Research Study Group

RESEARCH STUDY GROUP - IST-13/TG-002

VISUALIZING MASSIVE MILITARY DATASETS

REVISED TERMS OF REFERENCE - NOV 27, 1997

http://vistg.net

I. ORIGIN

The NATO Panel on Informatics Science and Technology (IST)

has adopted the Defence Research Group’s forum previously designated RSG.30, Panel 8. The new research study group is designated IST-13/TG-002, Visualization in Massive Military Datasets.

Background

Many people believe visualization to be the ability of computer systems to process data and display it pictorially. However, visualization is not the computer’s capability to generate pictures and graphs from data, but rather the human’s capacity to utilize effectively the output from the computer to understand the data.

The model below shows conceptually how the various building blocks of the visualization process relate to each other.

IST-13 interprets visualization as a human activity supported by technology. It is a means by which humans make sense of complex data. We consider visualization technologies, such as search engines, algorithmic processes and display devices and techniques, but only in relation to how they help humans to perform their tasks effectively. We emphasize the human use of computational systems to ensure that the right information is available in the form and at the time needed.

B. Scope

The aim of IST-13 is to explore methods for presenting to human users the implications of the contents of large, complex and varying military-relevant datasets of diverse kinds.

Significant issues include:

- What should be shown relates to the task requirements.

- What should be shown relates to the user's requirements.

- What can be shown is affected by time and resource constraints, also relating to the task and situation.

- What can be shown relates to engine capabilities.

- Can the system show what it should show?

C. Military Benefits

Military visualization problems arise constantly from the ever increasing volume, complexity and dynamic nature of data sets of interest that are beyond the capability of current presentation techniques to display effectively for human understanding and action. Effective visualization for the military will require the continuous development of new techniques - tools and methods - that will allow closer coupling and system level integration of the human’s capabilities into the overall human-computer system.

Anticipated military benefits are:

Greatly improved information accessibility, filtration, extraction, and understanding.

Significant improvements of: Simulation, Decision Support and Analysis systems.

Support for joint NATO activities through better interoperation and communication, particularly in PFP activities.

Promoting appropriate use of advanced visualization in operational applications and systems (Outreach).

II OBJECTIVES

Understanding and producing information from huge amounts of rapidly changing data is a key concern in military and defence operations. The IST Panel considers that flexible and intuitive visual interfaces will contribute greatly to military operations by improving human interaction with the dataflood and their capacity to extract and manage information from it.

IST-13 will conduct a work program including the following tasks:

a) Conduct critically needed short-term research into data visualization for military purposes.

This will include data visualization for Electronic Warfare ambiguity analysis, and means to cluster text reports for access and attention.

b) Begin necessary long-term study of automated knowledge discovery from major datasets for military purposes.

This activity includes text and speech corpora and can be expanded to include satellite imagery, battlefield data fusion, and visualizing geomatics data.

c) Maintain ongoing technology overview and review. This includes tracking visualization developments in Universities and the private sector, and forecasting what military needs will and will not be met in those sectors;

d) Bridge the "cultural" gap between users and technologists in the field of visualization; and

e) Support activity within the NATO countries by preparing, maintaining, and distributing online a database of ongoing research and development.

f) Support activity within the NATO countries by preparing, maintaining, and distributing online a listing of pointers to massive open source datasets suitable for testing and research into visualization and visualization technologies.

g) Support and foster the Network of Experts in Visualization for Massive Military-relevant Datasets.

III. RESOURCES

Membership

IST-13 members will be scientists with technical background in processing visual information and familiarity with defence applications.

Invited participation of experts from Industry is recommended, subject to the unanimous approval of all participating nations.

The status of IST-13 will be OPEN, subject to subsequent task-sensitivities.

IST-13 operates a Network of Experts [N/X], titled Data Visualization in Massive Military Datasets. One hundred and two scientists, engineers and officials currently belong. They are evenly distributed among Government, University and Private Sectors.

Technical

IST-13’s homepage is http://vistg.net

The Network-of-Experts Website is http://visn-x.net

IST-13 members have briefed IST-13 and its predecessor concerning several systems under development in their home laboratories and institutes. These are potentially useable for joint research and development among the participants.

CA: A 3-D interactive E-W visualization interface has delivered in prototype form and an early production system to a Canadian Defence Establishment, and a second production interface is anticipated. An application of a 3-D animated display is also being developed to help in analyzing large quantities of text.

DE: DDRE is carrying out development involved in virtual reality interfaces for mapping, on commercial software platforms.

BE: Researchers at the Royal Military Academy are working to extract relevant information from photographic material. For example, they might try to detect a hidden tank by surveying surrounding detail, possibly using multi-sensor fusion.

UK: There are projects on Textual Information Exploitation, and on Event Stream Analysis, both employing visualization:

Textual information extraction. A demonstrator to rank reports for anaysis has been built. Other tasks involve summarization and topic spotting.

For Event Stream Analysis and Visualization, DERA built Plumber, an open architecture for data mining and visualization, with local data analysis widgets and interfaces to Excel, MatLab, and generic databases.

Further British work: Extracting information and searching for for shape similarity in images. Databases include SPOT, IRLS, DD5, and SAR; aircraft and ships are the targets. As well, DERA has a "Knowledge Visualization Centre" using supercomputers.

ND: TNO-FEL has installed a Battle Lab, demonstrating several simulation and AI applications for training and operational use.

TNO (ND) and DERA (UK) are cooperating on the Advanced Information Systems Demonstrator..

CA: There is an experiment at DCIEM in which subjects coordinate on a very simple Command and Control problem. The analyses may be related to the DRA event stream analysis project

IV. SECURITY LEVEL

Discussions will be up to the classification NATO-SECRET. The meetings will be open to industry participation from the NATO countries as a general rule. Specific discussions may be considered closed on a case by case basis.

V. LIAISON

IST-13 will maintain liaison with the following other NATO Science and Technology Fora:

Speech processing

AI in C2

VR for training

Deliverables (Original Programme of Work)

IST-13 deliverables are listed below:

Year minus-1 et prev. (i.e., already delivered.)

Network-of-Experts [N/X] creation, and first meeting.
Proceedings of first N/X meeting.
Second N/X meeting.
IST-13 Website www.mmtaylor.net/private/rsg30/index.html
Human Aspects, Applications, Technologies for Dataset visualization (HAT paper) Table of Contents.

Year 1. (Current year, ends June, 1998.)

HAT paper
Proceedings of second N/X meeting
Third N/X meeting
Proceedings of third N/X meeting
Expertise database
Dataset pointers
First [short term] research deliverables–text clustering prototype
Halfway Review Milestones and Progress Report of IST-13, detailing the above, and progress in other task 3 research.

Year 2.

Fourth N/X meeting
Proceedings of fourth N/X meeting
Long term research on information discovery continues.

Year 3.

Fifth N/X meeting
Proceedings of Fifth N/X meeting
Final Report for the current work period, detailing the results of all experimentation (Task #3); and other tasks

Tasks	Year Present
	1997	0/	1998	1999	2000
1. Database of expertise		**
2. Dataset pointers			**
3. Short term research: - Text clustering			**prototype delivery
4. Long term research: - Information discovery
5. N/X network of experts	**Meeting Proceedings each year
6. Websites	IST-13 Website N/X Website
7. IST-13 Reports	*HAT paper TOC delivery			*IST-13 interim report	*IST-13 final report

* Milestones

** Deliverables