Virginia Goodwin

Virginia Goodwin
I’ve been collaborating with my colleagues across the Lab on various facets of how to do responsible machine learning. I really want the Lab to be a leader in the domain.

Please tell us about your area of expertise at the Laboratory. How are you applying it toward R&D that addresses a national security challenge?

Let’s be real; when I started at the Laboratory with a bachelor’s degree in physics, I did not have an area of expertise. I was competent to do basic analysis and coding that I was told to do. Fortunately, the Lab gives you opportunities to grow, under the guidance of the more senior staff. My areas of expertise have grown through the course of my career to include

  1. machine learning for classification — which is to say, creating algorithms that will correctly categorize new data based on similarity to previously seen data.
  2. computer vision — a subset of machine learning algorithms that focus on extracting information from imagery data, typically seeking to replicate human visual performance.
  3. systems analysis — this is usually where we get a squishy, loosely articulated problem from a sponsor, and we need to wrest that into a well-posed problem, figure out what the major technical challenges are, and propose potential solutions. These are the really interesting ones, like, “We know X new technology is going to become available in the next year. How will that impact our defense systems?” These problems are where you end up learning whole new domains of information that you never dreamed of needing to know.

Can you describe one of the best projects you've worked on at the Lab?

My favorite project has been the Collaborative-UAS for Hostile Attribution, Surveillance, Emplacement, and Reconnaissance (CHASER) program. I love this one because it’s my brain child; it started as idea I had based on current events and current state of the art, we fleshed it out within the group and pitched it for internal funding, we got awarded some funding, and we went out and turned an idea into a real, concrete thing that now exists.

It started from an air security incident where a drone (a.k.a an uncrewed aerial system, or UAS) flew into the active runway airspace at Gatwick airport right during the Christmas travel season, forcing the airport to shut down for two days. I read about that in the news, and I thought, “We can solve this. The technology pieces are there, it just needs to be put together.”

So we did it. We built out the CHASER testbed to develop autonomous drone missions, with an initial focus on being able to autonomously follow another drone, in order to identify the intruding drone pilot, to resolve situations like the Gatwick incident. We’ve now had a number of successful flight tests exercising different autonomous functions. It’s also just really fun to go out in the field and fly drones. Particularly during the summer of 2021, when we weren’t seeing each other in the office due to Covid, going out to the field with the team was especially beneficial. It’s been extremely rewarding to me to see this whole program that came about because I had an idea; two years of an entire team of engineers getting paid to do interesting work because I read a news article and thought, “We can fix this.”

What keeps you at Lincoln Laboratory?

I’ve been at the Lab since 2004. What has kept me here is twofold. One: reasonable work-life balance. Of course there are crunch times, but most of the time I work a 40-hour week, and I can flex those hours as needed. The Lab’s flexible work schedule is a real boon. I want to be able to enjoy my life outside of work, and the Lab allows me to do that.

The second reason is the Lab’s scope of different subject areas that allows you to guide your career and to explore different technical interests as they arise. Most of my technical work is similar — machine learning, data analysis, or systems analysis. But the specifics can vary wildly; I’ve done projects ranging from assessing GPS errors for en route air traffic, analysis of the impact of 5G cellular on small drones, analysis of homeland defense against cruise missiles, plus a significant amount of algorithm development for both computer vision and radar data classification. I’ve been in five different groups in two different divisions, and in each one I’ve learned so much new and interesting stuff.

Right now I’m dipping my toes into the domain of ethics in machine learning, and one of my pet projects is moving the whole Lab toward better documentation for machine learning models and data sets. This is totally beside the actual work, but the Lab allows you the freedom to go after what you want to do. I recently co-authored a paper on the benefits of systems analysis for machine learning, and it wasn’t part of funded work, but a colleague and I had the idea, fleshed it out, and away we went. I enjoy learning new things all the time. I know way more than I ever thought I would need to about GPS error sources, how cellular technology works, Army logistics, laser pointing accuracy, and a whole host of random topics that have come up over the course of all my different projects. I am frequently frustrated, but rarely bored.

What impact do you want to have on the world?

My current pet project is bringing greater transparency to machine learning (so-called “AI” but I could give you my whole dissertation on why that’s a bad phrase and I try to avoid using it). I’ve been pushing my teams within the lab to adopt documentation best practices for the machine learning models that we build, and I’ve been collaborating with my colleagues across the Lab on various facets of how to do responsible machine learning. I really want the Lab to be a leader in the domain. I believe that it’s both our individual responsibility as engineers, and also our shared responsibility as a national laboratory, to be a leader in developing and implementing best practices and mitigating potential harm from these technologies.

How do you like to spend your time outside of the Laboratory?

Right now I’ve got two young children, so they take up the vast majority of my time outside of work. We also totally jumped on the Covid puppy bandwagon, and since our adopted puppy is a rescue she requires a little extra love and attention. My husband and I are also executing a multiphase effort over many years to renovate our home, which is a lot. Hilariously, I find that the program management skills I’ve honed at work are extremely helpful in the home renovation program execution. A typical weekend involves shuttling the children to their various activities, some baking or cooking projects (in the immortal words of Jack Black in School of Rock, “Because I like to eat”), and the never-ending yard work and house work. Truly, a gripping suburban drama.