Modern grammar induction systems often employ curriculum learning strategies that begin by training on a subset of all available input that is considered simpler than the full data. Traditionally, filtering has been at granularities of whole input units, e.g., discarding entire sentences with too many words or punctuation marks. We propose instead viewing interpunctuation fragments as atoms, in...

Statistical Parsing with Context-Free Filtering Grammar Michael Demko Master of Science Graduate Department of Computer Science University of Toronto 2007 Statistical parsers that simultaneously generate both phrase-structure and lexical dependency trees have been limited in two important ways: the detection of non-projective dependencies has not been integrated with other parsing decisions, or...

In this paper we present a model to transfor a grammatical formalism in another. The model is applicable only on restrictive conditions. However, it is fairly useful for many purposes: parsing evaluation, researching methods for truly combining different parsing outputs to reach better parsing performances, and building larger syntactically annotated corpora for data-driven approaches. The mode...

This paper describes the semantic format of the UAIC Ro-Dia Dependency Treebank, based on the previous classical syntactic annotation. The discussed format exploits all the semantic information annotated in the morphological level. The transformation of syntactic annotation into semantic one is made semi-automatically, using a tool called Treeops, which is a converter of an XML format to anothe...

The paper is a first attempt to fill a gap in the dependency literature, by providing a mathematical result on the complexity of recognition with a dependency grammar. The paper describes an improved Earley-type recognizer with a complexity O(IGl2n3). The improvement is due to a precompilation of the dependency rules into parse tables, that determine the conditions of applicability of two prima...

A sequence of nodes in a syntax tree, each the head of the next, is usually called a projection path, or simply a projection. Projections in this sense are fairly basic phrase-structural entities, though hardly of central importance in most accounts. There are a few exceptions. For example, varieties of projection that I called c-projection and s-projection were significant in my earlier work o...

The frequent occurrence of divergences|structural diier-ences between languages|presents a great challenge for statistical word-level alignment. In this paper, we introduce DUSTer, a method for systematically identifying common divergence types and transforming an English sentence structure to bear a closer resemblance to that of another language. Our ultimate goal is to enable more accurate al...

This paper proposes a description of German word order including phenomena considered as complex, such as scrambling, (partial) VP fronting and verbal pied piping. Our description relates a syntactic dependency structure directly to a topological hierarchy without resorting to movement or similar mechanisms.

In this paper, parsing with dependency grammar is modeled as a constraint satisfaction problem. A restricted kind of constraints is proposed, which is simple enough to be implemented efficiently, but which is also rich enough to express a wide variety of grammatical well-formedness conditions. We give a number of examples to demonstrate how different kinds of linguistic knowledge can be encoded...