121 search results - page 6 / 25 » Improving branch-and-cut performance by random sampling |
ICASSP
2011
IEEE
14 years 2 months ago
2011
IEEE
We have studied two efficient sampling methods, Langevin and Hessian adapted Metropolis Hastings (MH), applied to a parameter estimation problem of the mathematical model (Lorent...
81
Voted
CVPR
2004
IEEE
16 years 15 days ago
2004
IEEE
Invariant features or operators are often used to shield the recognition process from the effect of "nuisance" parameters, such as rotations, foreshortening, or illumina...
AI
2004
Springer
14 years 10 months ago
2004
Springer
Feature selection, as a preprocessing step to machine learning, has been very effective in reducing dimensionality, removing irrelevant data, increasing learning accuracy, and imp...
FLAIRS
2004
14 years 12 months ago
2004
Novelty detection in time series is an important problem with application in different domains such as machine failure detection, fraud detection and auditing. An approach to this...
105
click to vote
EMNLP
2010
14 years 8 months ago
2010
Minimum Error Rate Training is the algorithm for log-linear model parameter training most used in state-of-the-art Statistical Machine Translation systems. In its original formula...