In Conversation With: Dr. Marie Piraud

Dr. Marie Piraud is currently a Postdoc at the Image-Based Biomedical Modeling group at the  chair for Computer-Aided Medical Procedures (CAMP) at TUM. She obtained her Bachelor’s and Master’s degree in Fundamental and Quantum Physics at Ecole Normale Supérieure de Cachan and Ecole Normale Supérieure de Paris, France. She continued to obtain her PhD in Physics in Laboratoire Charles Fabry, Institut d'Optique, Palaiseau, France. We recently met up with Marie and chatted about her academic journey, her current research and the struggles many females face with imposter syndrome in academia.

What are you currently researching at the CAMP chair in our department?

My research focuses on modeling image-based medical data in order to get insight into the underlying biological processes. I work on tumor growth and cancer development. More specifically, I’m currently working on one blood cancer called Multiple Meyloma.

The research has evolved from a few years ago; in the past more classical statistical and computer vision methods were used, but we now leverage deep learning techniques. With deep learning we can analyze the medical images and accomplish tasks such as tumor detection.

The research can be varied, depending on the disease and what the doctors with whom we collaborate want to achieve. Our ultimate goal is to assist them in making better treatment planning. As the way cancer develops is highly patient specific, recognizing the most aggressive sorts of cancer would enable doctors to plan better. Our modeling can help with decisions such as who you should treat, when to treat them and how you should treat them. Deriving different stages for specific diseases is one task, and assigning patients to different risk classes using simple but relevant criteria is another. The research we do as data scientists at our group ultimately attempts to optimize and simplify this process for the doctors, depending on their needs.

Is deep learning something you were doing before your current Postdoc position?

No, I wasn’t doing deep learning before, but I’ve come to really enjoy the computer vision and deep learning part of my research. The final goal of the research is not necessarily some type of computer vision output - but rather leveraging deep learning and computer vision techniques to gain insight into what is going on with the cancer and how it can be treated, which is achieved with modeling, and doing modeling was definitely in my skillset before I started working with deep learning.

So speaking of skillsets, how did you end up at the CAMP chair working with deep learning coming from a Physics background?

I was doing theoretical physics where the goal was to attempt to model complex systems, and trying to gain insight and interesting underlying natural laws of these systems. In theoretical physics you also perform many simulations: you start from the general rules that you know, make some approximations, create some effective models and then you simulate.

I was already coding a lot and I enjoyed programming my own models - my work was very interesting! However, I was specialized on a very specific topic, which gave me the impression I could only end up in a handful of universities in the end. As I didn't want to end up pigeonholed, I thought that a transition to data science and machine learning could be beneficial to my career. I also believed I could also succeed in this area, given my skillset as a physicist.

When I started applying to various positions after my first Postdoc, I figured I could also apply to do another one to help me in this transitional phase. Bjoern Menze, my current advisor, proposed this Postdoc to me. As he also studied physics he was understanding of my situation and encouraged me to come and transition to this new line of work at his lab.

What were the differences and new skills you had to acquire after transitioning over from a purely Physics based environment to the research you’re doing now?

For this position I had to learn a lot about statistical methods and read quite a lot of new books. With my scientific background, I was able grasp these new concepts and ideas quite quickly, but I still have a lot to learn!  

Another difference is the role of publications. In Physics I read many journal publications, and we’d mainly publish in scientific journals, but Informatics is very conference based. There are deadlines all year long and you try to push your papers to conferences as fast as possible. I personally have the impression that conference publications are much less well written. But on the other hand, as papers in Physics are published much slower, we weren’t always necessarily spending our time on valuable actions. It was not uncommon to polish publications for sometimes up to three months, where each and every word was weighed and debated, and every color of every curve was discussed for half an hour.

You were a Marie Curie Fellow, and now are a Marie Curie Alumna. Can you tell us a little bit about what that is and what it entails?

The Marie Curie organization award prestigious EU fellowships for PhD students, Postdocs, and Senior Researchers. The idea is to allow scientists to complete world-class research while undertaking mobility within Europe. You apply as an individual and are awarded a grant - in my case it was for a Postdoc position for two years. There’s also a community network behind the Marie Curie Fellow title and it is very competitive due to its prestige.

What factors do you think allowed you to be successful in applying for this prestigious grant?

