Why use natural language (NL) to communicate with computer systems? I have always thought the answer was obvious. Since NL is the most natural communication language for people to use [1], it would make computer systems easiest to use if they understood and spoke NL. I often get resistance to this opinion, but the resistance almost always takes the form of comparing current GUIs to current NL i...

Morphological analysis and disambiguation are crucial stages in a variety of natural language processing applications, especially when languages with complex morphology are concerned. We present a system which disambiguates the output of a morphological analyzer for Hebrew. It consists of several simple classifiers and a module that combines them under the constraints imposed by the analyzer. W...

conception of linguistic information is characterised by the rise of ‘meaning proper’ in the wake of the development of modern logic, mainly through the work of Frege, Russell, and early Wittgenstein. One of the hallmarks of Frege’s philosophy of logic is his anti-psychologism: in order to give logic its proper due, he claims, we need to separate it from ‘psychology’, i.e., we need to distingui...

Our basic goal is the development of more robust systems for extracting information from natural language text. A robust system is one which is able to extract at least partial information despite the presence of ill-formed or unexpected syntactic, semantic, or discourse structures. Our approach has two aspects: First, we incorporate a rich set of syntactic, semantic, and discourse constraints,...

The meaning of a wide variety of natural language expressions can be stated in terms of degrees on scales. This paper provides an overview of the contribution of scales to linguistic meaning. I discuss the ontology of scales, survey the role of scalarity within and across languages, and summarize recent research into the linguistically relevant features of scales, with the goal of developing a ...

Graphs are a powerful representation formalism that can be applied to a variety of aspects related to language processing. We provide an overview of how Natural Language Processing problems have been projected into the graph framework, focusing in particular on graph construction – a crucial step in modeling the data to emphasize the phenomena targeted.

A recurrent network was trained from sentence examples to construct symbolic parses of sentence forms. Hundreds of sentences, representing significant syntactic complexity, were formulated and then divided into training and testing sets to evaluate the ability of a recurrent network to learn their structure. The network is shown to generalize well over test sentences and the errors that do rema...

Designing a database has been a tough problem since the early days of business data processing. As domains of database applications extend to textual and multimedia information processing, the design problem has grown even harder although the recent object-oriented approach introduced a richer set of building blocks into the database arena[2]. Natural language (NL, for short) databases are part...

We consider a dependency-parsed text corpus as an instance of a labeled directed graph, where nodes represent words and weighted directed edges represent the syntactic relations between them. We show that graph walks, combined with existing techniques of supervised learning that model local and global information about the graph walk process, can be used to derive a task-specific word similarit...

In the past decades of NLP, there has been a steady shift away from rule-based, linguistically motivated modeling towards statistical learning and the induction of unsupervised feature representations. However, natural language components used in today’s NLP pipelines are still static in the sense that their statistical model or rule-base is created once, then subsequently applied without furth...