It is standardly believed that some occurrences of expressions designate singularly, while other occurrences of expressions designate plurally. For instance, the singular expression the student may be used on an occasion to talk about one particular student, while the plural expression the students may be used on an occasion to talk about several…

Naturalistic language learning is contextually grounded. When people learn their first (L1) and often their second (L2) language, they do so in various contexts. In this dissertation I examine the effect of various contexts on language development. Part 1 describes the effects of textual, linguistic context in reading. I employed an eye-tracking…

While information theory is typically considered in the context of modern computing and engineering, its core mathematical principles provide a potentially useful lens through which to consider human language. Like the artificial communication systems such principles were invented to describe, natural languages involve a sender and receiver, a…

Natural language is highly complex and can be challenging for some learners, yet the contribution of complexity to individual differences in language learning remains poorly understood. This poor understanding appears due to both a lack of consensus among researchers regarding what complexity is, and to on-line language research often employing…

The evolution of the Web brought new interesting problems to computer scientists that we loosely classify in the fields of social and semantic computing. Social computing is related to two major paradigms: computations carried out by a large amount of people in a collective intelligence fashion (i.e. wikis), and performing computations on social…

Natural Language Generation (NLG) is the process of generating natural language text from an input, which is a communicative goal and a database or knowledge base. Informally, the architecture of a standard NLG system consists of the following modules (Reiter and Dale, 2000): content determination, sentence planning (or microplanning) and surface…

Domain adaptation of natural language processing systems is challenging because it requires human expertise. While manual effort is effective in creating a high quality knowledge base, it is expensive and time consuming. Clinical text adds another layer of complexity to the task due to privacy and confidentiality restrictions that hinder the…

This thesis aims to define and extend a line of computational models for text comprehension that are humanly plausible. Since natural language is human by nature, computational models of human language will always be just that--models. To the degree that they miss out on information that humans would tap into, they may be improved by considering…

Much recent work in natural language processing treats linguistic analysis as an inference problem over graphs. This development opens up useful connections between machine learning, graph theory, and linguistics. The first part of this dissertation formulates syntactic dependency parsing as a dynamic Markov random field with the novel…

The goal of this dissertation is to introduce a method for deriving morphemes from Arabic words using stem patterns, a feature of Arabic morphology. The motivations are three-fold: modeling with morphemes rather than words should help address the out-of-vocabulary problem; working with stem patterns should prove to be a cross-dialectally valid…

The deep semantic understanding necessary for complex natural language processing tasks, such as automatic question-answering or text summarization, would benefit from highly accurate word sense disambiguation (WSD). This dissertation investigates what makes an appropriate and effective sense inventory for WSD. Drawing on theories and…

For over 30 years, reference resolution, the process of determining what a noun phrase including a pronoun refers to in written and spoken language, has been an important and on-going area of research. Most existing pronominal reference resolution algorithms and systems are designed to use syntactic information and surface features (e.g. number…

Topic models like latent Dirichlet allocation (LDA) provide a framework for analyzing large datasets where observations are collected into groups. Although topic modeling has been fruitfully applied to problems social science, biology, and computer vision, it has been most widely used to model datasets where documents are modeled as exchangeable…

In this study I introduce BioFrameNet, an extension of the Berkeley FrameNet lexical database to the domain of molecular biology. I examine the syntactic and semantic combinatorial possibilities exhibited in the lexical items used in this domain in order to get a better understanding of the grammatical properties of the language used in scientific…

Radiology data, in the form of images and reports, is growing at a high rate due to the introduction of new imaging modalities, new uses of existing modalities, and the growing importance of objective image information in the diagnosis and treatment of patients. This increase has resulted in an enormous set of image data that is richly annotated…