Jennifer Mankoff

Women@SCS conducted an interview with Jennifer Mankoff, Professor in the Human Computer Interaction Institute at CMU.

To start off, can you tell us a bit about your background and your journey before coming to Carnegie Mellon?
I did my PhD at Georgia Tech, and then accepted a faculty position at Berkeley. And I came here because of a two body problem. [My husband and I] both had positions in California, but my husband was unhappy. So we planned a move, since him commuting from Berkeley to positions in the south bay would have made it hard to spend much time together. We actually only applied to CMU, and got very lucky. So that's how we got here.

Why did you choose CMU for your faculty position? Why did you only apply to CMU?
Well we had a three month old, and I didn't really want to go on a wide job search with the baby when I was breastfeeding, and deal with all the travel, and all the craziness that was involved. One of my advisors from graduate school is faculty here: Scott Hudson. He knew that I was planning on going on the market the next year, and he suggested I should apply here since there was a position for technical faculty. So we did. And, my family is in New York. So, it was a move in the right direction in that sense, now that we had a baby. So we came and did an interview and the department was very supportive of the fact that we had a newborn. They gave me a room to breast-feed or pump during breaks between my interviews, and everyone was really great about the fact that we had to juggle this. So that created a good impression -- it's such a great department -- it was an easy case for us to come here.

What's your favorite part of your job?
I really love mentoring students. Working one-on-one with students of all ages to do research -- not necessarily that they all have to go into research -- but to work with them to solve problems and to help them achieve their goals both.

Could you talk a little about your research and what you do?
Yeah, I am a bit of a dilettante right now. In the past I have worked on health, sustainability, and assistive technology as application areas. Right now, my focus is on 3D printing of assistive technology. That’s my application area -- but I'm a technical researcher in human-computer interaction, which means that I like to build systems that both solve specific applied problems, but also that generalize to others.

So when I run into an issue in building an application, for example, 3D-printing assistive technology, I dig into it and develop general solutions. Say that I realize that measurement is a really tough problem--maybe I will go out and study how people measure, and look at how we can change the way we do modeling to accommodate measurement uncertainty in general (not just for assistive technology). A lot of my work right now is in 3D printing. So, another issue with creating new assistive technology, a lot of it has to attach to real world objects. How do we support attachment in a tool for any kind of 3D model?

Or, how do we even change the modeling process so that it has more software engineering-like components? For example, modules that are specific for attachment. How can we change the tools to make that possible? I really enjoy going back and forth between very applied things, and even things that are more about studying how people do things, and then going back and asking, ‘What can we do to create tools that will enable that?’

What is the most challenging part of your job?
The most challenging part of my job? Well, I would say I spend a lot of time raising money. That can be challenging. And balancing my job with my home life is also challenging at times, because I'm also a parent and I run a household with my husband. And I'm constantly trying to make tradeoffs between those two worlds.

Is there an achievement that you're most proud of? Or multiple achievements?
[Pause]...I guess it depends what world we're talking about. I’m very proud of being a good parent to my children. I'm also proud of having overcome some very difficult health issues and still continuing to be in this position. I'm proud of the work my students have done. They have taught me many things and drawn me into new areas. And of contributing in areas that matter. I tend to only do projects that I'm really passionate about. I choose them on that basis. That's how I ended up doing sustainability work and assistive technology.

And were you always interested in computer science?
Well, I started in college as a double degree student in viola performance and computer science. I was at a liberal arts college. I didn't have a major for a long time. And then I eventually dropped the conservatory, and focused just on the college. Then, I started majoring in computer science.

OK. So what drew you to computer science in the first place?
Well I had been exposed just a tiny bit to computers in high school and before. My mom had a punch-card based knitting machine. and my dad was really into mathematics (as a hobby). And, we had a programming class in high school. Then at Oberlin, I was fascinated by the VAX boxes that we had, and the fact that I could go into them and almost feel like I was exploring this unknown world with my hands. Almost as if I were reaching out to this world, and trying to figure out what was inside it. So I was very curious about them, and I really enjoyed programming. And so when I finally had to pick a major--I had at least five or six different majors that I was considering--I couldn't decide. I really just love learning. And so the way I picked computer science was, it was the one major for which I wanted to take every single required course. I was interested in them all.

