We consider the problem of Bayesian learning on sensitive datasets and present two simple
but somewhat surprising results that connect Bayesian learning to “differential privacy”, a …

We study the problem of subsampling in differential privacy (DP), a question that is the
centerpiece behind many successful differentially private machine learning algorithms. …

Obesity is well‐known as the second factor for tumorigenesis after smoking and is bound up
with the malignant progression of several kinds of cancers, including esophageal cancer, …

Time series forecasting is an important problem across many domains, including predictions
of solar plant energy output, electricity consumption, and traffic jam situation. In this paper, …

Training large neural networks requires distributing learning across multiple workers, where
the cost of communicating gradients can be a significant bottleneck. signSGD alleviates this …

Faced with distribution shift between training and test set, we wish to detect and quantify the
shift, and to correct our classifiers without test set labels. Motivated by medical diagnosis, …

Short-term probabilistic forecasts of the trajectory of the COVID-19 pandemic in the United
States have served as a visible and important communication channel between the scientific …

The Gaussian mechanism is an essential building block used in multitude of differentially
private data analysis algorithms. In this paper we revisit the Gaussian mechanism and show …

We introduce a family of adaptive estimators on graphs, based on penalizing the l 1 norm of
discrete graph differences. This generalizes the idea of trend filtering (Kim et al., 2009; …

Sparse Subspace Clustering (SSC) and Low-Rank Representation (LRR) are both
considered as the state-of-the-art methods for {\em subspace clustering}. The two methods are …