Experimental Results | ‘able 2: Segmented Model Scores. |
Experimental Results | We find that when using a ood segmentation model , segmentation of the lorphologically complex target language im-roves model performance over an unsegmented aseline (the confidence scores come from boot-3rap resampling). |
Experimental Results | So, we ran the word-based baseline system, the segmented model (Unsup L—match), and the prediction model (CRF—LM) outputs, along with the reference translation through the supervised morphological analyzer Omorfi (Piri—nen and Listenmaa, 2007). |
Models 2.1 Baseline Models | These are derived from the same unsupervised segmentation model used in other experiments. |
Models 2.1 Baseline Models | performance of unsupervised segmentation for translation, our third baseline is a segmented translation model based on a supervised segmentation model (called Sup), using the hand-built Omorfi morphological analyzer (Pirinen and Lis-tenmaa, 2007), which provided slightly higher BLEU scores than the word-based baseline. |
Models 2.1 Baseline Models | We therefore trained several different segmentation models , considering factors of granularity, coverage, and source-target symmetry. |
Related Work | The goal of this experiment was to control the segmented model’s tendency to overfit by rewarding it for using correct whole-word forms. |
Background | We also build in parallel a segmentation model to select 3,- from the set {new}, same}. |
Background | We build two segmentation models , one trained on contributions of less than four tokens, and another trained on contributions of four or more tokens, to distinguish between characteristics of contentful and non-contentful contributions. |
Background | No segmentation model is used and no ILP constraints are enforced. |