srivbane

This group works on large-scale machine learning and data analysis with application mainly in climate science. The datasets for these problems are substantial, requiring a large number of computations as they search through piles of explicit and implicit information. Further, the methods used will again take advantage of algorithms that can be distributed to multiple nodes and GPUs. The researchers are working on several projects during 2022:

• Deep Learning Methods for Sub-Seasonal Forecasting
• High-Dimensional Geometry of Deep Neural Networks
• Differentially Private Adaptive Gradient Descent
• De-randomized PAC-Bayes Generalization Bound for Deep Nets