There are various ways in which the large number of parameters contained in statistical machine translation models may be acquired. While early work emphasized the use of generative statistical models and methods such as the expectation maximization (EM) algorithm, the use of alternative machine learning paradigms such as discriminative training has been explored.
Machine Learning and its 9 sub-topics are the main subject of 550 publications.
Topics in MachineLearningWord Graphs | Parameter Tuning | Reranking | Large Scale Discriminative Training | Minimum Bayes Risk | Confidence Measures | System Combination | Incremental Updating | Neural Network Models
Last modified on March 06, 2015, at 06:37 PM
Hosted by the University of Edinburgh
Powered by PmWiki | Skin design theundersigned | mod by CarlosAB