI can't simply replace a human. Humans have a really strong visual system. We have a very highly cognitive system where we can glean insights and takeaways beyond just the pixels, or just the text. And so how do we actually design systems where algorithms augment a human, where a human can stay in the driver's seat, stay creative, but defer all these mundane or repetitive tasks that simply add friction to the computer? And that's what the keynote is about. >> And talk to me about when you're talking with organizations, where are they in terms of appetite to understand the benefits that natural language processing, AI and humans together, can have on visual analytics, and being able to interpret that data? >> Yeah. So I would say it's really moving fast. So three years ago, organizations were like AI, it's a great buzzword, we're weary because when rubber hits the road, it's really hard to take that into action. But now we're slowly seeing places where it can actually work. So organizations are really thirsty to figure out how do we actually add customer value? How do we actually build products where AI can move from a simple, cute proof of concept working in a lab to actual production? And that is where organizations are right now. And we've already seen that with various types of examples, like machine translation. You open up a Google page in Spanish, and you can hit auto translate and it will convert it into English. Now, is it perfect? Not, but is it good enough? Yes. And I think that's where AI algorithms are heading, and organizations are really trying to figure out what's in it for us, and what's in it for our customers. >> What are some of the cultural, anytime we talk about AI, we always talk about ethics. But what are some of the cultural, or the language specific challenges with respect to natural language techniques that organizations need to be aware of? >> Yeah, that's a great question, and it's a common question, and really important. So as I've said, these AI algorithms are only as good as the data that they're often trained on. And so it's really important, in addition to the cultural aspects of incorporating those into the techniques, is to really figure out what sort of biases come into play, right? So a simple example is there's sarcasm in language, and different cultures have different ways of interpreting it. There are subtleties in language, jokes. My kids have a certain type of language when they're talking with each other that I may not understand. So there's a whole complexity around cultural appropriation generations that, where language constantly evolves, as well as biases. For example, we've had conversations in the news where AI algorithms are trained on a particular data set for detecting crime. And there are hidden biases that go into play with that sort of data. So we're really, it's important to be acknowledged of where the data is, and what sorts of cultural biases come into play. But translation, simple language translation is already more or less a solved problem. But beyond the simple language translation, we also have to account for language subtleties as well. >> Right, and the subtleties can be very dramatic. When you're talking with organizations that are really looking to become data driven. Everybody talks about being data driven, and we hear it on the news all the time, it's mainstream. But what that actually really means, and how an organization actually delivers on that are two different things. When you're talking with customers that are, okay, we've got to talk about ethics. We know that there's biases and data. How do you help them get around that so that they can actually adopt that technology, and make it useful and impactful to the business? >> Yeah. So just as important as figuring out how AI algorithms can help an organization's business, it's equally important for an organization to be more data literate about the data that feeds into these algorithms. So making data as a first class citizen, and figuring out are there hidden biases? Is the data comprehensive enough? Acknowledging where there are limitations in the data and being completely transparent about that. And sharing that with customers, I think, is really key. And coming back to humans being in the driver's seat. If these experiences are designed where humans are, in fact, in the driver's seat, as a human, they can intervene and correct and repair the system if they do see certain types of oddities that come into play with these algorithms. >> Going to ask you in our final few minutes here, I know that you have a PhD in computer graphics from Northwestern, is it? >> Yep. >> Northwestern. >> Go Wildcats, yep. >> Were you always interested in STEM and data? Talk to me a little bit about your background. >> Yeah. I grew up in a family full of academics and female academics. And now, yes, I have boys, including my dog. Everybody's male, but I have a really strong vested interest in supporting women in STEM. And I actually would go further and say, STEAM. I think arts and science are both equally important. In fact, I would say that on our research team, there's a good representation of minorities and women. And data analysis and visual analysis, in particular, is a field that is very conducive for women in the field, because women tend to be naturally meticulous. They're very good at distilling what they're seeing. So I would argue that there are a host of disciplines in this space that make it equally exciting and conducive for women to jump in. >> I'm glad that you said that. That's actually quite exciting, and that's a real positive thing that's going on in the industry, and what you're seeing. So I'm looking forward to your keynote, and I'm sure the audience is as well. Vidya, it was a pleasure to have you on the program talking about intelligent visual analytics tools, and the opportunities that they bring to organizations. Thanks for your time. >> Thanks, Lisa. >> For Vidya Setlur, I'm Lisa Martin. You're watching theCUBE's coverage of WiDS conference 2022. Stick around, more great content coming up next. (bright music)