Women@SCS conducted an interview with Stephanie Rosenthal, teaching professor of computer science and artificial intelligence.
Women@SCS: Okay. So super broad starting off; could you tell us about your background and your journey before Carnegie Mellon, or through Carnegie Mellon too because you've had a long history here?
Professor Rosenthal: Yeah, so when I was in high school, I decided that I really liked robots, and so – I grew up in the D.C. area, and I said to myself "Okay, let me just start randomly emailing government agencies and asking if they would take an intern."
Women@SCS: Why government agencies?
Professor Rosenthal: At least in the late '90s, early 2000s, there weren't a lot of robot companies. I just looked up what places did things with robots, and after my junior year of high school, I did a summer internship with the Army Research Lab doing robotics, and then I interviewed at the Naval Research Lab (NRL) which was just a little further from my house. They hired me my senior year of high school and then for at least one internship in undergrad. They became my mentors. I learned what research was. I learned how to work on really hard problems that I'd never thought about. Reid Simmons at CMU was collaborating with NRL on a project to create a robot that attended a conference, and that was the project they put me on. The robot's name was GRACE, Graduate Robot Attending a Conference, G-R-A-C, and then the E for end of conference. Everyone at NRL was like "Hey, you should apply to CMU early" and, so that's what I did. So that's how I decided to go to CMU for undergrad in CS.
Women@SCS: Very cool.
Professor Rosenthal: So in 2003 I started at CMU as an undergrad in computer science, and it was really hard.
Professor Rosenthal: Just like everybody still says. That hasn't changed. I did a lot of research as an undergraduate first with Reid Simmons, then Susan Finger on smart whiteboards for aiding student collaboration, and then Anind Dey on smart tools that could benefit families. I applied for grad school and decided to stay because of the ability to collaborate. I was co-advised by Manuela Veloso and Anind Dey, and I finished in 2012. I worked at the Software Engineering Institute as a Research Scientist for 4 years, and then I decided I missed teaching. I joined Chatham University and helped them start a Data Science major for a few years. And now I'm back.
Professor Rosenthal: So I've grown up here.
Women@SCS: That's really cool.
Professor Rosenthal: Yep. I've taken all your classes – 213. 251, took it.
Women@SCS: You never forget.
Professor Rosenthal: No, you never forget it.
Professor Rosenthal: Mark Stehlik was my advisor. Carol was the head of Women@SCS. So it's been that way for a long time. It's nice to be back with familiar faces.
Women@SCS: Mm-hmm. What was your experience like with Women@SCS? Do you think it's grown as a community since you were here as an undergrad?
Professor Rosenthal: There were fewer women in CS at the time and there were about 15 or 20 women who were Women@SCS regulars when I was in undergrad, and there wasn't SCS4All. So we had a couple of guys in it. And because it was so new – I think it had been started within like the last four or five years of when I started at CMU – the original grad students who started it were also really still involved with us as undergrads, so I remember working a lot more with the grad students. I think you don't do that as much.
Women@SCS: Oh, cool. Yeah, we're trying to.
Professor Rosenthal: So I remember that a lot. [They’re] actually like people that I still see at conferences and things like that.
Women@SCS: The small world, I guess.
Professor Rosenthal: Yeah, well, CS is a small world. But CMU is a smaller room, I guess.
Women@SCS: [Laughs] Yeah, for sure.
Professor Rosenthal: I think you're doing a lot of the same things that you were before. We did road shows. I helped start the Tech Nights [program]. I ran SCS Day. I think I have the t-shirts for like the second, third, fourth SCS Day.
Women@SCS: [Laughs] Awesome.
Women@SCS: Vintage. So how did you figure out you wanted to teach and become a professor? Where did that start from?
Professor Rosenthal: So starting in high school, I saw everyone doing research, and I was like "Yeah, research is what I want to do." I always knew that's what I wanted. It was really cool to come up with your own ideas and work on them. And then in grad school, you start to realize that when you grow out of being a grad student, most of the time you're not doing research anymore. If you become a professor, you're coming up with ideas, but mostly you’re teaching. You're teaching your grad students how to become good researchers. You're teaching classes, of course. You're mentoring undergraduate students. Those opportunities to work with students made me realize that I really like that part of being a professor. But I don't think that I really knew that I wanted to teach in the classroom as what I wanted to do instead of research until I jumped into it at Chatham.
Professor Rosenthal: The point when someone figures some concept out is just really rewarding. I'm sure you feel it as TAs and even just helping a friend in class, so I find that really rewarding. I like the research part too. I still do research. I still advise undergrads in research, but I really like the teaching part, to watch people figure out new concepts. It's a lot of fun.
Women@SCS: What was it like teaching both CS and non-CS classes, and CS classes like 110 to like 281 and Autonomous Agents?
Professor Rosenthal: Yeah, I think the hardest part about teaching a variety of courses is remembering your audience.
Women@SCS: Right. Yeah. It's a context switch.
Professor Rosenthal: It's a context switch. I have to remember that in 110 that these students have potentially never really even worked on a computer - tablets and phones are different - and help them understand how computers work and why writing code could potentially make some of their own work easier. So that's one end of the spectrum, and then the other is Autonomous Agents 482. It's a junior or senior level course. The students have taken how many AI courses and how many programming courses, and so those students, they're fully capable of doing whatever you give them. For them, I think the question is more about teaching them state of the art concepts and bringing in people from industry or other fields, for example, to show how are these things used in the real world. The assignments are not just toy problems. And 281 AI is sort of somewhere in the middle, where you're mature enough to write your own code, and think through hard problems theoretically, but you're still learning, and so creating structure to help students figure it out, but not get frustrated. I think that's that middle ground.
Women@SCS: I see. Quite a balancing act.
