BodySLAM (Simultaneous Localization and Mapping)
01/2020 - Present
BodySLAM is an ongoing project with the Stanford IMMERS Lab to use mixed reality to support physicians during medical procedures. We are constructing a Microsoft HoloLens application which aligns a medical image with a patients body in 3D space, to support the superimposition of medical imaging data onto the patient during medical procedures.
Take a look at our working draft project writeup. (Co-authors: Kent Vainio, Christoph Leuze, Bruce Daniel).
Hacking for Defense (H4D) - Protecting Civilian Sites
03/2020 - 06/2020
Worked with the Air Force Research Laboratory to architect a speech-to-text platform for streamlining pilot-JTAC communications during close air support operations. Project sponsored by the DoD’s Joint Artificial Intelligence Center. Presented final report to Former Secretary of Defense Jim Mattis.
Take a look at our project report slides. (Collaborators: Xinlan Emily Hu, Emily Bishko, Matt Trost, Jason Ginsburg).
Generative Photocathode Materials Sourcing
09/2019 - 12/2019
Materials with a low photocathode emittance are of interest to particle physics researchers, as they enable the generation of high brightness electron beams while maintaining the ability to control beam parameters. In this project, we implement several variations of GAN to learn a distribution for the molecular structure of a material given a desired photoemittance. We also present a novel GAN architecture based off genetic policy optimization. Named GPO-GAN (Genetic Policy Optimization GAN), this model outperforms other standard GANs in this task.
Take a look at our project writeup. (Co-authors: Jason Zheng, Makena Low).