That's a good sign.
I figured it was. I mean it was sort of arbitrary, but it worked out for me. And then I had a couple of internships, Argonne National Labs, and then Bell Labs, when I was an undergraduate. And I really enjoyed my exposure to research in those settings. Then I went back in my senior year, to talk to my advisor, and tell him that I was thinking about applying to graduate school. And he said, “Oh, I thought you were going to be a doctor”. But when I told him I was going to stay in computer science, suddenly he had a whole list of courses I should take that I hadn't heard about before; compilers, and other things like that. So I did some very interesting coursework in my last year because I hadn't been told to do it before then.

So you kind of alluded to this before. I noticed that you wrote about Lyme disease on your blog. Has that affected your work?
I was diagnosed in 2007 after being sick for a year. And I was ill enough, that for part of my time here I was actually on a part-time disability. So it affected my work in a very direct way. It still does, to some extent--I have to be very careful to put health first, and I limit my travel. But it also led me to publish a couple of papers actually, on studies of people with Lyme disease, and to think about the overlap between chronic health issues and disability in new ways. It continues to be an area that I am--I've been trying...The problem is, what people need, is [for] doctors to study Lyme disease. And I haven’t always found the right way as a computer scientist to get involved with it. So I've learned a lot from it. And I think it's influenced my work. But I haven't necessarily succeeded yet, in finding the project that's going to give back to people with Lyme disease.

One of the studies we did was looking at how conflicting information online affects people's beliefs about their illnesses. That was a really fascinating study. There is actually a very large disagreement in the online world about how the disease proceeds. So we were able to show, in a qualitative sense, that depending on whether people were diagnosed quickly or slowly, it actually affected the things that they believed about their disease. And then over time, if they didn’t get better, they all converged on the same model of beliefs even though they might start out in different places. And also, in studying a Lyme disease support group, we found that it was an interesting opportunity to look at online content that was produced by end users who were trying to make that content accessible to each other. Most discussions about web accessibility focus on how an organization or business [should] have an accessible website. But more and more content is in blogs, and wikis, and support groups, and other things that no organization really controls.

And so we then went out and discovered that there’s actually, on Wikipedia, a group of people that’s monitoring the accessibility of articles there and trying to advocate for more accessibility as well. And so we began exploring what are the moderation techniques that are better or worse for making things more accessible, and what are the barriers to making end-user generated content accessible in terms of the tools that are available to them? So I think there’s a really interesting set of research that could be done there to follow up on that.

But I think [having lyme disease] also just changed … you know, having been through a chronic illness like that--which was actually the second difficult health situation I’ve been through--helps me also, I think, be a better mentor to my students and just aware of the challenges people face in balancing work and home life.

So also, back to… You mentioned that you were the only woman in your department when you were looking to -
[interjection] Most of the time I was an undergrad, yeah. One of one or two at most.

Yeah, so this is still a huge issue now. Do you have any things to say about it in general, or to a female who’d be interested in the field? Like, any general words of wisdom?
Oh, geez. I wish it were that easy!

[both laugh]

You know, I think one thing I would say is, it’s okay to go back and forth between not paying attention to the issue and paying attention to the issue. You know, there’ve been times in my life when I was actually frustrated by the fact that I was constantly noticing that I was the only woman, and it was easier to just be oblivious. And I think there’s times in your life, whether you’re oblivious or you choose to just put it to the side, when that’s okay. And then there’s other times when it’s been a focal point for me, and that’s empowered me. And so, I sort of used to beat myself up over, “Oh, should I apply for fellowships that are only for women, or should I pay attention to these issues?”. And I finally realized that I need to just sort of allow it to move through my life in different ways depending on where I’m at and what I need at the time. You know? And so I’ve stopped believing that there’s one way through that, if that makes any sense.

Right now, I actually just submitted a study that was partly driven by frustrations I was having with publishing credit and author order. So we did a very large study of bibliometric data for machine learning and HCI looking at author order and how it interacts with gender. It turns out not only is there a trend overall that women publish less, on average than men--which you can see documented in a lot of literature (for very complex reasons)--but also women are less likely to be in certain author positions than men. It’s very hard to tease out why that is and how much of it has to do with factors that have nothing to do with gender bias, but women are more likely to be first or middle author and less likely to be last author. So we were doing this study where we interviewed people about that, and we also looked at large scale quantitative data to try to document some of the relationships there. So I would say right now it’s something I would be paying more attention to than usual, and as a result, I did a lot of reading on gender issues in Academia. [I] really learned a ton about some of the trends and challenges, and it’s been a fascinating year collaborating with Jessica Hammer [a fellow faculty member in HCII] and an undergraduate, Anna Wong, and a graduate student, Kirstin Early, on this. We sort of reflect the full range of academic positions as we’ve been collaborating on this work.

