Meet the recruited faculty – Farnaz Adib Yaghmaie

How can learning-based methods be used to control complex systems in challenging and unpredictable environments? This is something that Farnaz Adib Yaghmaie is exploring. We spoke with her about this and several initiatives, including the new research center SEDDIT.

Farnaz Adib Yaghmaie, Assistant Professor in Automatic Control at Linköping University, joined the ELLIIT Recruited Faculty in January 2022 with a focus on reinforcement learning and control theory.

Please, tell us about your research areas.

I am particularly interested in the intersection of learning and control, this includes reinforcement learning, online control and learning methods for control. I enjoy defining and redefining learning problems to enhance control systems, and vice versa.

Where did you find your interest in these areas?

My background is in control-electrical engineering. While working on a research question during my PhD, I noticed that the problem could not be solved by classical control solutions and required a learning algorithm. I then continued working on learning solutions for control during my postdocs and became fascinated by the capabilities and mathematics behind those solutions.

In your own words, please summarize your work within these areas.

I am working on developing algorithms or designing agents that go beyond data and can reason about various situations. Such an agent continuously interacts with the environment, and through this interaction, it learns how to accomplish tasks. A good example could be a robot working alongside humans in an environment. Through interaction with the environment and other people, the robot learns to perform tasks successfully.

Do you have any active projects within ELLIIT?

When I joined ELLIIT as an assistant professor, I proposed a research topic that I am currently pursuing. The project focuses on controlling dynamical systems that are subject to adversarial attacks. For example, in the context of autonomous driving, an adversary might introduce small disturbances to the car, potentially compromising its performance. My research aims to develop strategies to control the car effectively in such scenarios.

What are your ambitions with ELLIIT in the future?

I would like to leverage the synergy within the ELLIIT environment to contribute towards reliable, safe, and efficient information and artificial intelligence technology. I believe that achieving this requires not only expertise from various fields but also an environment where people can share and brainstorm. ELLIIT provides such an environment, and I am grateful for the opportunity.

Is there anything else you’re working on that you’d like to highlight?

I am continuously developing and teaching PhD courses. I developed a 6-credit course from scratch on reinforcement learning (RL) and taught it in Fall 2024 at LiU. For details regarding lectures and sample codes, you can visit the course page. (link below)

I also teach a PhD course on RL for WASP, which is offered every second year. Currently, I am developing a PhD course on advanced robotics to be given in Fall 2025 or Spring 2026.

Additionally, I am involved in a new research center called SEnsor informatics and Decision-making for the DIgital Transformation (SEDDIT), where I work on foundation models and reinforcement learning. (link below)

And finally, what do you enjoy doing in your free time outside of academia?

In the summertime, I enjoy spending time in nature and exercising. However, in the winter, I prefer baking. In practice, I don’t have much time for either, as my free time is mostly occupied by my kids.