Dilsun Kaynar

Dilsun Kaynar

Women@SCS conducted an interview with Dilsun Kaynar, Assistant Teaching Professor at the School of Computer Science.

Women@SCS
The first question is really general: could you tell us a little bit about your background and your sort of path or journey before coming to CMU?

Professor Kaynar
Sure. I grew up in Turkey. I did a computer science degree in Turkey, which was called computer engineering (but it was essentially a computer science degree). At the beginning I wasn't sure if I liked my discipline. I can't say that I fell in love with it from the start, especially I think because it was a little bit intimidating to see some people who were already programmers and who knew how to hack the systems while I was a true blue blank slate. I had to learn everything from scratch. But on the other hand there were people like me; maybe after my sophomore year or so I started seeing things that I really found interesting, like I think formal languages, something that would be synonymous to the course 15-251 here.

Women@SCS
Yeah, that's a good class.

Professor Kaynar
It was a turning point for me. I really enjoyed the kinds of problems that actually fell under that umbrella of computer science. I had never followed that before. And there were some elective courses on logic - again, I enjoyed the compilers class. So I noticed that, depending on the area, I was actually finding some branches of computer science really interesting.

Women@SCS
That's awesome.

Professor Kaynar
And that's, I think, toward the end of my undergraduate degree. I knew I wanted to be in grad school. I had a scholarship to do whatever master's degree I wanted to do in Britain, under a legal organization who supports students. So I ended up choosing the University of Edinburgh in Scotland for my master's. Of course I applied and I got accepted to a bunch of schools. I chose that school because as I said, I was interested in logic, I was interested in theory. They had a strong tradition of those disciplines, plus I somehow found Edinburgh interesting to live in.
I ended up spending a really awesome year getting my master's degree, and during that time I was entirely sure that I wanted to be a PhD. So they offered me a scholarship; I went on to get a PhD at the University of Edinburgh. My research area was programming languages. You can study programming languages in many different ways. It doesn't need to be programming day and night; you can actually look at meaning, trying to find a structure in programs. And the part that really attracted me was using mathematical, logical ways to say something precise about what the program is actually achieving, or in general something that could be written as precisely as a program, maybe some protocol, maybe some abstract model of the system, defining properties that you want to achieve and then proving that you do achieve those properties.
So yeah, I did my PhD research in that area; then I got involved in modeling and verifying more complex kinds of systems, not just something that falls into the area of programming languages. I ended up doing a postdoc at MIT with Nancy Lynch, who is known for her work on the theory of distributive computation. And after that I came to CMU.

Women@SCS
Wow. Yeah, sounds like you had a lot of experience with some very specific topics inside of CS. My next question was: what excites you about distributed computing (which you touched upon with Nancy Lynch), but this was just kind of like a brief overview from your About page as a professor here. So if you want to talk about what excites you about programming languages and what made you decide to do research there.

Professor Kaynar
I mean programming languages can tie into distributed computing. It depends on how you look at it. The programming language research I did and distributed computing research I did was all actually about modeling some computing system, and stating properties of interest about that system, and coming up with methods to show that those properties are satisfied.
And for distributed computing you end up having harder things to prove. That was interesting for me: I wanted to learn more about it. One thing about distributed computing - and when I say distributed computing I'm using it in a very broad sense, where computation is actually logically distributed, not necessarily something happened in the States or something happened in Asia. That's a specific instance. But it could be happening on the same system and I could still be calling it distributed computing because of the logic of distribution.

Women@SCS
I see.

Professor Kaynar
One of the more interesting things about it is interaction - that lots of things are happening at the same time. So different locales of computation may be coordinating, or maybe computing. So how do you actually model them? So when you say model, come up with some abstract representation of your system in some language, right? And that's where programming language comes in: how do I actually create a model of my system that is unambiguous (just like a computer program is unambiguous)? And then with distributed systems we specify properties of them, and proving them tends to be really interesting and challenging. For example in 150 we have recursive functions, right? For now you're looking at sequential code, and the kinds of programs we look at are usually easier to prove by some kind of induction.

Women@SCS
It’s all induction.

Professor Kaynar
Yeah. But in distributed computing you may need other methods in addition to that. For example, say this system model looks really complicated; so let me come up with an abstract version of that which hides some of the details so that I can prove something about that. Then, later I’ll actually find another method to say my system is actually an implementation of that other more abstract version, in that all of its behaviors are also behaviors of the abstract version about which I can say something formally. So there are questions like that that I like working on. One of the remaining challenges in the field is in the discrete setting: many languages and many methods of verification have emerged in the discrete setting. But when it comes to dealing with computational systems that deal with physical reality, the methods don't really generalize immediately. So there is a lot of interesting work going on in that sphere. Cyber physical systems, for example, is a term for some kinds of systems that I find interesting to work on, even though I'm not doing much active research right now.

