Beth Mynatt | Bringing Researchers Together

This episode’s guest is Beth Mynatt, Professor of Computing at Georgia Tech. In addition to her pioneering work in Human Computer Interaction and Ubiquitous Computing, she has devoted significant energy in recent years to research administration. In our conversation, we talk about her own research in human-centered health technology, her role facilitating interdisciplinary collaborations at Georgia Tech, and her efforts at the national level as chair of the Computing Community Consortium.

Bio

Dr. Elizabeth Mynatt is the Executive Director of Georgia Tech’s Institute for People and Technology (IPaT), a College of Computing Distinguished Professor, and the Director of the Everyday Computing Lab. The Institute for People and Technology (IPaT) serves as a catalyst for research activities that shape the future of human-centered systems, environments and technologies to promote fulfilling, healthy and productive lives. Dr. Mynatt is an internationally recognized expert in the areas of ubiquitous computing, personal health informatics, computer-supported collaborative work and human-computer interface design. Named Top Woman Innovator in Technology by Atlanta Woman Magazine in 2005, Dr. Mynatt has created new technologies that support the independence and quality of life of older adults “aging in place,” that help people manage diabetes and cancer, and that increase creative collaboration in workplaces.  Mynatt is also the Chair of the CRA Computing Community Consortium, an NSF-sponsored effort to engage the computing research community in envisioning more audacious research challenges. She serves as member of the National Academies Computer Science and Telecommunications Board (CSTB) and as an ACM Council Member at Large. She has been recognized as an ACM Fellow, a member of the SIGCHI Academy, and a Sloan and Kavli research fellow. She has published more than 100 scientific papers and chaired the CHI 2010 conference, the premier international conference in human-computer interaction. Prior to joining the Georgia Tech faculty in 1998, Mynatt was a member of the research staff at Xerox PARC.

Transcript

Andrew Miller:             Beth, thank you for talking with us today. What first got you interested in computer science research?

Beth Mynatt:                I started out in electrical engineering and that was relatively short-lived in my undergraduate days because I became enamored with the concrete problems where computers could do something. It was much less abstract than the kinds of problems I was working on in ECE. Then, at that point, I went to a very fast progression I went through a very fast progression of isn’t artificial intelligence cool, so I was working in that area. I was also working in software engineering understanding how people were trying to build these systems. Then, I discovered HCI and that connected, for me, the difficult part of understanding what makes people tick, and then how to build computing systems that connect, amplify, or somehow engage that in useful ways.

Andrew Miller:             In your research you’ve attacked this from a number of different weight date directions. Your dissertation work was more focused on accessibility and you build the system called Mercator. Would you care to go back in time with me and give us an idea of what to that research was all about?

Beth Mynatt:                It really is going back in time because that was at the point that graphical user interfaces were just starting to appear. You have to go back to the world of … The very first little folder icon on the screen. What happened is I was loosely affiliated with a group that was just playing around with things in multimedia, because that was the term of the day. There were a whole bunch of people that worked as some form of computer users computer programmers who were blind because there were these cool things called screen readers that could take the text on the screen and send it to a braille device or send it to a speaker. The graphical interfaces threatened to do to destroy this entire area of occupation and work because no longer with the information on the screen text, but it was just images. There was no obvious way to translate that information.

That was just too much of a challenging problem for us to ignore. It just sounded like too much fun to work on. We tackled this notion of when you have pictures, when you have a graphical representation what is the best way to convey that for someone who is blind? We very quickly felt that just if you’re scooting the mouse along the screen and it’s saying, folder X and folder Y, and left, right, and up, and down are immaterial that trying to just describe visually what it looked like wasn’t the point. We build a system that got back into the guts of the interface in terms of how information was grouped and what were the interaction mechanisms selecting something, choosing from a list, inter-information in a text box, those little UI widgets that everyone was starting to work with, and then did it on a widget by which it level of how to make that information productively useful for a blank blind computer person who nevertheless it’s working with cited colleagues.

