University of Washington Department of Computer
Science and Engineering
June 1, 2004
“My colleagues and I believe that the combination
of Artificial Intelligence and Ubiquitous Computing holds great promise for
people with Alzheimer’s Disease, mental retardation, traumatic brain injury,
and similar cognitive handicaps.”
- Professor Henry Kautz, CSE
In order to respond to
social ideas and needs, CSE seeks high-impact projects that allow the faculty
to collaborate with interdisciplinary and community partners. These projects demonstrate to CSE students
the importance of tackling problems where society might benefit from technology
solutions.
Scope of Assisted
Cognition
The Assisted Cognition
project led by Professor Henry Kautz illustrates how CSE responds to social
needs with an interdisciplinary approach.
Some people are born with or develop cognitive disorders that limit
their ability to perform ordinary tasks of daily living, such as preparing
meals or using public transportation. And these problems are growing. For example, in the next 50 years, the
number of Alzheimer’s patients in the U.S. is projected to increase by
threefold, from 5 million today to 15 million in 2050.
In describing the
Assisted Cognition project, Kautz says,
“My
colleagues and I believe that the combination of Artificial Intelligence and
Ubiquitous Computing holds great promise for people with Alzheimer’s Disease,
mental retardation, traumatic brain injury, and similar cognitive handicaps. Researchers are developing sensors and
algorithms that can track users’ daily movements in their homes and
communities, recognize when they are performing common activities of daily
living, and detect unusual patterns of movement and activity that signal
distress and the need for help. Such
technology will help nursing homes, families, and employers provide security to
patients or employees who suffer from cognitive disabilities by means of
devices that provide warnings, prompts, or guidance when the user is lost or
having difficulty performing a task.”
Major funding for the
project comes from the National Institute for Disability and Rehabilitation
Research (NIDRR), in the form of a joint grant between CSE and the Department
of Rehabilitation Medicine. This grant,
entitled “ACCESS: Assisted Cognition in Community, Employment, and Support
Settings” enables the two departments to work together to develop a personal
guidance system that will help cognitively disabled individuals safely and
effectively use public transportation.
Dr. Kurt Johnson, the lead on the medical side, notes that problems in
using public transit are a major cause of isolation and unemployment among
people suffering from mild to moderate mental retardation. Such a personal guidance system will improve
the quality of life of clients without the great cost and inconvenience of
special point-to-point bus services.
Another focus area within
the Assisted Cognition project are systems that help people with cognitive
disabilities perform day to day activities within their own homes, so that they
would not need to move to a nursing home or other medical facility. “A particular motivation for this kind of
application is the need to help the baby-boomer generation ‘age in place’,”
Kautz says. “My own family’s experience
with Alzheimer’s Disease showed me that providing reminders and gentle
corrections throughout the day can enable a person with early to middle stage
AD perform most of the tasks they need for daily life – things such as
dressing, housework, and hobbies. Since
not every sufferer has a family member or other caregiver who can provide such
assistance 24 hours a day, I started to think about how an AI (Artificial
Intelligence) system could take over some of the burden.”
Integrating
disciplines within Computer Science
Kautz’s scientific
background is in an area of AI that focuses on methods for representing and
reasoning about commonsense knowledge.
When he began to think about creating computerized cognitive aids he
quickly realized that the task would require expertise in many different areas
of computer science. “The uniquely
collaborative culture of UW CSE made it possible to turn the vision of Assisted
Cognition into tangible research,” Kautz says.
Professor Gaetano Borriello, one of the founders of the field of
ubiquitous computing and an expert on small sensing devices, became a
co-investigator on the NIDDR grant.
Professor Dieter Fox, a robotics expert, realized that many of the
techniques used for robot navigation in unknown environments could be applied
to the problem of inferring the state of a person – that is, his or her
location and the kind of activity being performed – from direct sensor data.
“Such integrated research
is a two-way street,” Kautz says. “Not
only does the application benefit from the insights of multiple perspectives,
but the work can lead to fundamental advances in the individual disciplines as
well.” For example, the motivating
example of the Assisted Cognition personal guidance system led Fox, Kautz, and
graduate student Lin Liao to develop a novel approach to learning hierarchical
probabilistic models of human motion.
This work received the 2004 Best Paper award at the prestigious National
Conference of the American Association for Artificial Intelligence.
Immediate classroom
benefits
As with the UrbanSim
project, the use of ubiquitous sensing devices in the home or in public places
