Julia Gong, a junior at Stanford University, was a 2019 winner of the Adobe Research Women-in-Technology (WIT) Scholarship, which annually recognizes outstanding female undergraduate and master’s students studying computer science and related fields. WIT Scholars have the chance to interview for an internship at Adobe Research, and in summer 2019 Gong worked in our San Jose headquarters, exploring generative methods for creating art. We checked in with her to learn more about her experience.
Why did you apply to become an Adobe Research Women-in-Technology Scholar and intern?
I love the idea of teaching computers how to “see,” how to create, and how to aid people. I’ve gotten more and more interested in computer vision over the past few years because the work is not only visually satisfying, but also beneficial and powerful when used ethically. It can be applied to so many meaningful causes, whether medical, environmental, or artistic.
I applied to be a WIT Scholar because I wanted to further explore my interest in this area and have the guidance of mentors who know the field and have walked this path. Having recently become interested in generative models and the notion of computers creating art, I was ecstatic that I’d have the opportunity to intern at Adobe Research, the leading industry lab for this sort of cutting-edge, creative work that empowers artists rather than replacing them.
This internship at Adobe Research, along with the fantastic mentors and people I’ve met during this experience, has affirmed to me that I love computer science research for the knowledge, the impact I can create, and the people I work with. I’d love to pursue a PhD in computer vision after graduation, and my time at Adobe has given me the confidence that I have the potential to do good and make a difference in this field.
Could you describe your intern project? What motivated you to pursue this project?
I worked on developing a deep generative model for geometry-aware “caricature” generation—yes, as in the cartoon-like faces! I’m incredibly lucky to have wonderful mentors in Yannick Hold-Geoffroy and Cynthia Lu. Just before the internship, I’d become especially interested in generative models, such as GANs, that generate creative visual effects, particularly with applications to human faces, which I find so unique and interesting to work with. I’d just completed a project on facial sketch-to-image generation, so we began our brainstorming in related territory.
When conducting further brainstorming, we discovered that the current state-of-the-art for automated caricature generation didn’t properly disentangle geometry from style, and the geometric warping could also see great improvement. This formed the motivation for our project.
I am actually continuing my internship part-time into the fall in hopes of polishing our work and submitting it to one of the upcoming computer vision or graphics conferences, which is incredibly exciting to me. I can’t wait to see where this collaboration leads us!
Do you have any tips to share about how interns can have a good experience at Adobe Research?
First, don’t be afraid to pick a research topic that is daunting, unexplored, or uncharted territory! Research is about walking in the dark with a flashlight and finding the right ways to widen the beam.
Even as an undergrad, empowered by my amazing mentors, I was able to tackle and make headway on a project that certainly takes more than three months to make serious progress on. Research is a marathon, and Adobe makes the marathon a little less sweaty and a little more fun.
Second, Adobe Research has an astonishing density of people who are genuinely crazy-smart, yet willing to help anyone, an undergraduate intern included. Don’t be afraid to reach out to random researchers whose work you find interesting, or to experts other than your mentors who can help you with finding resources, tools, and papers, or just offer a different perspective on your project.
What’s special about interning at Adobe Research?
Adobe Research has been a remarkably vibrant and unique place at which to work. As an industry lab, it balances scientific inquiry with a product focus. There’s such a wide spectrum of work. While some projects are done with a product in mind, driven by customer feedback, and work directly with product teams, other projects are more on the side of exploration and basic science.
As someone who loves pursuing meaningful research that pushes the boundaries of science, but equally loves working with its applications and seeing real users interact with the end product, it’s encouraging to know that there’s a place in industry that balances both ends of the equation.
Adobe is a large software company that is simultaneously whimsical, creative, productive, impactful, motivated, innovative, stimulating, and fun. I don’t know if there’s any other place quite like it!