Andrew Miller:             It seems like this idea of designing for those who maybe aren’t considered by the traditional systems as a theme because you then, after some time, at PARC, the Palo Alto Research Center, which you could definitely talk about, shifted first to studying “aging in place” technologies for older people to remain active in their own homes. Then around, I don’t know, 2008, 2009, or so, switching almost completely into engaging in health and wellness. Do you see that shift? What do you see as the common theme and then where you made a turn?

Beth Mynatt:                I encourage this when I teach classes, as well, which is that people are actually pretty horrible at designing for themselves. And that’s counter-intuitive, but you make too many assumptions, and you don’t delve deeply enough in the problem to get really good design-insights.

And so, one of the best ways to do really good research, and do really good design, is to focus on users who are not you. And to use that as a forcing-function to deeply understand their needs, their goals, the barriers to using computing technology to achieve those goals, and then trying to address it.

So when I came back to Georgia Tech as a faculty member, we had started just the very first few days of the Aware Home Research Initiative. And, again, luckily realized that designing for ourselves in-home environments was not terrifically interesting. But if you looked at user groups who could truly benefit in substantial ways from advances in technology in the home, older adults was a clear target for us.

And the scenario was very much about an older adult perhaps living alone with decreasing cognitive and physical capabilities, but nevertheless, wanting to retain independence, retain quality of life, in their own home as much as possible. And that turned out to be such a rich research area, because once you embraced that group, and their needs as the challenge that you’re going after, it exposed challenges in caregivers and everyday cognition, and forgetting things, and social interaction, and how important that was, and how can you foster social interactions with older adults?

So as soon as we shifted from thinking about ourselves to thinking about others, the quality of the work increased exponentially.

And then, from that, I had taken my ubiquitous computing work from Xerox PARC, and looked at a variety of, what I called, “everyday environments”. So, the home was an everyday environment, where we looked at everyday computing. Offices were another type of everyday environment, and we looked at questions of collaboration and looked at questions of creative expression and creative work.

And so that lasted about a decade in my research group. And part of what drove me to make a pivot in my group was as, especially, mobile technologies, and sensors on the body, and sensors on the environment, because more pervasively available it opened up whole new avenues around healthcare that really just were not accessible to us in that first decade.

And so, again, it was a sitting on the forefront of where the computing field was, so I started with the invention of graphical interfaces and then basic technology in the home environment. Well now we had all this great mobile-sensing and analytics available to us and what was immediately available and interesting to me, at that point, was how people use these technologies to make sense of and improve their health on their own terms.

And that was really important because a lot of folks were building applications that were variations of “do what the doctor tells you to do.” And we know that people don’t do what the doctor tells them to do. And if you put that in a mobile device it really doesn’t change people’s behavior. But if you actually meet, again, people’s goals about what they want to accomplish around their own health and wellbeing, then you can support those goals and actually garner improved healthcare outcomes in the long run.

Andrew Miller:             Yeah, and so, for full transparency, I joined the Everyday Computing Lab in 2008, when this transition to healthcare environments and people managing their own health was, sort of, fully underway. And one of the other aspects, of this, that I noticed, was the involvement of all of these other people: these non-computer scientists we had, the main experts you were working with, nutritionists, and then in the work we did together there were school teachers, administrators, all of these people started being involved in the research. And, to me, that seemed like something new.

Was that always, even back to Mercator, were there always those people involved in the research process? Or was that an evolution?

Beth Mynatt:                No, I think it became more and more important domains that start to border on medicine, it was really important that we deeply understood and partnered with line-computer programmers when we did the Mercator work.

And I worked with some great folks who were actually employed with NASA, and that’s one of the reasons that NASA supported the work. So we got into the depth of “What does it mean for something to be accessible?” Especially in that real-world context where these people don’t work alone, but they work along-side colleagues. And so that’s what exposes that interesting tension of the design challenges. You can’t do the simple solution that works as a demo, but doesn’t work in an authentic settings.

