Parsing is a key task in natural language processing. It involves predicting, for each natural language sentence, an abstract representation of the grammatical entities in the sentence and the relations between these entities.
Abstract We present novel metrics for parse evaluation in joint segmentation and parsing scenarios where the gold sequence of terminals is not known in advance. The protocol uses distance-based metrics defined for the space of trees over... more
Abstract We present novel metrics for parse evaluation in joint segmentation and parsing scenarios where the gold sequence of terminals is not known in advance. The protocol uses distance-based metrics defined for the space of trees over lattices. Our metrics allow us to precisely quantify the performance gap between non-realistic parsing scenarios (assuming gold segmented and tagged input) and realistic ones (not assuming gold segmentation and tags).
Abstract A serious bottleneck of comparative parser evaluation is the fact that different parsers subscribe to different formal frameworks and theoretical assumptions. Converting outputs from one framework to another is less than optimal... more
Abstract A serious bottleneck of comparative parser evaluation is the fact that different parsers subscribe to different formal frameworks and theoretical assumptions. Converting outputs from one framework to another is less than optimal as it easily introduces noise into the process. Here we present a principled formal protocol for evaluating parsing results across frameworks based on multi-function trees, tree generalization and edit distance metrics.
Abstract Current parameters of accurate unlexicalized parsers based on Probabilistic Context-Free Grammars (PCFGs) form a twodimensional grid in which rewrite events are conditioned on both horizontal (headoutward) and vertical (parental)... more
Abstract Current parameters of accurate unlexicalized parsers based on Probabilistic Context-Free Grammars (PCFGs) form a twodimensional grid in which rewrite events are conditioned on both horizontal (headoutward) and vertical (parental) histories. In Semitic languages, where arguments may move around rather freely and phrasestructures are often shallow, there are additional morphological factors that govern the generation process.
Abstract Methods for evaluating dependency parsing using attachment scores are highly sensitive to representational variation between dependency treebanks, making cross-experimental evaluation opaque. This paper develops a robust... more
Abstract Methods for evaluating dependency parsing using attachment scores are highly sensitive to representational variation between dependency treebanks, making cross-experimental evaluation opaque. This paper develops a robust procedure for cross-experimental evaluation, based on deterministic unification-based operations for harmonizing different representations and a refined notion of tree edit distance for evaluating parse hypotheses relative to multiple gold standards.
This work aims to demonstrate that event structure and thematic relations are closely intertwined. Specifically, we show that in Modern Hebrew the choice of a morphological template has profound effects on the event structure of derived... more
This work aims to demonstrate that event structure and thematic relations are closely intertwined. Specifically, we show that in Modern Hebrew the choice of a morphological template has profound effects on the event structure of derived verbs. These effects are correlated with the thematic features marked by the templates, and are mediated by the aspectual classification of the lexical material provided by roots.
Abstract The grammar of Semitic languages is different from that of English and many other languages. Therefore, general-purpose statistical parsers are not always equally successful when applied to Semitic data. This chapter considers... more
Abstract The grammar of Semitic languages is different from that of English and many other languages. Therefore, general-purpose statistical parsers are not always equally successful when applied to Semitic data. This chapter considers the syntax of Semitic languages and how it challenges existing general-purpose parsing architectures. We then survey the different components of a generative probabilistic parsing system and show how they can be designed and implemented in order to effectively cope with Semitic data.
Abstract This paper presents a novel architecture for specifying rich morphosyntactic representations and learning the associated grammars from annotated data. The key idea underlying the architecture is the application of the traditional... more
Abstract This paper presents a novel architecture for specifying rich morphosyntactic representations and learning the associated grammars from annotated data. The key idea underlying the architecture is the application of the traditional notion of a “paradigm” to the syntactic domain. N-place predicates associated with paradigm cells are viewed as relational networks that are realized recursively by combining and ordering cells from other paradigms.
The purpose of this paper is to present Marvin Minsky's view of Artificial intelligence, based on three of his papers. This paper is organized as follows. In section 1 I provide a brief introduction to AI and to Marvin Minsky's work.