|
|
The increasing need to provide help to people in
making sense of the vast amount of information available to us
is widely recognised. Intelligent documents - documents which
provide
intelligent support for readers and providers of
information - can enable users to navigate effectively through
vast quantities of information. The presentation of information
can be customised in the
context of the requirements of the reader or the
information provider. The use of knowledge-based techniques in
the production of documentation can help to ensure consistency
and to improve quality.
Work in intelligent documents within the AI community at Edinburgh is concentrated in the Artificial Intelligence Applications Institute (AIAI) and the Natural Language Processing (NLP) group in the Department of AI. Close links are maintained with other groups within the Department of AI, including the Software Systems and Processes Group, and with researchers in the Centre for Cognitive Science and Human Communications Research Centre, particularly the Language Technology Group. There are over 50 researchers within the Edinburgh AI/Cognitive Science community working in areas relevant to intelligent documentation.
Formation, AIAI's knowledge-based layout system,
was commissioned originally by Pindar for use in production of
British Telecom's Yellow Pages directories. Designed to be extensible
to other layout
requirements, Formation is based on an interpreter
for the Layout Style Specification Language (LSSL), a domain-specific
language developed at AIAI for expressing layout style requirements.
The emphasis is on carrying out correctly the layout procedures
which have been specified, rather then using constraint-based
techniques to search for an optimal solution. The Formation GUI
offers dialog-based
modification and reconfiguration of styles.
Formation was delivered to Pindar ready for full
production use for BT Yellow Pages in late 1996. Planned future
development includes the creation of styles for laying out different
types of document,
e.g. catalogues, newspapers, brochures, and the development
of new features which enable Formation to create single-page documents
such as advertisements, based on the user's requirements for content
and look and feel.
Further Information: formation
AIAI worked with Aldus Europe for two years from
1990 to 1992 on the Paradigm project. Paradigm aimed to develop
functionality for the Pagemaker product which would assist users
in creating well-designed PageMaker templates for business documents
by interacting with an expert system driven by a set of document
design rules. The functionality envisaged was vaguely similar
to that provided by the
autocontent wizard found on today's Microsoft PowerPoint
product. AIAI was involved in performing the initial feasibility
study for the concept and then in providing consultancy to the
main project.
Highly technical documents, such as regulations,
can be difficult to navigate. For example, it can be very difficult
for someone unfamiliar with the Building Standards Regulations
for Scotland, having found the relevant sections for a particular
building project, to use them in a correct, logical and efficient
sequence. Designed originally to support the work of the Building
Directorate at the Scottish Office through making the regulations
and their associated Technical Standards more navigable, PLINTH
supports authors and readers of technical documents.
PLINTH allows authors to construct semantically augmented
hypertext networks for documents and their associated design rationales,
and to write intelligent navigation and consultation rules. It
assists the
reader to display different structural, logical and
rhetorical views of a document; to explore the design rationale
behind a document to understand its purpose; and to browse and
consult documents aided by
the intelligent navigation rules supplied by the
authors. In this way, PLINTH demonstrates that semantic tagging
can be used to provide different perspectives on the same information,
and to control the order in which it is presented to the user.
Further Information: PLINTH
The NLP group and AIAI supported the Military Aircraft
Division of British Aerospace Defence Limited on a research and
development project called GhostWriter. Its primary objective
was to describe, develop and demonstrate a prototypical authoring
environment which would illustrate the kinds of proactive support
required by authors of technical publications. From a business
perspective, GhostWriter had to demonstrate ways in which the
production of technical documents could be made more efficient
and reliable.
The cornerstone of the GhostWriter environment was
the provision of meaningful assistance for an author to affect
the composition of a knowledge-based model of a plan of actions
in a language-neutral form, such that this could be used as the
common basis for generating instructional texts in different languages.
Using this approach, British Aerospace could avoid the necessity
of translating a text from
one language into another.
The prototype has been used to show the interactive
and semi-automatic production of a significant portion of a complex
maintenance procedure to perform a functional test of the main
hydraulic power supply system of the Falcon 900 aircraft.
