Laurent Lessard is the newest addition to the Optimization theme at WID. He was previously located at UC Berkeley where he worked on decentralized control, large-scale optimization, and computation-based management of complex engineered systems. Michael Ferris, the Optimization theme leader, says this about Lessard: “Laurent is an excellent researcher at the interface of optimization and control; well trained, innovative, with impressive technical skills. We believe he will facilitate further collaborations between our strong groups focussing on data analytics, signal processing and the theory and algorithms of optimization.” As Lessard prepared for his start at WID, we asked him about his Tools for Discovery.
What do you work on?
My background is in control theory, the study of feedback and its effect on dynamical systems. I will be joining the optimization group at WID as well as the ECE department starting in the 2015-16 academic year. My work is focused on developing mathematical theory, algorithms, and analysis techniques to better understand and control large and complex systems.
My favorite example of a complex system is the self-driving car, which will undoubtedly see its official debut in the near future. A self-driving car must constantly gather and aggregate data from cameras, radar, GPS, as well as other internal sensors. Then, steering/throttle/braking adjustments are made automatically and the process is repeated. Another emerging complex system is the smart power grid, which will simultaneously draw from variable power sources such as wind and solar and cater to a variable user demand while accounting for constraints such as transmission line capacity.
The challenge with complex systems is that they overlap with several traditional disciplines: from computer science to electrical engineering to mechanical engineering. While a mechanical engineer might see a self-driving car as a more sophisticated version of a car, a computer scientist might instead see it as a sophisticated piece of software that happens to be in a car. I believe that both viewpoints are valuable, and that the mechanical engineer and computer scientist can learn a lot from one another.
My hope in joining the optimization group at WID is that I can bring some breadth to the group with my controls knowledge while at the same time learning and collaborating with other area experts in the group. I think the interdisciplinary nature of WID is very well suited for what I do, and I look forward to the challenges ahead!
What are your tools for analysis?
My most trusted tool is pen and paper. As ideas crystallize, there is a gradual transition from written notes to a typeset manuscript. Whether the idea amounts to a new solution or approach for a particular class of problems or a new method of data analysis, the next step often involves some form of computer validation, modeling, or simulation. My go-to software package is MATLAB, but depending on the need, I have recently found myself using a variety of other languages, including Mathematica, Python, and Julia.
“I am always in search of a more concise explanation, a more elegant approach, or a figure that better illustrates an idea or concept.”
Tools for Writing?
I use LaTeX for writing, note-taking, and even presentations. The main reason for this choice is that my work is largely mathematical and LaTeX handles mathematical typesetting beautifully. LaTeX documents are also easy to share, which facilitates collaboration.
Effectively communicating research findings can be challenging, particularly when the audience has a diverse technical background. Therefore, I spend a considerable amount of time editing, revising, and seeking feedback from others on my work. I am always in search of a more concise explanation, a more elegant approach, or a figure that better illustrates an idea or concept.
Tools for Collaboration?
I find great value in building a network of peers with a wide range of research interests. For me, this happens by attending seminars, conferences, and workshops, as well as giving talks about my own work. I also maintain a personal academic website so interested fellow researchers can see what I’m up to and contact me easily.
Having a network gives me context and perspective in research, something that is often lost if one works in isolation. For the quick sharing of ideas, I use a variety of tools depending on the situation: Dropbox, Google Docs, Skype, Slack, and E-mail. That being said, I prefer face-to-face interactions if possible, because my collaborators are my colleagues and my friends. Research is a business to many, but it is also a human endeavor!
Your ultimate tool for discovery?
My ultimate tool for discovery is to keep an open mind. New inspiration often comes from unexpected places, so I have found it incredibly valuable to expose myself to as many new people and ideas as possible. WID’s collaborative model for research exemplifies my view of how research should be done and I can’t wait to get started!
— curated by Nolan Lendved