Syntactic Models of Alignment for Machine Translation - 2005
Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only! Grab it Explore a comprehensive lecture on syntactic models for aligning parallel text in machine translation, delivered by Dan Gildea from the University of Rochester in 2005 at the Center for Language & Speech Processing, Johns Hopkins University. Delve into tree-based alignment models that utilize syntactic information for one or both languages, as well as models that infer structure directly from bilingual text. Examine theoretical aspects of Synchronous Context Free Grammars as a translation model, including a method to factor grammars for reducing synchronous parsing complexity. Learn about various concepts such as inversion transduction grammar, string and tree alignment, grouping, complexity, cloning, and lexicalized alignment. Gain insights into hybrid machine translation approaches, alignment errors, dependencies, and headword choice. This 1-hour 15-minute talk provides a deep dive into advanced topics in computational linguistics and machine translation.