We propose to study links between three important classification algorithms: Perceptrons, Multi-Layer Perceptrons (MLPs) and Support Vector Machines (SVMs). We first study ways to...
Logistic Regression (LR) has been widely used in statistics for many years, and has received extensive study in machine learning community recently due to its close relations to S...
Jian Zhang, Rong Jin, Yiming Yang, Alexander G. Ha...
Efficient training of direct multi-class formulations of linear Support Vector Machines is very useful in applications such as text classification with a huge number examples as w...
S. Sathiya Keerthi, S. Sundararajan, Kai-Wei Chang...
Variational data assimilation consists in estimating control parameters of a numerical model in order to minimize the misfit between the forecast values and some actual observatio...
Luigi Nardi, Charles Sorror, Fouad Badran, Sylvie ...
Model compensation is a standard way of improving the robustness of speech recognition systems to noise. A number of popular schemes are based on vector Taylor series (vts) compen...