The PEER Student Committee is pleased to present our next spotlighted researcher in the "Meet the PEER Students'' Series. The series features students and postdoctoral researchers who conduct exciting research projects, engage in leadership activities, and perform exceptional work. This month, we met Dr. Sifat Muin, a postdoctoral researcher at PEER.
Dr. Sifat Muin is a postdoctoral researcher at the Pacific Earthquake Engineering Research (PEER) Center. She finished her PhD in Civil and Environmental Engineering at UC Berkeley from Prof. Khalid Mosalam's group in 2018. She earned her BS and MS in Civil Engineering from Bangladesh University of Engineering and Technology and UC Berkeley, respectively. Her research interest lies in earthquake engineering with an emphasis on structural health monitoring, machine learning, and response-based damage detection. She published over 25 research publications in the most reputed platforms in the field. Dr. Muin is an active member of Structural Extreme Events Reconnaissance (StEER) network and Earthquake Engineering Research Institute (EERI).
What made you interested in earthquake engineering?
My interest in earthquake engineering grew gradually. The first time I was exposed to earthquake engineering was back when I was doing my undergrad. In my senior year we had to do a thesis, and the topic for me was pushover analysis. Then I came to the U.S., where I got more involved in different kinds of earthquake engineering projects. For example, during my masters, I studied the response of long-period structures to far-field earthquakes. Then I did an internship at PEER where I was working with the ground motion modeling team. As I got more exposed to earthquake engineering projects, eventually that kind of hooked me on to earthquake engineering.
What are your research objectives?
My main research objective is to enhance post-disaster resiliency. Resiliency means getting back to the pre-event state. One of the main aspects to get to resiliency is to assess the condition of structures, whether they're safe to resume operation or not. The current way of assessing this is that inspectors personally go and check and tag structures. This process takes time. My research objective is to make this process automated, so it doesn't take as long as usual. What I'm trying to do is to get data from sensors installed on structures and develop methods using machine learning tools that will monitor and assess the condition of the structure right after the earthquake. The overarching goal here is to develop a holistic monitoring and assessment system that, in real time, will help make decisions for the disaster-stricken community. I also feel very strongly that my research should impact the community, so translating my research to practice is something else I am really striving for. That’s another research value or research objective.
What are your career goals?
I want to stay in academia and keep doing research. It’s hard, so I’ll keep working and see if it’s something I can achieve. If not in academia, then I’ll try to have some research position at a national laboratory or in industry. Research is something I want to keep doing. This is my goal, and whether it's in the Academy as a faculty or whether it's as a research scientist somewhere else, that's something the future will tell.
How do you think the use of artificial intelligence will impact structural engineering practices in the long term?
I see a tremendous amount of potential for artificial intelligence in structural engineering. What it will bring to the field is the rapidness. Structural engineering is a mature field with tried and tested methods, but everything takes quite a bit of time. I think this is where artificial intelligence is essential. Let's take the example of regional simulation, where developing individual models for each structure is a non-trivial task. It's also not feasible. However, we can develop machine learning algorithms that can look at images of the structures and give information about certain components, which you can use in your model. You can also use transfer learning to get information from the existing models and then incorporate image-based machine learning tool outputs and develop the models. This will make the whole process a lot faster than what it is currently. The other example is data-driven structural health monitoring, where I'm working: acceleration data can tell you how the structures are doing. The implications are huge, so I think we should keep trying to bring more artificial intelligence into structural engineering, for our own sake.
What activities in your research have been the most challenging?
I’d say that it was writing, because compared to other aspects of research such as coding and data analysis, writing was less exciting, and it took a lot of effort even to get started. I’m still learning, but I also realized that it is something essential, because, if you're doing research but no one knows about it, it loses its impact. It is very important to let people know what you're doing. In my experiences in writing, the hard part was learning how to write specifically for research. There is a learning curve to it, you have to know the right tools to use to help you write, the formats and style matter, and you have to develop a habit for writing. Getting the first publication is the hardest, but you learn to go through the process of submission, handling reviews, and resubmission, which improves your manuscript. Then, as long as you build a habit to keep on writing, you will publish more and more.