Most data-driven dependency parsing approaches assume that sentence structure is represented as trees. Although trees have several desirable properties from both computational and...
Tree-based translation models, which exploit the linguistic syntax of source language, usually separate decoding into two steps: parsing and translation. Although this separation ...
We implement a Korean syntactic analyzer which decreases many ambiguities in syntax parse trees using segmentation and semantic connection units. We use dependency grammar for par...
Current parsing models are not immediately applicable for languages that exhibit strong interaction between morphology and syntax, e.g., Modern Hebrew (MH), Arabic and other Semit...
Conventional sentence compression methods employ a syntactic parser to compress a sentence without changing its meaning. However, the reference compressions made by humans do not ...