Bernie Boscoe

Bernie Boscoe

Assistant Professor, Computer Science
Southern Oregon University

About

I am an Assistant Professor of computer science at Southern Oregon University in Ashland, Oregon. I collaborate with domain scientists investigating our world, building pipelines to do better science. Current projects include AquiLLM, a RAG-LLM to study tacit knowledge in research groups, funded by the Sloan Foundation and the NSF. AquiLLM case studies include astronomy, environmental science, and computational acoustics. Related work includes GreenCrossingAI with environmental scientist Dr. Karen Mager here at SOU , and the Astrophysics Data Lab at UCLA with astrophysicist Dr. Tuan Do.

Projects

Selected Publications

Photometric redshift predictions

Photometric Redshifts for Cosmology: Improving Accuracy and Uncertainty Estimates Using Bayesian Neural Networks

Evan Jones, Tuan Do, Bernie Boscoe, Jack Singal, Yujie Wan, Zooey Nguyen.
The Astrophysical Journal, vol. 964, no. 2, IOP, 2024. doi:10.3847/1538-4357/ad2070.
arXiv:2306.13179 [astro-ph.CO]

US-RSE conference presentation

Navigating the Integration of Machine Learning into Domain Research

Bernie Boscoe, Tuan Do.
US-RSE Conference, 2023
[Zenodo]

Astronomy data pipeline

Elements of effective machine learning datasets in astronomy

Bernie Boscoe, Tuan Do, EvSelectedes, Billy Li, Kevin Alfaro and Chris Ma.
NeurIPS Machine Learning for the Physical Sciences Workshop, 2022
arXiv:2211.14401 [astro-ph.IM]

Selected Talks & Posters

  • A Retrieval Augmented Generation Tool for Research Groups. Chandler Campbell, Bernie Boscoe, Tuan Do. Gateways 2024, Bozeman, Montana
  • Arrays from the sky: Astronomy and Data Science. Willamette University, Salem, Oregon, 2023
  • Scaling Up: Incorporating HPC into undergraduate CS education courses using Gateways. Gateways 2023, Pittsburgh, PA
  • To Have and Have Not: Addressing inequities for learners accessing computational science environments. International Conference on Computational Science, London, UK, 2022
  • Three workflow add-ons to improve machine learning reproducibility in astronomy. Astronomical Data Analysis Software and Systems (ADASS), Cape Town, South Africa, 2021
  • From a Legacy System to an Open Source Library: What can we learn about sustaining software tools over decades? SIAM CSE21

Contact

If you are an SOU student interested in my line of research, please contact me at boscoeb@sou.edu.

You can also find me on GitHub.