Women@SCS
You said that you're not really doing active research right now, but could you talk about your past research projects or what you would want to work on?

Professor Kaynar
When I say I'm not doing active research, I'm just saying that I'm not actively writing a grant proposal or trying to publish papers as much as I would be doing a few years ago. So my most recent research – I actually still am a part of a grant on that project, and I do some work on it. It's in the area of accountability. So you may say, "Okay, what does accountability have to do with distributed computing?" But as I said, I'm interpreting distributed computing in a quite liberal sense here.
One example of the distributed computing system could be a system where some agents engage in a security protocol. In this system, there are certain behaviors that may be regarded as a security violation. So if a violation happens, somebody has to be held accountable. That somebody is not necessarily a person but some abstract agent in that interacting system. Who is to be held accountable then?
Well, maybe an action happens and is a security violation, but maybe that agent was forced to do it. So maybe they are actually not to be held accountable. Maybe there was something else in the system that actually gave that agent no choice. Looking at this question also brought us to the notion of causality, so what caused the violation. Intuitively that makes sense – whatever caused the violation should be the source of blame assignment. But then we realized, what is cause? There is a lot of literature on what causality actually is – if you go and ask different people, you will find different opinions, but in a computational setting it becomes even less obvious.
For example, the definition of causality we adopted is a counter-factual notion of causality, which means had this thing not happened the violation would not have happened. We thought that formulating this could be a building block for holding someone accountable. The most recent paper that I worked on was on this topic.

Women@SCS
Wow, that's really cool.

Professor Kaynar
Actually we found ourselves talking to people from the philosophy department here, and we have a very technical philosophy department here. You can look at their web pages. They knew everything we were talking about. We were talking about system models that were non-deterministic and using some things that we thought were unique to computer science, but they were with us all the time.

Women@SCS
So like a lot of the other faculty here, you've had quite a journey amongst various institutions. So why choose Carnegie Mellon and stay here for your faculty position?

Professor Kaynar
Yeah, Carnegie Mellon is definitely one of the best places that any computer scientist would want to be. But on the other hand where you end up ends up being the result of a sequence of events. I was a post-doc at MIT; I had already been there for a few years. I was going to look for the next thing to do for myself. And then my husband, who is a doctor, had received an offer from UPMC. It looked like a good opportunity for him, and it looked like a good opportunity to make a switch for me. Pittsburgh ended up being a place that could offer the optimum for both of our careers, anyway.

Women@SCS
It's a great location.

Professor Kaynar
Indeed. I came as a post-doc here, and then one thing led to another and I found myself as teaching faculty. I came in 2006 but I started my faculty position in 2012.

Women@SCS
Cool.

Professor Kaynar
It's a very big computer science community, as you know. I mean the computer science department is only a part of it. And I was doing research at CyLab for a while, which does security research.

Women@SCS
Interesting!

Women@SCS
What is it like teaching and working with first years? What is that experience like and what did you find surprising or unexpected?

Professor Kaynar
Right. I mean 110 and 150 are very different because one can have non-majors. Well regardless, I think for me having a chance to influence how students look at computing is very attractive, regardless of whether the student is a non-major or a major. We basically instill in you some habits of thinking or approaching problems. I find teaching at this entry level more interesting actually than teaching other levels for that reason. I was told that the intellectual level of students would be very high quality. I can call it not surprising because everybody told me, but I can say that I'm still surprised.\
In terms of working with first years, sometimes I notice that I have to be more careful about teaching first-years than teaching people at other levels. For example, this influences the way I structure the course. You try to keep the assignments frequent, the feedback frequent. Sometimes a little bit of handholding may be necessary. By that I mean sometimes students may actually not realize that they should actually be getting some additional help, that things may go worse from that point, so that kind of monitoring has to happen a little more closely. Sometimes they realize they actually are stuck somewhere but they don't know how to exploit all the support structure that's provided for them. Those are things that are specific to working with first years, I think.
Something I found very surprising, actually, although I'm not sure if it's specific to first years, is that everybody does something outside of just their schoolwork. I mean schoolwork is tough, but it's something very regular for a student to be involved in another extracurricular activity and devoting a lot of time to it. I see that this is probably the way high schools educate students nowadays, or the admissions process. They don't give up other things to just focus on coursework. Or maybe at some point some prioritization, of course, is necessary. Otherwise things might become more challenging. But people find it very natural to be involved in service, involved in their hobbies, some arts, sports, etc. I like that a lot, as long as it's manageable for the students.

Women@SCS
That is interesting. I never noticed that before.

Women@SCS
Yeah, so what is your favorite and least favorite part of your job?