Professor Rosenthal: Yeah. [Laughter]
Women@SCS: So what is your current research like that you're working on?
Professor Rosenthal: So, what I didn't tell you before is, when I left grad school, I went and I worked at a startup in Pittsburgh called Bossa Nova Robotics, and what they do is inventory for retail stores. They build robots that scan the shelves of retail stores and they can tell you what's empty on the shelf so that someone can restock it, things like that. I worked there for about a year. When I was working there, I realized that debugging big systems is just really hard. You work on little cases, or you have auto graders in school, and then you get out into the real world, and your robot just like falls down.
Professor Rosenthal: And you're like "Well, there are probably 10,000 lines of code. How do I fix that?" And you work through it and start to build tools to help you fix those errors more efficiently.
Professor Rosenthal: When I left Bossa Nova, I decided that the research that I was really interested in is explainability. Lots of people here are working on that now, and a lot fewer were working on it about five years ago. This idea that a system, especially one that interacts with people, should be able to explain itself to you turns out to be really important. You could think about debugging purposes for a programmer versus a user who may not care exactly what's going on, they just want to make sure they're safe. There are lots of reasons why someone would want an explanation.
Professor Rosenthal: I started looking into this idea of explainability and providing context for robots. I was working with Manuela Veloso and our students Sai Selvaraj on how to summarize paths that robots take and also with Sidd Srinivasa and Henny Admoni and students Rosario Scalise and Shen Li on how to talk about path preferences and also similar looking objects. I still do some work on robots and explainability. If you're interested, let me know. Quick plug.
Professor Rosenthal: The bigger project that I'm working on now with Reid Simmons and Moshe Mash is trying to understand how to model data scientists as they build machine learning models, and then use those models to help data scientists learn new techniques, be more efficient and/or struggle less. As I was teaching at Chatham, I realized that explaining and teaching were very similar so one of my goals is to leverage new explainability research to determine what help to provide and when. So essentially trying to look at different ways to model what people do and then explain it, either for teaching or for other purposes. So yeah, that's what I'm working on.
Women@SCS: [Laughter] Pretty cool that there’s a natural bridge there. So switching gears a bit, how would you describe your experiences of women in computer science in general?
Professor Rosenthal: So like I said, I started in high school getting interested in computer science and robots. I went to a summer camp, probably would have been in 2000, and it was like robotics or mechatronics or something, and for whatever reason, they accepted 100 people into the program, and there were 90 guys and 10 girls.
Professor Rosenthal: I don't think that you realize what gender parity is here, like you have 50 percent women in your classes, until you go into a situation where you're like "I'm the only one." I had that before I even came to CMU. In fact I think I was a judge for a robot competition in high school also, and someone asked me that question. Why would I want to go into computer science if there were so many men? And I sort of turned to them and said "Well, I didn't really notice." [Laughs] It didn't bother me, but I think there are always times when you're going to feel isolated, whether it's because of your gender or for some other reason, and so trying to come up with ways to meet people on their level is an important skill.
Professor Rosenthal: I think there were 30 percent women when I was an undergrad, so it wasn't 50 percent yet. I think CMU today is better for women for mentorship as well compared to what it used to be. I hardly remember having any women professors in my classes in CS. Having role models, including professors, is important. It’s important to try to find the people who support you and I think everyone likes to find someone who is similar to them. [Laughs]
Women@SCS: Yeah, naturally.
Professor Rosenthal: Um, so hopefully I can be that for people now.
Women@SCS: Yeah, definitely.
Women@SCS: The next question is, how do you manage the time to do everything you do as a professor and also doing research, and do you have any time management strategies you’d like to share?
Professor Rosenthal: [Laughs] Yeah, so I'm a big fan of lists. I just threw out my big stack of Post-It lists, but you can see I just have lists everywhere. I have three kids under six, so time management is necessary at work so I can be home for them when they need me.
Women@SCS: Oh, wow. Wow.
Professor Rosenthal: Really necessary. My strategy is to keep work at work, try not to do work when my kids are around or awake, with the exception of when they're watching TV and not paying attention to me anyway. [Laughs]
Professor Rosenthal: So the strategy is to have a designated work day, and I will work my butt off during those hours to get everything done. I make lists to help me figure out what to do. What's the top priority today? What do I need to get done? Start crossing it off the list. I heard from someone else to schedule "deep work time" where you shut off your e-mail and your phone and you just get work done, and unfortunately, or fortunately, with teaching, a lot of the work is handling things online so that strategy doesn’t work for me.
Professor Rosenthal: So I think that is my strategy, just like really trying to stay focused.
Women@SCS: Cool! Those are really good tips. I guess those would be even more important if you do end up teaching those two courses next semester.
Professor Rosenthal: It's important to also remember that everyone has their own things that they're dealing with, whether or not it's about kids and two classes, or if it's like the one really hard class that you're working on, and that's it. I think any of those things can be hard at any particular moment. I try not to compare myself to anyone else.
Women@SCS: Right. Yeah. So do you ever get any spare time?
Professor Rosenthal: [Laughs]
Women@SCS: And if so, what do you like to do with it, or what's your recharge activity?
Professor Rosenthal: Honestly, it's just really hard to schedule that time. It's important and it's really hard. So what I did for this semester, because I'm only teaching Mondays and Wednesdays, is with a friend of mine I scheduled for Tuesday mornings a stained-glass making class for us outside of work, and I just don’t come in till noon on Tuesdays. I don't get that me-only time on the weekends necessarily, so that’s what’s working for me right now.
Professor Rosenthal: Some of my glass robots on the wall, I made those too.
Women@SCS: Oh my God. They're so cute.
Women@SCS: Thank you so much.
Women@SCS: Thank you so much.
Professor Rosenthal: Have a great day.