With a few exceptions, discriminative training in statistical machine translation (SMT) has been content with tuning weights for large feature sets on small development data. Evid...
One common predictive modeling challenge occurs in text mining problems is that the training data and the operational (testing) data are drawn from different underlying distributi...