Brown CS News

“It’s a paradoxical process,” Adriana Schulz says of her work in computational design at the intersection of formal methods, programming languages (PL), and machine learning (ML). “If I’m designing a bridge, it can’t all be new ideas and inspiration, but there’s more to it than precision and constraints. How do you reconcile that? Formal methods and PL allow correctness by construction, and ML is a great way to enable ideation and exploration. I’m passionate about combining both things, the formalism and the creativity.”

Adriana is one of two recent hires in the multi-year CS With Impact campaign, our largest expansion to date. This fall, she joins Brown CS as associate professor within a year of receiving a Sloan Research Fellowship, an NSF CAREER Award, and the ACM SIGGRAPH Significant New Researcher Award, as well as being named to MIT’s 35 Under 35.

“But in the beginning,” she says, “I was in it for the math. I wanted to be a mathematician, and I grew up in Brazil, and I wanted a real job, so I went into engineering instead of pure math.”

After earning a Bachelor in Electronics Engineering at Federal University of Rio de Janeiro, Adriana was drawn to computer graphics while completing a Master’s degree in Mathematics at Instituto de Matemática Pura e Aplicada, seeing it as mathematical but offering a chance not just to prove things but to implement, with tangible and inspiring results. 

“Beautiful renderings were exciting to me,” she says. “After finishing my Master’s, I decided to do a PhD at MIT, and I found that fabrication was the next level up: not just pretty pictures but making and building. My doctoral work was on classical tools for design and fabrication, so ML came much later, but as a faculty member at University of Washington, I really started to branch out of my comfort zone and collaborate a lot, pushing different areas of the field. And what stuck was PL and ML.”

“PL made a lot of sense,” Adriana says, “because digital fabrication is programmable fabrication. One of my first papers I was really passionate about was a language for optimizing carpentry designs with compiler techniques. ML was more of a challenge: it was so exciting to see automatically generated images and video, but manufacturing requires a lot of precision, and interpretability is tricky. I really like their intersection, because they’re two domains that are quite antithetical in some ways.”

Academia wasn’t without its challenges for Adriana, but the introduction to her doctoral thesis ends on a note of triumph: “The main discovery of my PhD was this imperfect yet amazing world.” It’s a striking statement, so I ask for the benefit of future CS students: what are her best practices for moving through negativity into greater enchantment? 

“Part of maturity,” she answers, “is becoming capable of loving something despite its imperfections. I think there’s a big difference between being an advocate and being a leader. The world is full of advocates, everyone standing up to to scream for some cause, and that part isn’t hard. Something I learned in developing my passion for academia was that it’s easy to say that the system’s broken, but it’s much harder to lead a program, a department, a school, to have to juggle and keep moving things forward.” 

One of her attempts to move things forward and pay them forward is the ACM SIGGRAPH Community Group for Women in Computer Graphics Research (WiGRAPH), which Adriana co-founded and currently directs. It dates back to 2016, when she started hosting the Berthouzoz Women in Research Lunch at the SIGGRAPH conference.

“There’s a lot of support once you decide to go on the job market,” she says, “but a lot of women give up before that point. I’m really passionate about our Rising Stars program, which provides funding and mentoring two years prior to that decision. It’s personal to me, because my advisor took me to SIGGRAPH and pushed me – if not for his support, I wouldn’t be here.”  

Her current research, Adriana explains, is aimed at nothing less than the next manufacturing revolution. At the same time that manufacturing becomes more programmable, enabling fabrication of items in smaller and smaller batches, or even batches of one, we’re able to specify material properties at incredibly high resolution, creating objects that behave differently from anything in the past. 

“I’m very interested,” she says, “in the questions of computational design that will drive the revolution forward. Representations are the key things that drive computation, so how do you represent a design or a design space so you can truly explore it? Is it a program, a data structure, or a combination of abstractions?”

Adriana sees industry collaboration as essential: “They have problems to solve and I don’t want to sit in an ivory tower. As one example, life cycle assessment is a big metric when companies decide the environmental impact of a product, but it’s so hard to compute that it doesn’t even enter the design cycle and gets calculated only after the product is made. I’m interested in where computer scientists can come in and improve that.”  

“Once again,” she says, “it’s about the paradox of design, the contradiction of design. We talk a lot about design for manufacturing, but the idea of producing more and more is really wasteful. A lot of the algorithms I’ve built are about compromise, navigating a landscape of tradeoffs, and environmental impact is a really important way to do that. I’m also really excited about accessibility, about designing systems and tools to make them available to more people.”

Adriana is happy to return to New England because her husband is from the area and she’s eager to have her kids grow up around their cousins and grandparents. She’s excited to be joining Brown CS because of the collaboration. 

“Brown has the longest continually-running graphics group in the known universe,” Adriana says, “and as a researcher, that’s thrilling. Daniel Ritchie is the world leader in neurosymbolic reasoning tools for computer graphics, and there isn’t enough room to list everybody I want to start working with or resume working with: James Tompkin, Will Crichton, Shriram Krishnamurthi, not to mention the roboticists!” 

She’s also very excited to be working with Brown’s undergraduates: “They really get involved in research early, and I love to teach. I’ve been teaching a course on computational fabrication for a few years, and I like that it’s different from an algorithms course where you’re taught three sorting algorithms, and the third one is the fast one you should actually use. Which algorithm should you use in a design context? It depends. There isn’t one perfect answer when you’re building systems, and I really want undergrads to be inspired by that, working together on something we can all use.” 

And the eventual audience for those systems? As large as possible, Adriana says.

“When the pandemic hit,” she remembers, “we started a digital fabrication center at University of Washington that used laser cutters to 3D print face shields, prototype new gown designs, and more. Doctors would come to us with their needs, and we’d produce the items they wished they could make. We brought them into the conversation. My grandma won’t operate a 3D printer, but she can be part of a similar conversation.”

“Do you mind a Ratatouille reference?” Adriana laughs. “The critic in the movie says it’s not true that everyone can cook, but in the end, he tastes Remy’s food and understands: it’s not that anyone can be a great artist, but a great artist can come from anywhere. That’s what digital fabrication is doing and what I want to enable.”  

For more information, click the link that follows to contact Brown CS Communications Manager Jesse C. Polhemus.