So I think that’s always been the tenor of how I approach inter-disciplinary research in HCI, you find those subject matter experts, those informants, those individuals that deeply understand the people and the challenges and the goals that they’re after and the challenges that they face, and then you bring them in as first-rate collaborators on the team.

And so when we started to work on health and wellness, we started to work exactly with the group that you described. People wanted to improve their nutrition so we worked with nutritionists. People wanted to be able to set achievable goals around physical activity, so we worked with trainers and coaches.

For your research project, which you know well, we wanted an understanding of physical activity by adolescent kids. Well, much of their social engagement and the place where you can intercept and work with these kids, is in a school-setting. So we had to understand what was possible, in the school-settings. And the fact that the role of the teacher, the educators, in those environments, has a profound impact on whether your technology innovations are gonna be successful, it’s part of the authenticity of that research program.

So, for me, interdisciplinary HCI research is always finding that expertise and creating the trust in what’s necessary so you can have a robust research collaboration.

Andrew Miller:             Yeah! And if I were to characterize now in your research group, is that you’ve taken that one step further, where you now have these research partnerships where the clinical collaborators are also using your collaboration to answer some of their research questions, from their perspective, which is even more than doing good user-center work. That seems to be this intersection of two parallel research agendas. Is that something that you actively look for or do you still feel like mainly they are consultants and participants?

Beth Mynatt:                No, that’s a really good question.

So I think what I’ve been able to achieve in the research program, as I’ve cultivated over the years, is we gain expertise, and folks trust us that we do good stuff. That we’re able to then move from, like you said, using subject-matter experts to help us understand a problem, to moving to research-collaborators that …

For example, on the clinical side, the research-collaborators want to improve health-outcomes for breast cancer patients. So in the current project, we’re looking at this from an HCI, designing personalized mobile technologies, and so we have our goals and what we hope to achieve and accomplish from a human-centered perspective, and then they have a, very much, a clinical outcomes perspective. And now we’re able to align those two teams up.

And, you know, what was great about that was, it enabled us to break through and get funding from The National Cancer Institute, and to be part of a larger agenda around combating breast cancer. That’s a pretty far leap for us from the very beginning of just starting to work on how individuals want to improve their own health.

Andrew Miller:             So maybe shifting gears here for a little bit, another aspect of your work, at Georgia Tech and beyond, has been running these research institute and centers that try and connect people together.

So, the Aware Home Institute, I think, was your first experience in this. Could you give us just a whirl-wind tour of, sort of, what you’ve learned across all of those that you feel is how to get researchers to connect with each other, or what is the core function of these kinds of organizations?

Beth Mynatt:                That’s an important question.

So, whirl-wind tour of 15 years of research administration.

Andrew Miller:             Thank you!

Beth Mynatt:                What we’ve learned until all of the stumbles around the Aware Home Research Initiative, and then coming in a few years later to run the GVU Center, and then starting this new institute for people and technology, it comes down to some really core principles, but it’s amazing how difficult those are to achieve, in the beginning.

Which is that you create a place to convene researchers from different fields. So, the Aware Home, for example, was computer scientists and engineers, but also professors of psychology and professors of design. And you have long and sometimes painful meetings and discussions with them to understand their language, to understand what they value, to understand what sets them off. And then use those discussions to figure out joint ways that you can work together where you both desperately need each other to be successful.

I had the great fortune of working with Wendy Rogers, in psychology, as my counter-point to my computer science and HCI background for the Aware Home, and Wendy brought with her tremendous expertise of understanding what the needs of older adults were, and what works.

Not only that, but how do you validate that, and evaluate that, in authentic ways that is credible to her community so that not only computer scientists can think that something is cool, because what it does in the Aware Home, but a bunch of psychology folks can also get equally excited about the types of impact that you can have around cognitive aging and how the needs of older adults can be addressed.

