Corpora: job suitable for statistical computational linguist

[James Cussens]

Dear folks,

It may well be that you know someone (perhaps yourself) suitable for
the following post. A statistically inclined computational linguist
would fit the bill - something I did not stress sufficiently in the
advert.

James

James Cussens 				             jc at cs.york.ac.uk
Department of Computer Science 		      Tel  +44 (0)1904 434732
University of York			      Fax  +44 (0)1904 432767
Heslington, York YO10 5DD, UK		 http://www.cs.york.ac.uk/~jc


**********************************************************************

    POSTDOCTORAL RESEARCHER AT DEPARTMENT OF COMPUTER SCIENCE
			  UNIVERSITY OF YORK

		Induction of Stochastic Logic Programs

**********************************************************************

James Cussens has recently been awarded an EPSRC-funded 1-year
research grant starting 1st Oct 2000 to investigate methods for
inducing stochastic logic programs. A postdoctoral post is therefore
available within the Artificial Intelligence group.

This post requires a post-doctoral researcher with a background in
statistical analysis of complex models. Ideally this would be combined
with experience of logic programming. The post would suit someone
interested in extending existing probabilistic models such as Bayes
nets and probabilistic context-free grammars. Knowledge of statistical
computational linguistics would be an advantage as would some
familiarity with inductive logic programming.

The appointment is for a period of one year starting 1st October 2000.
Funds on the grant allow us to appoint at the maximum point on the
Grade IA scale (24,479 GBP per annum).

It is expected that James will be working with David Page (Departments
of Computer Science and Biostatistics and Medical Informatics) at the
University of Wisconsin for the latter half of the project. The
researcher appointed on this project should be able to join James on
this visit should they wish to do so.

Informal enquiries may be made to: James Cussens (jc at cs.york.ac.uk,
Tel: +44 1904 434732).

MORE INFORMATION ABOUT THE PROJECT CAN BE FOUND AT
http://www.cs.york.ac.uk/~jc/research/slps/

CLOSING DATE FOR APPLICATIONS WILL BE 12 JULY 2000.
Interviews are expected to be on 3rd August.

Formal applications can be made by sending THREE copies of a letter of
application and a full curriculum vitae, together with the names and
addresses of three referees, to the Personnel Office, University of
York, Heslington, York YO10 5DD, UK. Please quote reference number
6035. Email applications will NOT be accepted.

Overview
--------

Probabilistic models are currently an important focus of research in
both Artificial Intelligence and Computational Linguistics. In the AI
community, attention has focussed on Bayesian nets and related
graphical models. In computational linguistics n-gram models, hidden
Markov models and stochastic context-free grammars have been used
widely as part of the `statistical natural language processing
revolution.'

In both communities there is interest in extending current methods to
incorporate domain knowledge (e.g. linguistic knowledge) and/or
relational data. Inductive logic programming is an approach to machine
learning that does allow domain knowledge to be incorporated, but
where probabilistic methods are relatively undeveloped.

Stochastic logic programs (SLPs) are logic programs with added
parameters which define a log-linear distribution over proofs. It has
previously been argued that they effectively combine many of the
strengths of statistical and logical models. The project intends to
put this argument to the test by combining and developing work in both
computational linguistics and uncertainty in AI with the goal of
designing and implementing algorithms for learning both structure and
parameters of SLPs from data and background knowledge.


Department
----------
The Department of Computer Science has a record of high achievement in
research and teaching.  It was rated Grade 5* (this is the highest
possible rating i.e. attainable levels of international excellence in
a majority of sub-areas of activity and to attainable levels of
national excellence in all others) in the 1996 Research Assessment
Exercise, and Excellent (i.e. demonstrably very high levels of
achievement and best practice) in the 1994 Teaching Quality Assessment
Exercise.  The University won the 1996 Queen's Anniversary Prize for
Higher and Further Education for the work of the Department of
Computer Science, and a recent ranking of UK Computer Science
departments placed the department equal first in a ranking of 83
institutions.

The Department's continued expansion and new facilities are
accommodated in a purpose-built, 4200 square-metre building, opened in
September 1997, which occupies the highest point on campus.  The most
prominent of the Department's facilities is a 1.6 million pound
Silicon Graphics Origin-2300 supercomputer with 32 processors and 8
gigabytes of memory.  This provides one of the best resources for high
performance computing at a UK academic institution.

The Artificial Intelligence Group is one of the largest and most
productive in the UK.  The research of the Group is concerned with the
theoretical principles of artificial intelligence and their
application to real-world domains. Most of the Group's research focuses
on the areas of automated reasoning, machine learning, natural language
processing and intelligent user interfaces.