We present an approach for synthesizing transformations to enhance locality in imperfectly-nested loops. The key idea is to embed the iteration space of every statement in a loop ...
Many learning applications are characterized by high dimensions. Usually not all of these dimensions are relevant and some are redundant. There are two main approaches to reduce d...
Abstract--We present an explicit formula for B-spline convolution kernels; these are defined as the convolution of several B-splines of variable widths and degrees . We apply our r...
In this paper we present a novel approach to acoustic model training for non-audible murmur (NAM) recognition using normal speech data transformed into NAM data. NAM is extremely ...
Feature-space transforms such as feature-space maximum likelihood linear regression (FMLLR) are very effective speaker adaptation technique, especially on mismatched test data. In...
Jing Huang, Karthik Visweswariah, Peder A. Olsen, ...