Our work references Phytomechatronics, a speculative framework for imagining technological advancements that nourish rather than extract from biological systems.
Below you’ll find short videos made by our undergraduate researchers. To learn about our most recent exhibition check out The Generative Tree.
Name: Grace Garavan
Year: Senior
Majors/minor: Anthropology and political science major, global studies minor
Research overview: “Motivated by the intersection between technology and the environment, I’ve spent the last year exploring the use of AI in identifying environmental trends, with an emphasis on climate change phenomena. Assisting Jenn Karson, founder of the Plant Machine Design Group at UVM, I photographed an original dataset of over 10,000 leaves that were eaten by an invasive species, Lymantria dispar (spongy moth caterpillar), during the outbreaks of 2021 and 2022. This Damaged Leaf Dataset was used to train a Generative Adversarial Network (GAN) to visualize patterns of damage and simulate the regeneration of this iconic Vermont foliage. Each leaf was collected, cleaned, pressed, then photographed and cataloged through this process. The AI algorithm can speculate on what the leaf looked like before it was eaten by the spongy moth caterpillar, or what a whole leaf would look like if the caterpillars ate it. We ‘ve presented this research to the public through gallery exhibitions and an interactive interface demonstrating how the AI algorithm processed each leaf and ‘healed’ the damage through layers of data.”
What has excited her: “I’ve been really excited about the idea of using technology to repair, rather than damage, our relationship with nature. By focusing on an invasive species in New England affected by the regional impacts of climate change, we hoped to showcase how AI can ‘solve’ pressing environmental issues while also being mindful of the environmental toll that such computing requires. The framework of phytomechatronics, which fascinates me, is about imagining technological advancements that nourish rather than extract from biological systems—a difficult balance to strike.”
What she’s gained: “Through my involvement in this project, I have gained both an immense appreciation for the patience required in experimental research and a love for arts-based methodologies. Photographing over 10,000 leaves took months, but I believe there was value in being able to see each piece of data individually, and conveying this data to the public through both artistic and scientific representations allowed people to interact and connect with it more. I plan to attend a master’s program in anthropology after I graduate, and I believe my appreciation for both the research process and the creativity in communicating its result will serve me well.”
Picking up where Maya Griffith, Liv Welford, and Cam Wodarz left off, Callie Levitt is fine tuning our machine learning model, a customized version of CycleGAN