Further Information: Ghostwriter
In the IDAS project, the NLP group investigated how
natural language generation and hypertext techniques can be used
in a novel way to produce technical documentation. In IDAS, text
is derived from a single non-textual documentation database, rather
than text being itself the primary representation used. This
makes documentation adaptable and reusable, avoiding some of the
problems
with traditional technical documentation. The initial
IDAS prototypes were for users of Automatic Test Equipment. Work
in this area, in collaboration with the University of Sussex,
Racal Research and Racal Instruments, has continued with the CORECT
project.
CORECT is a system to help in the design of an electronic
system in response to a call for tenders from a customer. The
goal of CORECT is to produce tools to assist the marketing and
definition of complex
systems by allowing guided but free-format entry
of requirement facts, collating and checking consistency, tracking
correlation of implementation proposals and aiding the communication
between parties co-operating in this process. The role of the
University of Edinburgh is mainly to implement the generation
of documents for the different types of people involved in the
design process. An interesting
theoretical issue that arises here is whether (and
how) it is possible to allow users to specify their own document
requirements (in terms of form and content).
Racal will use the CORECT tool to support all aspects
of their business, beginning with the marketing side. Requirements
capture is a generic problem, and the CORECT approach will be
generalisable to
other application domains. Racal in collaboration
with the other project partners will undertake this generalisation
and technology transfer to other Racal companies, to ensure that
the CORECT tool will
be usable across the whole range of complex electronic
products & systems.
Further Information: CORECT
The ILEX project, an EPSRC-funded project involving
the Natural Language Processing Group within the Department of
AI and the Language Technology Group at the Human Communication
Research Centre, seeks to generate labels for items in an electronic
catalogue in such a way as to reflect the interest of the user,
while opportunistically furthering the aims of the information
provider (e.g. education). The project collaborators include the
National Museums of Scotland, and ILEX development is currently
aimed at providing an electronic tour round the gallery of 20th
Century Jewellery at the National Museum of Scotland.
ILEX focuses on automatic text generation. The ILEX
team want to consider the viewer's (or reader's) level of expertise,
to take into account the discourse history, to enable the use
of comparisons
etc. with objects seen previously and to allow the
viewer freedom in choosing the order in which objects are seen.
The system is to have its own agenda of educational goals, which
it must realise by talking about the objects in the tour; its
behaviour will be similar to the opportunistic education strategies
used by a teacher or tour guide.
Further Information: ILEX
There is increasing demand for personalised and customised
services. Together with the increasing variety of printing technologies
and the increased availability of new publishing media, this has
resulted in widespread interest within the printing and publishing
industries in tools and techniques which support their production
processes.
AIAI is interested in documents that can actively
support the "reader" or "user" in carrying
out their task. The aim is to allow documents to be created from
underlying models with tailoring to the needs to the user/reader
in support of the task they are carrying out. The GhostWriter
project made a good start with this type of support. AIAI is interested
in the type of systems that can create text, procedures, checklists,
images and graphics, and present them for viewing in an appropriate
way for the user taking into account their skill level, access
to information technology and working environment. This type
of system would support the work of the user/reader in carrying
out the task they have in hand. Presenting appropriate material
very specific to the task.
AIAI led the DTI flagship project Enterprise, which
aimed at providing a method and computer toolset to help in capturing
and analysing aspects of a business, using these to identify and
compare options for
meeting business requirements, and to support their
business processes.
Underlying the development of the Enterprise Tool
Set is the Enterprise Ontology, a technology of terms and definitions
relevant to business enterprises. The interface between the user
and the
Enterprise Tool Set is the Task Manager. It supports
the user in performing tasks through providing help with following
procedures. The Task Manager plans user tasks, and identifies
appropriate agents
through matching the capabilities required for tasks
against the capabilities of available agents; it thus provides
the user with support within the context of the resources currently
available.
AIAI is continuing its applied research in the area
under the EPSRC Business Process Change Programme.
Further Information: Enterprise
| Index | About AIAI | People | Projects | Technologies | Business Services | Publications | Info. Services |