Maritime Information Markup

1
Maritime Information Markup
Arizona State University U.S. Coast Guard
NSF Digital Government dg.o2001
2
Overview

• Introduction
• Maritime information
• Ontological engineering
• Markup language
• Digital charts and other information
• Prototype demonstration
• Benefits and Summary

3
Maritime Information

• Maritime information is of various types
  database, diagrams, pictorial, text, structured,
  unstructured.
• Charts pictorial depictions based on vector (or
  raster) data, with overlays of features' (e. g.,
  navigation aids, obstructions).
• Text documents describing what the charts cannot
  depict, using a structured narrative format.
• Port information and port status reports.
• Weather reports and forecasts, warnings, text
  messages, updates about navigation aids, etc.
  This could be either structured or free form text.

4
Computational Ontology

• In AI, a computational ontology is a
  "representation of a conceptualization". It is a
  collection of (defined) concepts that exist in a
  domain and the relationships between them.
  "Computational" means that it can be processed by
  software.
• Why do we need it? To allow software to
  'understand' the relationships between things,
  for example, that a 'silo' is (for boaters) a
  'landmark.
• The knowledge we need comes from standards
  documents digital chart databases lexicons and
  symbology definitions and other canonical'
  documents.
• Concepts and relationships are identified by (1)
  semi-automated scraping from standards documents
  and web sites (2)database relations entry by
  hand from lexicons and symbology definitions.

5
Computational Ontology

• Results c. 170 classes from standards, c. 100
  from sample digital chart c. 500 from lexicons
  and symbology, 30 more from markup.
• This is a draft product the number of classes
  will change (fall?).
• Reuse of other ontologies was explored early on,
  but it turned out that not much could be re-used,
  largely because the word senses were wrong' for
  our domain. For example, a bridge is a passageway
  for cars...but a potential obstruction for boats.

6
Taxonomy
7
Markup Language

• Maritime Information Markup Language (MIML) is a
  markup language for documents containing maritime
  information.
• Why do we need it? The documents are densely
  packed with information they are
  semi-structured linguistic methods are unlikely
  to be useful because of the lack of linguistic
  clues rarely does anyone need all the
  information in a part of a document.
• Markup is currently being done by hand.
• How is it being used? Currently, for retrieval.

8
(No Transcript)
9
Digital Chart Data

• What is it?
• Answer Information about individual features.
• What is being done with it?
• Answer Currently, retrieval on demand, soon,
  inferences based on values, and distribution.
• Other information 'Coast Pilot', weather
  forecasts, etc.

10
Putting it all together

• Standards chart database lexicons symbology
  definitions sample documents ---gt ontology
• Ontology (class and attribute names) ---gt MIML
  tags
• MIML data retrieval, C P, weather information,
  etc. ---gt smarter applications, better
  information processing

11
Prototype

• The prototype information retrieval demonstration
  that accepts questions and provides answers in
  the form of information retrieved from different
  sources. The prototype demonstrates the use of
  the taxonomy to provide a unified interface for
  retrieval of all types of information.
• It retrieves information from (1) the Coast
  Pilot (2) a feature database generated from
  digital nautical charts (3) web sites. It also
  generates some information (tide predictions) on
  an as-needed basis.
• Demonstration this evening.

12
Plans for the Future

• The most important item for research is to feed
  the results forward' into other applications.
• The most important item for the applications side
  is to use the draft ontology and markup in an
  effort to create a markup standard for this area.

13
Benefits

• The bulk of United States imports and exports
  travel by water. In addition, inland waterways
  carry a significant proportion of domestic
  freight.
• This project will help mariners get more
  complete and timely information vital to safety
  facilitate development of advanced navigation
  aids for coastal navigation software help with
  updating base material (e.g., chart databases,
  the Coast Pilot) from change notices (e.g., Local
  notices to Mariners).
• The computer science research aspects impinge
  mostly on artificial intelligence, the Semantic
  Web and related technology, and geospatial
  information.

14
Summary

• This talk has described a 'feed-through' project
  consisting of a transition from ontological
  engineering to information distribution
  technology via markup languages.

15
Culprits

• Raphael Malyankar
• Nicholas Findler
• Leo Leong
• Rashmi Iyengar
• Suresh Lalvani
• Jayagowri Renuka
• Helen Wu

• Jay Spalding, US Coast Guard
• Data supplied by NOAA Maptech
• NSF Digital Govt. program