Interactive Linear Models in Survey Sampling

Sampling Rate Conversion in the Discrete Linear Canonical Transform Domain

Sampling rate conversion (SRC) is one of important issues in modern sampling theory. It can be realized by up-sampling, filtering, and down-sampling operations, which need large complexity. Although some efficient algorithms have been presented to do the sampling rate conversion, they all need to compute the N-point original signal to obtain the up-sampling or the down-sampling signal in the tim...

Linear-Time Gibbs Sampling in Piecewise Graphical Models

Many real-world Bayesian inference problems such as preference learning or trader valuation modeling in financial markets naturally use piecewise likelihoods. Unfortunately, exact closed-form inference in the underlying Bayesian graphical models is intractable in the general case and existing approximation techniques provide few guarantees on both approximation quality and efficiency. While (Ma...

Sampling from the posterior distribution in generalized linear mixed models

Performance measure congruity in linear agency models with interactive tasks

This note demonstrates how performance measure congruity and noise determine an agency’s total surplus within an linear agency framework with multiple tasks. It provides a decomposition of agency costs, leading back to a congruity index previously proposed in the literature. In addition, it generalizes this index to a more general cost function, thereby highlighting the context specificity of t...

A Survey of Sparse Linear Models

We live in a data-intensive world and every day a lot of new data is being generated in domains like the internet, genetics, and multimedia, among others. This paper explores recent advances in sparse linear models, mainly in terms of algorithmic development and to a lesser extent exploring ways to apply them to large datasets. We discuss mathematical motivation behind these algorithms and also...