So, the real trick is to create a joint agenda, and then do everything that you can to support that. And that can be time-consuming and exhausting. It means that I’m working with Wendy on writing papers that, maybe, none of my HCI colleagues would ever see, but I’m contributing to her success, and contributing to making contributions in her community, and then she’s doing the same for me. And we would have great debates because I would come at it with an HCI design perspective, and I would have probably 20 different design variables in my head as I’m trying to craft a great user experience for these technologies, and Wendy’s head would explode because she would be like, “Okay, what are the fixed variables, and what are we evaluating, and how do we use an experimental framework to evaluate this?”

And so, we had to … I remember negotiating that dance of, okay, here’s where Beth is kind of a free-ranging, user-centered, designer, and here’s where we’re nailing things down enough to do a rigorous laboratory evaluation.

And so we had to understand each other’s paradigms for how we do research. And then, from those experiences as an investigator, I then went to try to craft communities where I could help make that happen over and over again for other teams.

So, in the GVU Center, it was about bringing together, again, designers, engineers, psychology, a bit around business and policy, and the arts, to say, “okay, how do we amplify human abilities through the creation of new technologies, but doing so in a way that, again, contributes to all of these different fields?”

And it’s about creating teams, creating C-Grant opportunities. A lot of the trick is you create research funding opportunities where people have to be equal partners, on both sides, to be competitive for the grant, and then you hold them accountable to that as they do that work.

And then creating venues and ways of talking about the research so that people learn from each other. The GVU is famous for its Brown Bags on Thursdays at lunch, which aren’t really brown bags, because we provide food, but in the first six months, people brought their own food, and it’s been called “Brown Bag” forever, “Brown Bag Lunch Seminars”.

And we’ve had collaborations start because one person comes in and says, this is the world that I work in, and the kind of research and things that I want to do, and someone else from a completely different field stands up and says, wow! That’s really interesting. Maybe we could work together on that.

But that wouldn’t have happened if we didn’t create these interdisciplinary research seminar series that brought those folks together, and fed them, which helps keep them in the room.

Andrew Miller:             As a former grad student of the GVU, I can attest to the value of feeding people.

Beth Mynatt:                The food matters.

So, I did that for five years at GVU, and really understood the boundaries of doing that in an academic research environment, and then was given the opportunity to found the new Institute for People and Technology at Georgia Tech, which essentially took that formula for GVU. And then we did our best, and continued to work to expand that to include an even wider range of disciplines.

So we do a lot more work with public policy now, and then, in a wider range of basic research through applied research at economic development. And in both of those cases, what you can do is take the goals of anything that we would’ve done in GVU, and try to close the loop to making that a reality in the world today.

So, for example, we can design new medical or healthcare interventions, but then the policy folks also pick that up and they look at the barriers to those interventions ever becoming a reality, such as reimbursement regimes and regulations. Or, it’s a great tele-medicine design but if you can’t have cross-state tele-medicine, then no one is ever going to use it. So that’s great work by the policy folks, which is to understand the impact of policy and how it supports innovation, or inhibits innovation, and so we put together out teams that are doing that deep human-centered work, and then connect them with the policy folks, and then the business folks, and everything that’s necessary so that great ideas become a commercial reality, or become a reality one way or the other, whether it’s for profit or not for profit.

So it’s a more expansive set of inter disciplinary partners. And I will admit that the basic research, and the applied-research folks, seem to have just as much trouble getting along with each other, initially, as your engineers and your designers.

So there’s the disciplinary gaps, and then there’s where you are on that spectrum of research, and we’re now trying to manage both of those dimensions.

Andrew Miller:             Do you consider much of what IPaT facilitates to be computing research? Is it something else? Should we still care about those labels?

Beth Mynatt:                It’s computing, at-large, but if I called it just computing I’d have a bunch of folks quite angry with me, because it involves architecture, and it involves business, and it involves policy, and so, if I just labeled it computing-research, a bunch of other disciplines would just simply say, “Hey, I don’t hear myself in that language.”

And so, we have a slightly bulky mission statement that talks about computing systems, and environments, and experiences, because not everyone thinks of themselves as an expansive computing researcher, even though there’s probably some information technology in there somewhere.

