In many prediction tasks, selecting relevant features is essential for achieving good generalization performance. Most feature selection algorithms consider all features to be a p...
Su-In Lee, Vassil Chatalbashev, David Vickrey, Dap...
Abstract: This paper presents a new statistical methodology to simulate the effect of both inter-die and intra-die variation on the electrical performance of analog integrated circ...
Carlo Guardiani, Sharad Saxena, Patrick McNamara, ...
Network coordinate (NC) system is an efficient and scalable system for Internet distance prediction. In this paper, we propose 3 two-layer NC systems HNPS, HBBS and HIDES derived f...
Chengbo Dong, Guodong Wang, Xuan Zhang, Beixing De...
— Both theory and a wealth of empirical studies have established that ensembles are more accurate than single predictive models. Unfortunately, the problem of how to maximize ens...
The Cross-Entropy (CE) method is a modern and effective optimization method well suited to parallel implementations. There is a vast array of problems today, some of which are hig...
Gareth E. Evans, Jonathan M. Keith, Dirk P. Kroese