Modelling Human Syntax by Means of Probabilistic Dependency Grammars - 2007

Explore a comprehensive lecture on probabilistic dependency grammars and their role in computational linguistics. Delve into the limitations of context-free dependency grammars in accounting for linguistic phenomena such as non-projective word order, secondary dependencies, and punctuation. Examine a generative dependency model that addresses these challenges and learn about error-driven incremental parsing algorithms with repair. Discover how the dependency model assigns probabilities to partial dependency analyses and the need for incorporating time-dependence into the model for left-right incremental text processing. Gain insights from Matthias Buch-Kromann, head of the Computational Linguistics Group at Copenhagen Business School, as he shares his expertise in dependency treebanks, probabilistic dependency models, and computational models of human parsing and translation.