Research

Our lab’s research applies techniques from machine learning, signal processing, and optimization to solve inverse problems in computational imaging, focusing on the following areas. See our papers for the full picture.

Volumetric grid

3D reconstruction and representation

Grids vs INRs · Plenoxels · K-Planes · ThermalNeRF · Neural Microfacet Fields · 3D Field of Junctions · GA-Planes · Compression Error Bounds
Grids vs INRs project Grids vs INRs code Plenoxels code K-Planes code ThermalNeRF code Neural Microfacet Fields code 3D Field of Junctions project 3D Field of Junctions code GA-Planes code Compression Error Bounds code

Brain scan and Radon transform

Medical imaging

PaDIS-MRI · Provable TR · Probable, Fast TR · Propeller MRI · Perfusion Photon-Counting CT · DAB-quant
PaDIS-MRI code Probable, Fast TR code Propeller MRI code Perfusion Photon-Counting CT code DAB-quant code

Gradient ascent and MLP

Machine learning foundations

Spectral Bias · Sparse NN Recovery · Grids vs INRs · Compression Error Bounds
Spectral Bias code Sparse NN Recovery code Grids vs INRs project Grids vs INRs code Compression Error Bounds code

Diffusion model run on a brain

Generative priors for inverse problems

PaDIS-MRI · Distribution-Shift Uncertainty · Protein-Space Inverse Problems · Failures in DPS · KLIP
PaDIS-MRI code Distribution-Shift Uncertainty project Distribution-Shift Uncertainty code Protein-Space Inverse Problems code Failures in DPS code KLIP code

Classifier mislabeling under distribution shift

Identifying and tackling distribution shift

Fourier Lens on Robustness · Mistakes on ImageNet · Distribution-Shift Uncertainty · KLIP
Fourier Lens on Robustness code Mistakes on ImageNet code Distribution-Shift Uncertainty project Distribution-Shift Uncertainty code KLIP code