Hi! I am a Computer Scientist with the Machine Intelligence Group at Lawrence Livermore National Laboratory (LLNL), a Federally Funded Research Institute. My research interests broadly include computer vision, machine learning, and high dimensional data analysis.
My CV (pdf, updated Oct 2021).
I’m currently thinking about: (a) understanding and improving ML model behavior under controllable and uncontrollable distribution shifts (b) solving ill-posed inverse problems with novel data priors (c) applying ML to high energy physics, healthcare, and x-ray imaging, design optimization etc. I also serve on the program committees for several ML and vision venues (NeurIPS, CVPR, ICLR, ICCV, ICML, AAAI).
At LLNL, I lead the Open Data Initiative with the goal of releasing LLNL’s high impact scientific datasets for machine learning — so far, we have 10 datasets and counting!
We are looking to hire multiple postdocs and graduate interns to work on exciting machine learning research (both applied & fundamental). Reach out if you are interested
Updates
- (new!) Recent workshop papers at ICML’22:
- Improved Medical Out-of-Distribution Detectors For Modality and Semantic Shifts at Principles of Distribution Shift (PoDS)
Studies the effect of optimizing for inlier vs outlier consistencies for OOD detection - Data-Efficient Scientific Design Optimization with Neural Network Surrogates at Active Learning in the Real World
Proposes a new uncertainty estimator for accurate design optimization with very few training samples. - Accurate Calibration of Agent-based Epidemiological Models with Neural Network Surrogates at AI4COVID.
Shows a simple method to match observed data to complex epi models with NN surrogates - Machine Learning-Powered Mitigation Policy Optimization in Epidemiological Models at AI4COVID.
Studies complex policy optimization with different epi models using ML
- Improved Medical Out-of-Distribution Detectors For Modality and Semantic Shifts at Principles of Distribution Shift (PoDS)
- I was chosen as a Highlighted Reviewer for ICLR 2022
- Two papers accepted at ICASSP 2022 — (a) Sparsity to improve attribute discovery in StyleGANs and (b) Generalization error prediction using NN anchoring.
- Recent workshop papers at NeurIPS’21 — (a) Geometric priors for ICF at ML4PS Workshop (b) Characterizing distribution shift with multi-GAN alignment at DistShift Workshop.
Rushil Anirudh 