One of the things that resulted from that was that I made a point of asking the SCS to put me up for full professor, which I would not have asked for if I hadn’t been doing this work. Over the course of this study, I’ve also gone through the full professor process and been approved for that just this summer (2016).

So it’s interesting how those things interact, right? Yet, at other times of my life, it really was fine or maybe even better for me to just put all of that aside, and just focus on being a computer scientist. So I think that’s one piece of it for me. The other piece is to seek out mentors who are understanding, and are willing to stand up for you, and share those experiences where you can. I haven’t explicitly done anything to make this happen, but my group has a huge number of women. I think the majority of PhD students I’ve had over the years have been female, and also the majority of undergraduates that I’ve mentored. And that’s a lot of students, at least over a 100 by now. So I think people naturally seek that out mentorship that addresses these issues when they can, and I like to think that’s a good thing. I like to think that I am able to support those students because of that shared experience.

Yeah, that’s fantastic. Can you tell us a little bit about the courses that you teach?
I teach a course on the practical side of data science which looks at the human issues that impact the quality of the results that you can get from data; all the way from ‘what is the right question to ask and the right data that goes with that question?’ to the potential for bias in sampling, to the ways in which data needs to be cleaned, to the kind of things that you might do to interpret data, to visualization, and ultimately how to present back to people. Whether it’s an interactive system, or a report, or whatever it is that you’re doing. So we just look at the whole gambit the ways in which you need to think about data--in terms of psychology, preference, design, and also the computation that goes with that, hand-in-hand across the whole pipeline of dealing with data.

So that's a course that I really enjoy teaching every spring. In the fall, right now, the course that I teach is called Software Structures for User Interfaces, and that's much more of a toolkit-centric course. So, whereas [in] the data pipeline you're creating portfolio items on a Google Appspot, in the fall course, what you're doing is deconstructing the tools that we use to build user interfaces, and understanding how their internals work so that you can do a better job of building on top of them, and expanding what they do.

When you were thinking of going to grad school was that mostly because of the research that you had done? Or did you know that you wanted to teach? How did that work?
Well, when I was an undergraduate, I went to the very first conference on women in computing. It’s called the Symposium on Women in Computing. It was held before Grace Hopper existed. And they contacted the heads of departments all over the country, and said if you have an undergraduate woman studying computer science, she can come for free to this conference.

So I was selected--not surprisingly, given the number of women in my department--and I went, and there were a lot of very senior women there who would talk about their research, and also their lives, and the challenges they faced as women in the field.

And I was--I'll be honest--completely bored with everything but the research, and also very annoyed when I got back. Because after that, every time I walked into a classroom, I noticed that I was the only woman--which gets back to the point I was making about how it’s better sometimes to just ignore that. But truly, it’s such a good thing I went, because it showed me how interesting research is and I wanted to know more about it. And so I convinced some friends of mine in college to drive to the ACM -- the yearly CSE conference which all of the faculty in our department went to every year.

And so we told them that we can take our midterms while on the road -- which is what they normally scheduled during that event -- and then we drove to Arizona from Oberlin; in, like, 36 hours, shifting it four hours at a time, and showed up, and said, “Could we please student volunteer?” [laughs] And they said, “We've already filled all of our student volunteer slots.” So we said, “Well, we're here!” So they very kindly said, “If you sign up to be ACM members, we'll let you SV [student volunteer]”. And so they let us SV, and I just had a blast, and again, was just fascinated by the research. And I came home, and got my very first issue of CACM and it had a front page article on the ‘digital desk’ in it; which became a research area for me actually (and also, something I built a similar thing to in graduate school) because I thought it was so interesting (and because it helped with a repetitive strain injury I had).

So I went to Georgia Tech, and loved it. And I still wasn't sure if I was going to stay in academia. I was actually planning on going into industry because I hated public speaking. I hated it. I was a terrible public speaker. I would go through maybe 15 practice talks before talking at a conference, and ended up in tears after many of them. That perhaps drove my advisors crazy, though they didn’t show it.

Part of why I had left the conservatory [as an undergraduate] was also a little bit of a fear of being on stage. I loved playing in orchestras, but not really solo. And so, I thought, well why would I take a faculty position somewhere where I would have to basically get up in front of a crowd of people and talk twice a week or more, all year long [i.e., teach]. But as I was approaching my job search I had a life changing experience on a side project I was involved in that changed my perception of public speaking.

