Skip to Main Content
Table 6 

Evaluation results for different reranking configurations. Beam = paraphrase selection beam (% of first best probability); Mean n = mean number of paraphrases per act; Total n = total number of paraphrases used for evaluation; Nat = mean naturalness rating over the generated paraphrase set. The last 3 columns indicate the significance of the difference in naturalness according to a two-tailed Wilcoxon rank sum test (*p < 0.05, **p < 0.01, ***p < 0.001).

graphic
 
graphic
 
Close Modal

or Create an Account

Close Modal
Close Modal