raises ethical questions that faculty can discuss with their students. As they help people, these devices also
could be viewed by some as instruments of surveillance. In a sense, the development of these devices
brings the issues of privacy into the forefront where they can be examined in
the classroom.
In addition to exposure
to important issues, students begin to understand that computer science
research is not just about hardware and software. We can imagine and create new ways in which computing positively
transforms our very lives.
Corporate partnerships
In the future, technology
research will increasingly involve partnerships with private firms outside the
University. For example, Intel is a
very important partner in the Assisted Cognition project with the UW. As Kautz points out,
Intel
is providing funding for the project through the Intel Research Council. Intel Seattle Research (IRC), located right
in the University district, is working very closely with us. Many of our students are interning there
over the summer, and we are jointly carrying out much of the technical
work. Moreover, Intel has helped
establish a national organization called the Center for Aging Services
Technology (CAST), which is dedicated to bringing together researchers and
practitioners in a wide variety of different disciplines and organizations.
Viewpoint of the team from Rehabilitation Medicine
Research Associate Mark
Harniss and Research Scientist Pat Brown from the Department of Rehabilitation
Medicine are leading the effort within ACCESS to determine how the devices
should interact with users and caregivers in order to be acceptable and
effective. In particular, Harniss and
Brown point out that most of the disabled people who are potential users of the
devices are their own guardians.
Given the importance of
developing devices that increase the confidence and independence of users,
Harniss and Brown are involved in a needs assessment and testing process that
involves potential users and caregivers.
For example, they are asking such questions as:
The
importance of engaging the community in the definition and directions of the
project led the ACCESS team to form an Advisory Board that includes parent
representatives and members of the Washington State Divisions of Developmental
Disabilities and Vocational Rehabilitation, the Washington State ARC (Advocates
for the Rights of Citizens with Developmental Disabilities), and Columbia Legal
Services.
Impact of project on
students
Even though the Assisted
Cognition Project and partnership between CSE and the Medical School are in the
early stages of development, they are already having impacts on students who
begin to envision their professional lives in light of work on the project.
For example, Don Patterson,
a graduate student who works on the Assisted Cognition Project, illustrates how
this project fits into his own philosophy of his professional responsibilities,
and how it is preparing him to become a professor of Computer Science.
In the United States
Navy, Patterson was able to use his life-long interest in computers to operate
database software that runs the complex systems on ships. However, the Navy experience was also
frustrating because there wasn’t the opportunity to improve the databases and
because the project didn’t fulfill his strongest ambition as a computer
scientist: “to improve the quality of life of people who are affected daily by
computer systems.”
In the spring of 2004,
Don worked with Lin Liao, fellow CSE graduate student Krzysztof Gajos, a group
of CSE undergraduates, and their advisors Kautz and Fox to create a working
prototype of a personal guidance system.
The prototype, dubbed “Opportunity Knocks” because it alerted the user
to possible danger by a knocking sound, was demonstrated to members of Congress
at a special session of the national conference of the American Association of
Homes and Services for the Aging (AAHSA) in Washington, DC. The user interface for Opportunity Knocks is
a cell-phone that is equipped with GPS (global positioning system) and a color
screen, as shown in Figure 1.


Figure
1: Opportunity Knocks is a prototype personal guidance system designed to help
people with cognitive disabilities use public transportation safely and
effectively. This figure shows a series
of screen shots of the cell-phone interface as the user (i) selects a desired
destination, (ii) selects the desired mode of transportation, (iii) is alerted
to the fact that he failed to get off the bus at the property stop (the message
is accompanied by a sound alert), and (iv) is given new directions on how to
get on back to his destination.
Summary of Assisted
Cognition
As David Notkin, Chair of
CSE, points out in the Introduction to this case study, CSE has “fluid
interactions” with partner communities.
In the case of the Assisted Cognition, students benefit directly when
they work in the project that has interacts with the Medical School and
industrial partners like Intel. They
see partnerships formed within and without UW, develop synergy, and produce
technology of great benefit to people.
In short: collaborative projects like Assisted Cognition help train a
generation of scientists dedicated to creating computing system that directly
improve the quality of life of people from all sectors of society.
FOR MORE INFORMATION
VISIT:
http://www.cs.washington.edu/assistcog/
PHOTOGRAPHS OF PEOPLE
FROM ASSISTED COGNITION PROJECT:

Henry Kautz

Dieter Fox

Gaetano Borriello

Kurt Johnson