Home of live rats
and obviously completely out of date: last updates, Jan 23 1995
- and don't trust anyone over 30.
Good question. I have problems sometimes, trying to
decide how to describe myself,
or considering ways in which I would like to be described by others,
or whether this is even possible for various arcane
philosophical reasons.
(But, of course, others have had this
problem before. Bite me,
Mr. kripke
)
Alternatively, you could look at a set of bookmarks (only slightly out-of-date) handily compiled by Netscape (that's "MoZilla" to you).
Or of course, you can mail me and ask.
I am a gradual student. I graduate. that is my current telic role in the great lexicon of life. I have, in the past, both graduated, and not graduated. Never cylindrically.
Statistically, many of you who will get here know much about the graduate school shtick, probably more than you ever thought you wanted to. For those who don't, I do research. What's that mean? It means i re-search --- I look at things people have looked at before; and then I laugh. A haughty laugh, it spews from my lungs as if...
I'm getting off the subject. Flame off.
there's much reorganizing to be done around here.
I live here.
life in the fast lane
Ostensibly, I am a computer scientist. Computer science is a strange
field, in that it doesn't necessarily have a subject matter. (It has
been said by many that any subject with "science" in its name isn't
one.)
Among other things, computer science is the study of how to answer difficult questions using computers --- how to formulate the questions, what one needs to know to answer them, how do you interpret any answers you might get, etc. Well, that last one isn't often addressed as well as we might like, but we'll leave it on the list.
As a computer scientist, I study language -- people's language, how they use it, and how it might be that it can be used to talk about things. I don't do this from the perspective of the traditional linguist, but rather as corpus linguistics, using available computing tools and techniques (and inventing those not yet available) in order to structure available language data in such a way that interesting generalizations are visible.
I could also argue that I have an experiential learning perspective, and that I'm working on language acquisition. But that's another story [1].
These partial analysis sytems use syntact cues and knowledge of argument frames/selectional characteristics to derive interpretations for unknown terms in text, as well as to resolve known ambiguous terms by identifying relevant characteristics of their context.
This paper places these partial analysis systems in a theoretical framework by relating them to an argument-based functional (compositional) theory of semantics. The paper claims that a full phrase-structure grammar representation for NLP tasks is perhaps impractical, and outlines a pattern representation that characterizes local syntactic behavior.
The paper also presents simple corpus-based methods for acquiring these patterns. the patterns derived by these techniques are shown to be similar to those used in a working NLP information extraction system (Diderot).
The web (WWW, 3W, the spider of Gaul, whatever... ) is getting huge. Not that it's not large already, just that it's getting out of hand. To help you out in finding stuff that may or may not be out there, here are a few services that can guide you around:
T.S. Waterman ........................... waterman@cs.brandeis.edu
