SciFly
March 2005 - July 2005
SciFly is a customised flyer generation system, built on top of
the Myriad platform. Myriad is the Information Engineering Lab's
generic platform for creating contextualised information retrieval
and delivery systems. The heart of Myriad is the Virtual
Document Planner (VDP) planning engine.
SciFly
is a demonstration system for CSIRO’s Myriad framework for information
retrieval and delivery. SciFly’s first public outing was at CeBIT 2005 in Sydney, which was held on May 24-26.
SciFly delivers customised content based on user-selected interests. In
the current system, content is assembled into a dynamically generated document
and printed on-the-spot. Simultaneously a PDF document is emailed to the
user along with a plain text summary in the body of an email.
This means that users receive relevant information presented in contextually
relevant formats:
- Paper flyer for immediate perusal
- PDF version for retention of information and later reference
- Plain text summary for mobile device access
The current demonstration of SciFly captures user details such as name,
email address and affiliation from a bar coded tag (as supplied to conference
delegates).The user then navigates a touch-screen menu system to select
topics of interest. The topics in this case are based on the capabilities,
projects and application domains of the CSIRO ICT Centre.
Once all data is collected, SciFly assembles content tailored to the specific
set of interests expressed by the user, dynamically fits the content to
the space constraints of a double sided A4 page, and adds relevant contact
information, web links and higher level context.
Importantly, SciFly does not just assemble pre-configured content, but
dynamically adjusts the amount and detail of content based on the range
of topics selected.
Technical Details
Paper flyers are produced as PDF documents. The PDF documents
are generated from XHTML input. This is done in two steps. The first
step is to use an XSLT transformation to convert the XHTML into
an XSL:FO formatted document. Using the Apache FOP processing engine,
the XSL:FO document is then transformed into a PDF document.
Email content is generated from the same discourse structure as
the paper output; that is, the content itself is not replanned for
the different output devices. Instead, the global presentation reasoning
(the information assembly stage as we call it - see ALTA Workshop
paper on Information
Assembly for Adaptive Display for more information about the
process) is re-run over the same content in the same discourse structure
in the context of the different output device. This results in the
content presentation being adapted to suit the differing needs and
constraints of textual email.
Read More about SciFly
Myriad and the Virtual Document Planner
Myriad is our software platform for building
applications that make use of contextualised information retrieval and delivery.
The core component of Myriad is the Virtual Document Planner or
VDP, which is a typical AI planning engine that performs goal-decomposition planning.
The VDP is a generic engine for delivering tailored information.
It was originally designed to produce documents dynamically,
integrating data from multiple sources and customising the
content for a user. We are currently extending the VDP to
handle both textual and graphical data planning for controlling
a Graphical User Interface. The architecture of the VDP is
a typical Natural Language Generation (NLG) architecture where
the linguistic resources are separate from the planning engines.
The engines use discourse rules to select and organise the
content, ensuring that the content remains coherent. Design
rules then determine an appropriate way to realise the content
and structure of the presentation on the relevant output medium
(eg HTML web page, Palm Pilot, Mobile Phone screen, speech,
multimedia display).
Topical Information Gathering for Easy Reuse (TIGER)
In general, the focus of the research in TIGER is on Question
Answering. Generically, Question Answering systems can be described
as systems that, given an arbitrary question posed in natural language,
scan available data sources and retrieve answers to the question.
The system then presents the available answers to the user in a
useful presentation.
In TIGER, however, we don't aim to investigate general purpose
question answering, nor do we attempt to address problems in all
dimensions of question answering. Instead, we have framed our investigation
by concentrating on how Question Answering can assist expert users
performing specific tasks with defined information needs.
Additionally, within this framework, we focus our efforts on answer
formulation (constructing useful and coherent answers), rather than
on other aspects of question answering or information retrieval.
This means that while work in information retrieval, and search
is a necessary component of any question answering system, we will
attempt to use existing technologies and methods for these (eg.
possibly P@noptic and/or
Google for the search
component).
More information is available at the Tiger
Project Internal Website (CSIRO access only).
Past Projects
Data Fusion, Data Management and Situation Assessment
May 2002 - July 2005
The Data Fusion, Data
Management and Situation Assessment (DFDMSA) project builds
upon the VDP architecture. Here the VDP is used to manage
the display of relevant information for Air Combat Operators
in performing surveillance and combat tasks.
An overview of the DFDMSA project's aims was published in
CoResearch in April 2003:

A research relationship between CSIRO and Boeing is likely
to have a significant impact on Australia's future early
warning and defence capabilities. The Australian Defence
Force's Project Wedgetail is providing a world-leading airborne
early warning and control system (AWACS) for Australia.
Boeing has funded $1 million of CSIRO research aimed at
developing better information-management systems to ensure
strategic advantage in surveillance and combat situations.
Several research teams for CSIRO Mathematical and Information
Sciences [now the CSIRO ICT Centre] are collaborating
on the project that brings together specialists in siutation
awareness, natural language generation and data integration.
Air Combat Operators (ACOs) are the key ground staff who
communicate with aircraft crews and inform them of developments
in a rapidly changing operational environment.
This means being aware of the location, direction and
identity of potential threats and targets and knowing which
are most likely to require closer monitoring or action.
The air crew can be places in serious danger if the ACOs'
awareness of the situation declines because of information
overload. The project aims to develop technologies that
handle large amounts of data and provide full situational
awareness to the ACO and the flight crew.
Extract from CO:RESEARCH No. 395, April 2003
The Operators' environment is characterized by large volumes
of time-critical information, and the need to integrate data
from a number of different sources. This leads to both visual
clutter, and a high cognitive load for operators in the current
system. To overcome this overload of information, and to better
support operators in performing their required tasks, we introduce
an adaptive Graphical User Interface (controlled through a
scripting API) and the VDP as a task-specific information
planning engine.
Technology wise, the GUI is implemented in pure Java, and
makes use of BBN's excellent OpenMap
framework for the display and management of geo-spatial
information. Our scripting API is built around the BeanShell
java source interpreter, and allows the VDP to control both
what is presented on the graphical the display (the content),
and the way it is presented (both modality and interaction
properties).
You can read more about our approach and implementation
in our published papers.
This research was funded by Boeing
(US) as part of the Australian Defence Force Project
Wedgetail for Airborne Early Warning and Control (AEW&C).
The project was successfully completed in July 2005, and Boeing and CSIRO are exploring opportunities
to further develop the technology.
Adaptive Interactive Mobile Systems (AIMS)
February 2002 - June 2002
AIMS
was a Chief's Special Project which developed algorithms, architectures
and software for adaptive interactive mobile systems, to enable
faster, smarter and more robust customised information profiling
and delivery. Specifically, the FLOCK
software developed deals with the email overload problem, employing
intelligent agents cooperating in a Knowledge Factory
to automatically, flexibly and adaptively categorise email.
The FLOCK software attracted much interest, and was the subject
of a front page article
in The Australian newspaper.
|