From: stuha@ida.liu.se (Sture Hagglund)
Subject: CFP: Expert Judgement, Human Error, and Intelligent Systems
Date: 12 Dec 91 01:00:57 GMT
Organization: CIS Dept, Univ of Linkoping, Sweden


Crossposted from news.announce.conferences.

CALL FOR PAPERS
RESEARCH WORKSHOP ON
EXPERT JUDGMENT, HUMAN ERROR, AND INTELLIGENT SYSTEMS
VIENNA, AUSTRIA, AUGUST 3-4, 1992

NEED: Many have concluded that biased judgment and human error are the main
causes of most industrial catastrophes, transportation accidents, medical
misdiagnoses, forecasting failures, etc. Humans frequently either: (1) plan to
carry out tasks in an initially faulty fashion (judgment bias) or (2) commit
slips/lapses in carrying out correct actions (human error). Yet for many
reasons,  proficient human performers more often repair their biases/errors
before any adverse consequences take effect. Their "bias/error debugging" style
of reasoning leads to the most adaptive,  creative, and surest ways to avoid
more such accidents. This is the benefit of keeping humans in the problem
solving loop.

        The cost of this style of reasoning is that errors and bias tendencies,
and their consequences,  remain more often than one would like. To reduce these
costs, does not require eliminating human bias/error. Rather, an increasing
number of psychologists have proposed that it is better to tolerate
biases/errors and to help humans minimize their consequences. The open issue of
interest to this workshop is how to go about reducing the consequences of human
biases and errors. This concerns the explanation abilities of intelligent
systems.

GOAL: The goal is to convene a research workshop at the European Conference on
AI (ECAI'92) in Vienna, August 3 and 4, 1992 to bring individuals from the
psychology, J/DM, AI, and HCI communities together to discuss the topics based
upon their research. The program committee is particularly interested in the
evaluation of alternative, interdisciplinary approaches to reducing human error
and bias. Preference will be given to researchers who have built actual systems
who have run relevant experiments, and/or, who offer a provocative but
well-founded viewpoint.

RELEVANT TOPICS: Of particular interest are any stimulating position statements
that contribute to the goal. Some examples include, but aren't limited to, new
ways to overcome the many research problems confronting the following topic
areas:

BETTER PSYCHOLOGICAL MODELS OF ERRORS AND BIASES
> Better models of fallible expert reasoning and judgment
(e.g.,skill-rule-knowledge levels, recognition primed decisionmaking, bounded
rationality, confirmation bias) -- How can models be more predictive?
> More precise/formal  models of human error (slips, lapses, accidents, group
errors, phenologies, etc.)
> Reducing biases and errors in real world, safety critical, complex systems
> Making decision analysis methods more cognitively natural & domain knowledge
enriched (embedding decision models in knowledge bases, embedding rules in
behavioral decision models, etc.)

IMPROVED WAYS TO INTEGRATE AI AND HCI INTO ERROR/BIAS REPAIR TASKS
> Getting expert critiquing systems, expert advisors, tutors, etc. to "say the
right thing at the right time"
> Designing more naturalistic, knowledge based human-computer interactions
(e.g., managing trouble, handling misunderstanding, discussive vs. directive
explanations, persuasiveness, user modeling)
> Exploring multi-media and other usability enhancing factors (e.g. hypermedia,
"ecological interfaces", interaction devices, I/F evaluations). Recovering when
users fail to understand a correction.
> Strengths/weaknesses of deep text generation, plan formulation, user modeling
and intention sharing.
> Computer supported cooperative work, theories for augmenting collaboration,
cooperative explanations.

Those wishing to participate should send 5 copies of a position paper (5 to 20
pages) by April 26, 1992 to Barry G. Silverman, Institute for Artificial
Intelligence, George Washington University, 2021 K Street, NW, Suite 710,
Washington D.C. 20006 USA. FAX: 202 785 3382. Acceptance notification will be
sent by June 7, 1992.

Organizing Committee: Jens Rasmussen (honorary), Risoe National Labs
(J_Rasmussen@risoe.dk); Mike Donnell, Geo. Washington U.; Gerhard Fischer, U.of
Colorado (Gerhard@boulder.colorado.edu); John Fox, Imperial Cancer Research
Fund; Sture Hagglund, U. of Linkoping (stuha@ida.liu.se); Erik Hollnagel,
Computer Resources International (Erik.Hollnagel@spd.cri.dk); Barry G.Silverman
(Barry@gwusun.gwu.edu); Masoud Yazdani, Exeter U. (Masoud@dcs.exeter.ac.uk).
ECAI'92 Contact: Werner Horn - TEL: 43-1-53532810 (Werner@ai-vie-uucp).

Sponsored by ECAI'92 in cooperation with AAAI.