As background, I was disabled in graduate school while doing my Ph.D, by a very bad case of a repetitive strain injury that began the summer after my first year. My thesis work was done under the constraint of typing only about two hours a day, even though it involved building a user interface toolkit for recognition-based input. I was able to do half an hour a day during my second year, and I worked my way up to two hours by the time I graduated. So I was very good at writing things out on paper (which did not hurt), and then being very efficient when I could type.

So I was working on a project with people who are locked in and using a sensor implanted in their brain to control a computer. This was a side project; how could you turn down an opportunity to do something like this? These were people who were so severely disabled that they can’t communicate, except by maybe looking at a word board. If you could look at their eyes, and tell what letter they were looking at, you could build up a word over time. There was a neuroscientist who developed one of the first brain implant technologies, and he had implanted a sensor in the motor cortex to tell when neurons fired. They were using it to communicate. I was invited to help out with--I was actually trying to work on some biofeedback-based training code, and also was thinking about how to build interfaces that could be used if you had a very low bandwidth, error-prone input.

But one of the days I visited our primary participant, Tim, I had my viola with me. Tim was a motorcycle riding guy, who had never been to a classical music concert and loved rock music and dumb blond jokes. We would bring dumb blonde jokes to meetings with him, because that cheered him up. And who am I to have an opinion about what he wanted to hear if he was in that position?

And the colleague who had invited me to be on my project told me to take [my viola] out and play it. I decided to play the prelude to Bach’s 3rd suite, I believe it is. I got 2 notes into it, literally. And he burst into tears within 2 notes. And so I had this moment, when I realized that I needed to put my ego aside and just keep playing. It didn’t matter if I played it well or badly. It just mattered that I allowed him to have the experience he was having. That was the very first time I had a performance where I realized that it was my ego that was getting in the way of letting the audience enjoy what I had to offer.

And somehow that changed things. After that, I was able to stop worrying about what I was experiencing when I performed, and instead give what I had to the audience, and let them take it as they chose to. So that helped a lot with my feelings about public speaking. And, then I still could only type 2 hours a day.

When I put those two things together and thought about what I would do at an industry job: if I wasn’t typing, I would be in meetings all day. Versus in an academic job: I could mentor students, which I already knew I loved, and do research, and it just sort of... So I ended up applying to one industry job and everything else academic.

One final question. Do you have any advice for current students right now? Should I go to industry? Should I go to grad school? That sort of thing.
Well, you know, I’ve been very lucky. A lot of the decisions I’ve made were more about “I’m not ready to close the door”, and less about knowing I wanted to walk through it. I knew if I went to grad school and I didn’t like it, I could leave at any time and get any job I’d get as an undergraduate. So that worked really well for me, to just keep doors open, try experiences out, and see where they took me. So I think when it’s possible to do that, that’s a helpful way to make decisions.

I’ve also come to realize over the years that those decisions are much less important than the paths you take between them, that’s just my philosophy. So I guess along those lines, for me, I’ve been lucky enough to be able to always do what I’m most passionate about, and I think that has gotten me through some very difficult times. Whether being a parent, and being able to look a crying toddler in the eye and say, “It’s important that Mommy go to work now and leave you with whomever in order to do that.” Or, getting up when I was very sick with my lyme disease, and having something to work on that I cared about. It’s been hugely helpful to have projects that I’m passionate about.

I’m also in a field where I can work on any problem, because it’s an applied field. And that means, that if at one point I was intrigued by ethnography and social problems, and I [could] go out and give myself a poor man’s training in that, long enough to be able to do that kind of work. Similarly with 3D printing--and parts of that, that have nothing to do with HCI, but it’s fun to learn, and I just think that it helps. I’m lucky to be in a domain, in a field, and a job, where I can experiment with many different things, because I was never driven by a single passion where I was going to solve one problem and worked on it for many, many years straight.

So I don’t really have an answer to your specific question. It’s different for every person whether they should go one way or another, and I’ve had students who did everything from seminary, to law, to graduate school after working with me. And they’ve all, I believe, succeeded, and done what was right for them. I believe it is important to figure that out, and take the time it takes to learn that. Because I would never have predicted at any point--maybe a few weeks into my faculty position--that I would be doing this. After I was already a faculty member, maybe then I would’ve predicted that I would stay and keep doing it, right? But up until then, I had no idea that I was going to be here, and I didn’t know if it was the right thing or the wrong thing.