COMPUTATIONAL LINGUISTICS
ERCIM News No.26 - July 1996

Human-Centered Information Systems


by Gary W. Strong, Yi-Tzuu Chien, John Hestenes and Steve M. Griffin
The U.S. National Science Foundation (NSF) and Defense Advanced Research Projects Agency (DARPA) have co-operated in two human-centred initiatives, Human Language Technology (HLT) and Human Language Resources (HLR). A new initiative is the multi-agency, multi-modal communication initiative 'SIMULATE: Speech, Text, Image, and MULtimedia Advanced Technology Effort' (MMC). These multi-modal, speech and language initiatives are laying key foundations for new multi-agency research and development initiatives in Human-Centered Information Systems (HCIS). This document summarises on-going research and opportunities in human language technology in terms of the HCIS goals and research and technology threads.

In the last several years there has been a growing national awareness that the ways in which humans interact with and relate to information systems are critical to achieving high levels of service and functionality from systems.

Improvements in computing and communications per se are only a part of what is needed to reach the potential of new information technologies. While there are predictions for computers with 1,000 to 100,000 times the speed of today's systems there are as yet few organised R&D programs focusing on how individuals, groups, and communities will interact with them in institutional settings to extract useful information for decision making and for general knowledge expansion. Speed and connectivity are not enough to make systems useful as information engines. Many questions about representation, interactivity, cognition, corpora, and agents remain to be answered both before and after systems are implemented.

Many of the functional failures of major operational systems in the military, corporate, and educational arenas have been tentatively recognised and couched in such vague terms as operator error. Such phrases hint at knowledge gaps regarding how humans, groups and organisations interact with computing and communication systems. Current system design methodologies and requirements definition still tend to focus on the architecture of software and hardware systems, discounting higher level system issues, ones that address individuals, groups, communities, and the nature of collaboration environments.

The possibility of achieving new capabilities through human-centred information systems and technologies is an opportunity for broad societal impact. It holds the promise of redefining national competitiveness and productivity, for improving the general workforce, and creating a more informed and educated citizenry as we enter more deeply into the information age.

Technical Content and R&D Agenda of HCIS

The HCIS goals are to augment human capabilities to manage and use information through advances in five primary research and technology areas. Research programs or projects may aim at advances in either a few technology threads or, perhaps, in all of the following, depending on the specific research targets: Some relationships between these research threads and the three most recent inter-agency human language initiatives are:

Representation: Interactivity: Cognition: Corpora: Agents: Relationship to CIC Strategic Focus Areas

The Human-Centered Information Systems research and development agenda is also directly related to the National Science and Technology Council's Council on Information and Communication (CIC) Strategic Focus Areas (http://www.whitehouse.gov/White_House/EOP/OSTP/NSTC/html/cic/cic-plan.html) that expressly involve individuals and groups of users: User-Centered Interfaces and Tools, Virtual Environments, and Human Resources and Education. It is also has important relationships with the other three Strategic Focus Areas: High Performance Systems Software Technology, High Performance Computing Research, and High Confidence Systems.

Without government investment, certain expectations of advanced systems will continue to be unmet (such as speech recognition for ordinary citizens) and reasonable performance of human-agent systems (as in distributed, knowledge-based surgery and medicine). If the federal government does not invest in human-centred information systems and applications over the next decade, then this research and development will not be done and future government missions and national leadership and international competitiveness will be jeopardised.

The material in this paper represents the views of the authors and does not reflect National Science Foundation policies. Short descriptions of the HLT and HLR projects can be found at:
http://www.cs.tufts.edu/~jacob/isgw/
http://www.cise.nsf.gov/cise/IRIS/ISPhome.html

Please contact:
Gary W. Strong - National Science Foundation
Tel: +1 703 306 1928
Fax: +1 703 306 0599
E-mail: gstrong@nsf.gov
or strong.chi@xerox.com
return to the contents page