POSSIBLE MTP PROJECTS 2007 /
8
Project A:
Title: Mapping between
Semantic Web pages
Abstract
The vision of the Semantic
Web, as set out in (Berners-Lee, Hendler, & Lassila 2001), describes a Web friendly to both humans and
machines; where natural language text conveys knowledge to humans, and
corresponding (instantiated) ontologies provide a form of easily
accessible knowledge (a loosely
structured KB) to machines. Further,
rules will also be available which specify how to map a concept from one
ontology, o1, to a concept in another ontology,
o2.
Assuming these mappings are
stored in a central mapping repository, the principal task of this project is
to produce an algorithm for inferring new correspondances
between ontologies which have not been explicitly stated by any mapping rule,
but which can be inferred by through the use of intermediary ontologies and
their associated mappings. The
ontologies themselves may also be stored in a repository.
For example, lets say a given
page is represented by ontology o1, which is associated with a set of
mappings (in the repository) to change the concepts in o1 to one of
several ontologies, lets call them ox and oy.
So effectively, o1 is associated with a set of mapppings
from o1 to ox and o1 to oy;
lets call these Mappings(1, x) and Mappings(o1, oy) respectively. If
we wish to map to an ontology, o2, representing another page’s content,
we would first check if the repository contains Mappings(o1,
o2) . If not, we could check if the repository has Mappings(ox,
o2) or Mappings(oy, o2), in which case we would know
we would be able to map from o1 to o2 in two stages. Of course, in general this mapping process will
explore a larger number of mapping stages.
The ability to determine correspondances like this is desirable for many
reasons. For example, when an organisation
is searching for the cheapest supplier of a certain product or service, a
program will be able to use these ontologies and mappings to compare the prices
of that product/service from multiple suppliers, without also returning prices
for similar but different products like current Web based price comparison
sites.
The
project would also have to produce / retrieve a number of small ontologies
& sets of mappings to support this activity.
Programming Language: Java
Supervisors: Derek Sleeman
& David Corsar
Project B:
A Web based Aid for Engineers, Scientists and their
Students: a Web-based system to perform Dimensional Analysis
The Physical Sciences
(including Engineering) use and manipulate a great number of measurements, and
there are clear rules about which of these entities can be added and subtracted. Just as it makes no sense to add “3 apples
and 2 jumbo-jets” it makes no sense to add “3 square centimetres” and “10
seconds.”
To avoid these unacceptable
computations a system of dimensional analysis has been developed where every
physical quantity is characterised by 5 dimensions, namely
Length (L), Time (T), Mass
(M), Temperature (K), and Electrical Current (I)
So in this notation a volume
of a cube is L*L*L or L^3
Speed = Distance/Time = L/T
or LąTą etc
The aim of this project is to
develop a system which given a set of definitions of physical entities can then
advise a user whether an expression like:
AREA ^ 2/TIME =
VOLUME/LENGTH/TIME is
valid.
Potential Users Students in the
Physical Sciences and Engineering still need to do these computations and would
find some kind of calculator useful.
Secondly researchers in an area of AI known as Scientific Discovery
frequently generate equations and these need to be checked to ensure they are
dimensionally valid/consistent.
Outline Project:
NB A more detailed note on Dimensional Analysis and an outline
design for the system is available on request.
Programming Language: Java (and various
packages)
Supervisors : Derek Sleeman and Edward Thomas