I got very good advice and help from friends who had been awarded it before. Since the grant is awarded by the EU, the most important factor was to show that my project could tie links within the European research community. For my application I didn’t only focus on the scientific part, I tried to also focus on how it would help me help the research community, and how it could tie new links and new collaborations within Europe.

What draws you to stay in academia and where do you see your career progressing? Do you want to be a Professor?

I don’t absolutely want to become a professor, not at any cost, but I really like research and the freedom it provides me. The intellectual stimulation is something I also really value. However, I definitely think that it is possible to find jobs in industry which can have similar types of challenges.

Both Physics and Informatics are fields that have a higher percentage of males in them. Do you think this is an issue? And if so, what do you think would make these fields more appealing to women?

I think there is a lot to gain from having more female students. And I think women in general can be very thorough: they can go into great details and be very diligent. They can also be very stubborn, in a good sense, which can be helpful in research.

I think having more women in our field would naturally attract women, so we need to make the field more women-friendly. I think women want to have a balanced life, so working in an environment that encourages a healthy balance will help. I think that having role models can be very beneficial; despite my achievements, I am still a bit insecure. And while my confidence is improving with time, when I started my PhD I was very insecure and feared that it would be a total failure. Having a close female role model would have helped me tremendously.

During my PhD, I had a good advisor and I would say to him, “I can’t make a paper out of this, it’s nothing” and he’d respond “It’s research, and research is most of the time incremental.” Despite my hesitations, I’d submit the papers, and most of the time they were accepted and  published. After a while, as you do this several times through your PhD career, you realize: this is what it is. Sometimes there are big breakthroughs that seem like big shifts to you, but most of the time research really is just small incremental discoveries. You have to put your insecurities aside and try to publish despite them.

I’ve found that the insecurities I’ve had have gotten better with time and with practice. I also don’t want it to be totally away, because I feel like having too much self-confidence can also get in the way.

Do you think this type of insecurity, some could say it is imposter syndrome, is more prevalent amongst women?

Definitely. I have the impression that guys don’t even ask themselves these questions sometimes. There are of course exceptions, but I feel that the average male PhD student doesn’t doubt themselves as much as the women do.

Another experience where I’ve seen this is in the Marie Curie actions program. For three years I was involved in an outreach program where I went to the French high school of Munich. The idea was to expose female students to hard sciences and to help them figure out what they want to study in university. I went there three years in a row and I tried to be a role model, to give off an impression to the girls that “I’m young, I’m normal, but I do hard science, and I like it”, and hoped it would help at least some of them to pursue a science related career.

We would sit in small groups with the girls and had discussions where they could ask questions for an hour. I’d usually ask them if they like physics and math and I would get responses like “Oh, yes, I like it!”. But they would tell me that they didn’t want to go into science in university, and when I asked why, they’d always respond with something like “I’m better at languages, so I’ll do that”, or “I’m not good enough”. When I’d inquire about their grades they would usually say something totally reasonable, not necessarily the best - but totally reasonable marks. Then I’d ask the girls - “the guys in your class with the same grades, what do they want to do?” and the response was of course “science”.

But the guys just don’t ask themselves, “Am I good enough?”. The girls, ask themselves that, and the answer is usually degrading - their response is always “No, I’m not, because I’m not the best.” The girls would always feel they’re not good enough. I feel like this is something society has ingrained in females from a young age, and this is a very big problem.

Outside of your busy Postdoc work, what do you do in your free time for fun?

I try to do some sport - I like climbing and yoga. I also enjoy going out to eat with my friends. My free time has also improved a lot in the last few years! During my PhD, I was very hard working and didn’t have much free time, but I realized that if you don’t know what your own limits are, you should find out. It’s good to make breaks to refresh your brain.

I definitely still tend to work hard and late when I have energy - but I never work on weekends anymore. I had to learn that! At the beginning of a PhD, you feel like you always have to work, but it does not necessarily improve your efficiency. Everybody has to find their groove, and I have to make a small break every weekend.  When I come back to work it’s with a more critical eye to what I did the week before and I find that I am much more productive.

What advice do you have for women who are thinking about pursuing a PhD in Informatics or any other hard science?

When choosing your career path or your PhD topic try to find something that really gets you excited and don’t overthink - let others judge if you are good enough for a position. And finally, being happy at work is a big part of being happy in life.

Published November 2017