Jake Beattie
[Extra Bits] [CV]
About Me
My name is Jacob (Jake) Beattie, and I am a Computer Science and Engineering PhD student at The Ohio State University. I have the wonderful pleasure of being advised by Dr. Tanya Berger-Wolf and contributing to the Imageomics Institute and AI and Biodiversity Change (ABC) Global Center.
I completed dual Bachelor’s in Computer Science and Mathematics at Iowa State University in 2023. Go Cyclones!
My research is focused on the intersection of interpretability methods and scientific applications such as ecology, biology, and more. Lately, I have been specifically focused on how to use interpretability methods to accelerate scientific discovery.
Updates
- 12/2025 - I will be attending NeurIPS in San Diego. Please reach out if you'd like to chat about interpretability, AI for science, or anything else!
- 11/2025 - Released a preprint in collaboration with Sam Stevens, Towards Open-Ended Visual Scientific Discover with Sparse Autoencoders.
- 06/2025 - I am pleased to announce that a project I led focused on improving clinical trial screening with RAG following my graduation from undergrad has been accepted into Machine Learning: Health! You can find it here
- 06/2025 - I will be attending CVPR from 06/11 - 06/15 in Nashville! Stop by CV4Animals to say hi and take a look at our poster
- 05/2025 - Habitat-Driven Vocal Variation in Hawai'ian Birds, a paper that I helped write, was accepted into the CV4Animals workshop at CVPR!
- 01/2025 - I travelled to Hawai'i as part of the Experiential Introduction to AI and Ecology course. Over two weeks (and with the massive help of NEON staff and UH Hilo's Pat Hart and Amanda Navine), we collected some pretty awesome acoustic data on Hawai'ian birds.
Research
My main research interests lie at the intersection of AI, Explainability, and applications in Ecology/Biology. Over the course of my PhD (and beyond), my goal is to find some fundamental observation about how ML models structure their representations, and leverage this observation to accelerate ecological research and scientific discovery. To this end, I have already begun to work with incredible ecologists and biologists through the Imageomics Institute and the ABC Global Center. On the horizon, I have some projects that begin to work towards these theoretical observations as well.
Recently, I have focused on discovering morphological traits using sparse autoencoders.
Papers
Papers that I am particularly proud of and that encapsulate my overall research contributions well (currently exhaustive)
Towards Open-Ended Visual Scientific Discover with Sparse Autoencoders
Samuel Stevens, Jacob Beattie, Tanya Berger-Wolf, Yu Su [Preprint]
Habitat-Driven Vocal Variation in Hawai’ian Birds
Namrata Banerji, Ekaterina Nepovinnykh, Jacob Beattie, Hikaru Keebler, Leonardo Teixeira Viotti, Amanda Navine, Patrick Hart, Mike Long, Shea Uehana, Eissas Ouk, Evan Donoso, Avery Dean, Ann Carey, Tanya Berger-Wolf, Kaiya L Provost [Paper]
ChatGPT augmented clinical trial screening
Jacob Beattie, Dylan Owens, Ann Marie Navar, Luiza Giuliani Schmitt, Kimberly Taing, Sarah Neufeld1, Daniel Yang,Christian Chukwuma1, Ahmed Gul, Dong Soo Lee, Neil Desai, Dominic Moon, Jing Wang, Steve Jiang, and Michael Dohopolski [Paper]
Extra Bits
In my free time, I enjoy baking. If you happen to live in Columbus, OH, and desperately want a homemade loaf of bread, I can make one for you (in exchange for a few bucks for a poor grad student).