... Automatic generation of weather forecast texts using comprehensive probabilistic generation- space models Natural Language Engineering, 14(4):431–455, 2008 J Bilmes and K Kirchhoff Factored language models ... sizes, using random sampling and active learning Differences for training set sizes of 20 and 40 are all significant (p < 05) terestingly, while models trained on 100 utterances outperform models ... training using active learning significantly improves generation performance on sparse datasets, yielding results close to the human gold standard using a fraction of the data Phrase-based generation...