So it’s important to craft language where people see themselves. And they hear their values affirmed in the language of the center. And we have to pay attention to that in our vision statements, and our calls for C-Grant and research proposals, and the way that we communicate the results of our research back out to the world. It’s easy to fall in the trap of only using language that speaks to a sub-set of the communities involved.

Andrew Miller:             So, as if that wasn’t enough, at the same time that all of this has been happening, you have also become the chair of the Computing Community Consortium, which, I read from their statement, “is an NSF-sponsored effort to engage the computing-research community in envisioning more audacious research challenges.”

Beth Mynatt:                Yes!

Andrew Miller:             So, I’m wondering if you could, especially for people who are outside of the US-context, if you can just give us an explanation of what is the CCC? And how did you become involved in it?

Beth Mynatt:                So, we came into being alongside, timing-wise, just about when Obama was elected president. It was a tremendous ride of so many initiatives and interests coming out of the White House, there was tremendous interest of saying, what can computing research do for the country?

And so, I became a quick learner in many aspects of computer science that I hadn’t paid attention to since I went through graduate school, and worked across the council, and with the different parts of the community, to help put forward these visions. And luckily for me, since I’ve already talked to you about health, the Obama administration was very interested in health-information technology. So I ended up having a role in the council, focused on how the roll-out of health IT and electronic healthcare records needed to be supported by robust research to make these systems effective and usable for clinicians, as well as patients.

Andrew Miller:             And so, the idea is that this is a body at the US federal-level that helps set policy, but then also gather researchers together for agenda-setting events, or producing white papers? How would you characterize what mainly is the work of the CCC?

Beth Mynatt:                So, we’re about envisioning audacious research challenges for the community. And it was actually formed at a time that we were afraid the computer science community was getting a little conservative. We had a reputation at the NSF that we were really good at turning and shooting the arrows inward, which meant that we were really good at criticizing each other’s proposals as opposed to bonding together as a community to explain to the NSF and others why our research was so important, and what it would achieve.

And so this was an attempt to intervene in that cultural shift, as our research, across competing, was starting to look a little more incremental and conservative, and that’s certainly not how you’d describe the competing research ecosystem today.

And so, we work with members of the community. That’s academic researchers, professors, industry researchers, but it’s also folks at different agencies that support competing research, to be able to answer the question, what does our community need to do in this area?

And so, health IT was one example. And so, I helped a forum, one of many workshops, that brought together clinicians, medical researchers, and HCI and computing researchers, and put them all in the same room, and go through that really difficult process of language and respect, and who does what, and how could we work together to help foster new collaborations that would come out of this group, when new funding became available?

And so, the CCC helped do that in health. It helped do it in robotics. It helped do this in Big Data, in security, and many areas. The important thing about that process is that we will hold, for example, workshops where everybody comes together and they argue, they debate, they present, and the important thing is that the work doesn’t stop there. But then people from that workshop will continue to write a summary of the workshop, and then they will write a white paper that’s more of a call to action in a certain area. And then CCC folks and others will then walk around, for example, DC, to different funding agencies to help created up-tick for those ideas and create new resources in that space.

Andrew Miller:             So you mentioned the hard work of getting people in the room together. I’m wondering if there are other connections that you see between your work within the university, and then your work at this national-level? Or if there’s lessons that you are learning in either direction?

Beth Mynatt:                So, I think trying to connect my Georgie Tech work with the national work has been an interesting opportunity for me. Everyone teases me around here, I tend to use scrabble metaphors of double and triple word-scores. So, if I can work on a problem here at Georgia Tech, in terms of fostering research, and then I can take all the knowledge from that and then take that to DC, to argue for why that research should happen, that’s a double word score, for me.

And then I do the same thing. I go and learn things at the DC level, learn things in the national context, and then come back and have that information influence the things that we do here.