Professor Kaynar
My favorite part, I think, is the fact that I do something that I really value, find meaningful and enjoy. If I open that up a little bit, I like to read – I mean my job is reading things that I find interesting, writing about things that I find interesting, interacting with people that I find interesting. And that's my job!
In a field like computer science where problem solving is very much integral to the process, I have to think on my feet sometimes. It's okay if it is just gathering some thoughts and organizing what I want to say; it's another thing if I have to solve a problem in 30 seconds. I think that's the least favorite part of what I do. But, you know, that can be fun too.

Women@SCS
Albeit stress-inducing.

Professor Kaynar
Yeah, but we're all humans and everybody knows that you can make mistakes. It's just that feeling I don’t like, where you want to be helpful at that point but you just can't deliver. If I actually sit in the my office, I will probably produce a very good answer in five minutes. However, I can't do that on the spot sometimes. That's frustrating. But again, this doesn't happen often, and it's a tiny, tiny thing with respect to all of the other things.

Women@SCS
I see. So you said one of the favorite parts of your job was writing about things you find interesting. Can you tell us a little bit more about your experience writing and publishing the book that you have?

Professor Kaynar
Oh, which book are you talking about? The Theory of Timed I/O Automata?

Women@SCS
Yes!

Professor Kaynar
I have another book which is actually based on my PhD thesis.

Women@SCS Either one of those.

Professor Kaynar
These books are more like monographs on a very specialized topic that don't fit into other formats of publishing. I really enjoy the process of breaking my own thought process into units that I can explain. That's done regardless whether it's a monograph or some other type of scientific writing, and I really like that. Sometimes you don't enjoy it in the moment, but when the entire thing comes together, it's very rewarding.
In academic life you have opportunities to do this at different scales. The book is the largest scale, but even when you write some lecture notes, or some little note on a slide or whatever, if you do it well, you can address a lot of problems. I find it very rewarding.

Women@SCS
That's awesome.

Women@SCS
Kind of switching topics a little bit but what advice do you have for women in a male-dominated field like computer science? How was your undergraduate and graduate school experience?

Professor Kaynar
I've never researched this but funnily enough, in Turkey computer science was not as male-dominated as it is here. Maybe by chance, maybe by historical reasons, I don't know. So I'm not very typical in that sense. But in Britain and here of course it was. And most of the time I wouldn't even notice it, that I would turn around and, "Oh, I'm the only woman." Usually I don't notice it until somebody calls attention to it.
I've been lucky enough to find myself in supportive environments. So I haven't really suffered from it. It would be unfair for me to say that, but I totally understand that there are issues about equality or subtler things that do not surface easily, like some feeling of being excluded.
What I can say is this: today is definitely better than yesterday. People are becoming more aware of these issues, and there are better support structures. For example, I don't know when Women@SCS came into existence, but I see that you realize that it's there. You're taking advantage of it as a community that's aware of the issues.
What I can say is just notice that people are working to make it better for women and try to make use of that support structure as you're doing now. Try to find your own voice and don't shy away based on your being a woman. Just always sit at the table, remember your self-worth. And I'm not sure if Women@SCS does this, but there are mentoring programs, for example. Formally or informally, there are many senior women who would really be happy to guide you.

Women@SCS
Definitely. This is, again, switching gears a little bit but on the topic of time management how do you find the time, or how do you manage to find the time to do everything that you're interested in?

Professor Kaynar
So I don't think anybody has the magic bullet for that. Sometimes things end up falling through the cracks. But I try to deal with it by being organized and disciplined.
For example, I'm a morning person, and I try not to waste any time in the mornings. I send my children to school, and the first thing on my mind is coming to the office. I get myself here and I focus on work because I know the morning is the most productive time for me.
And I try to follow some rules that I have instituted. For example, I try to respond to emails in a timely manner, but I try to avoid reading them when I'm really concentrating on this piece of work. I would sometimes also rearrange my schedule so that I have undivided blocks of time. I find that having fragmented time doesn't work so well for me. When I get into something, I would like to take advantage of that concentration for a while.

Women@SCS
Yeah, I guess it’s kind of specific to the person.

Women@SCS
Is there any advice that you would want to give students in general to help them succeed in their undergraduate career?

Professor Kaynar
One observation I can make is that people who end up here were really high achievers in high school, mostly. You're getting a really rigorous education here, and there will be times that you struggle. There will be stumbling blocks, so you should not take it personally. You should really try to pinpoint the source of that problem and fight with it. I think you have to find that strength to regroup and approach it again and take it from a different angle. I mean sometimes it's realizing that you have to drop this one other thing that you've been devoting a lot of time to.
Because people are really intelligent, capable, and talented, sometimes they just spread themselves too thinly. Maybe just following the teacher in high school might have been fine to have the best grades. However, at the college you probably have to do more to engage with the material at a different level so that you can really absorb it. And it may take you a while until you find your own way of doing that.

Women@SCS
Thank you so much for your time, Professor.

Professor Kaynar
Thank you for giving me this chance.

Women@SCS
Not at all. The pleasure is ours, of course.