Andrew Miller:             And I’m wondering, also, just pulling back, now that you have this national level perspective, and you’ve been in the research-administration side for many years … I’m assuming there are aspects of the way this interdisciplinary grounded research, or contextual research, gets done, that you wish more researchers, themselves, would be able to see?

Beth Mynatt:                So, it sounds funny to say this, but I actually think we need more intra-disciplinary work in computing, where human-centered researchers, HCI folks, are working alongside security researchers or AI researchers. And I’ve heard a call for this more and more on the national level, even in the past year. I keep hearing about human-centered AI, for example.

And I think, like many academic areas, as we’ve grown we’ve become more siloed. And so, while it’s great that people like me reach out to schools of medicine and do healthcare research, we also need people like me to be reaching across to creating usable security solutions, or creating AI systems that are understandable and actually amplify human decision-making in some powerful ways. And because we don’t have those intra-disciplinary collaborations happening within the computing field, we’re losing out on a lot of interesting questions that our field could be addressing. I think there’s a lot of things that we can do, but we need to forge those collaborations. And that’s proven to be a little harder than I thought when I first went into this.

Andrew Miller:             So, finally, a lot of us in the future of Computing Academy, and probably the people listening to this, are just getting our careers started. What would you tell someone, like me, who’s just in the early stages of an academic career or a research career? What should I pay attention to? What should I be looking out for? Who should I be looking to partner with as I try to expand my research into interdisciplinary research?

Beth Mynatt:                So, for folks … Of course, I still get to advise brilliant PhD students, who are going out there into the field. I still think the work necessary to establish interdisciplinary partnerships is important, and it’s achievable for someone early in their career. You don’t have to wait until you get tenure to do this kind of work. But you want to establish those partnerships of equals, and understand, and budget, for the work that need to be done to make that partnership work.

And so there’s maybe aspects of how you’re budgeting out a research program, I’m thinking about the work that needs to be done that includes the travel and the engagement with the community groups, or the healthcare groups, or the teachers, or the schools, or whoever that partner is, to be able to make that work. It’s not all just about you and your research and what it’s gonna take to get that next paper out the door. So, learning how to anticipate that, and build your research team around that. And there may be ways of constructing a group with undergraduates, masters, and PhD students, and post-docs that allow you the flexibility in the team to make that a reality.

I think it’s important to realize that just cause it’s great, human-centered designed, doesn’t make it great human-centered research. And I drill this every year in my teaching, that following a discovery-process around human-centered design will always provide you insights to be a good designer, but what you need to make sure that you’re doing is understanding the current state of the art and the research and continuing to make contributions to that research.

And so you have to keep that researcher-designer parts of your brain active at both times, and talking to each other. And I think our field, sometimes, has trouble explaining how our research program will work, ’cause we kinda wave a hand and say “human-centered design will come here.” And there’s a big hole in the research proposal waiting for that mystery to be solved. And we have to do a better job of structuring the data, the questions, our approach, our anticipated results, so that others, who are not proponents of human-centered design, can see how the research will take place in a rigorous way.

And then I think the other thing that is important to do, and it’s important to take the time to do, is to be able to reflect back. This isn’t something that you do in your first year, or your second year, but third or fourth year on the way of doing a set of research projects. For example, in heath care and wellness, it’s been important for us to reflect back to, let’s say, the different models of health behavior, and the theories of how people’s action is informed by understanding of their interpretation of risk in health. And to engage in that dialogue of theory and research results, and again, that’s a way of keeping the work grounded in research.

But it’s a challenge to do because it’s a type of synthesis that, I think, other fields engage in more often than the computer science, HCI, fields do. We tend to keep moving quickly from project, to project, to project, and we need to take the time to come back and do that synthesis, because that’s how you enable to field to move at large leaps at a time, and it enables you to bring more and more people with you on that journey.

Andrew Miller:             Well, I can’t think of a better way to end.

So thank you so much for talking with us today.

Beth Mynatt:                Oh, it’s been a pleasure. You managed to capture probably three of my lectures in my class next year on the kinds of things that I talk about. So, this has been great fun.