What do you get for $150000? - Computer Science

Download Report

Transcript What do you get for $150000? - Computer Science

IMPOSSIBLE
PROBLEMS
Are Good For Us
Chris Brown
Graduation 2004
Appealing First Impression!
The Witch’s Cottage
Seems like fun at the time…
Louie’s Bar and Grill
Looking Back, It Was Great.
The University of Rochester
Looking Back, It Was Great.
A Testimonial
--- K. Ross 2001
“Where other undergraduate programs
would avoid assigning `impossible' problems to
its undergraduates. Rochester uses them as tools
in its curriculum to teach students how to
approach a problem that is truly hard.
As a graduate of this department, I know how to reduce a problem to its most
elemental components... I know how to ask the right questions about a
problem...where precisely is the complexity of this problem? What assumptions
does my approach make and are these reasonable? Is this problem reducible to
something I can solve? Can I generalize this phenomenon?
Learning to ask these questions... learning how to approach overwhelmingly
difficult problems... these are all skills that I use constantly in my career, that I
find invaluable to my career, and that I developed solely as a result of the kind of
curriculum implemented in the Department of Computer Science.''
Impossible Computer Problems*
• Understand stories & word problems.
• Use natural language for questions and
answers to www.
• Surveillance video: detect food consumption in
computer lab (use sound too?)
• Tell person’s age from image.
• Build robot to answer questions, serve hors
d’oeuvres, read nametags, and locate disaster
victims. In Edmonton. And in Acapulco.
* © Randal Nelson
Mabel the Mobile Table
Timeline
• Spring 2001: Schmid
proposes idea; I say
“Impossible!”
• A.Y. 2001-2: Mabel I
hardware($20K), IBM
loan, course.
• Summer 2002: NSF REU,
Resnet support,
Edmonton Competition.
(Host, Rescue)
• A.Y. 2002-3 Mable II
hardware ($30K), Abbyy
gift, Dean’s travel grant,
course.
• Summer 2003. NSF REU.
Resnet support, Acapulco
Competition. (Host,
Rescue)
• Fall 2004: Robot Rivals,
Presentation to Trustees
• Spring 2004: Robocup
track in Art. Int. course
Face Finding
Rehearsal
Edmonton
C
O
M
P
E
T
I
T
I
O
N
Acapulco
Publications
• American Assoc. of Artificial Intelligence
2004
• Int. Joint Conference on Artificial
Intelligence 2003
• UR Journal of Undergraduate Research
• Rochester Review Article
Current and Future Placement
•
•
•
•
•
•
•
•
•
Schmid ---USC
Meisner ---RPI
Turner ---Brown
Sweetser ---Brown
Elsner --- UR Sr.
Calarese --- UFl
Purav --- UR grad
Peramunage --- UR Sr
Altekar --- Berkeley
• Kollar -- MIT
• Feil-Seifer ---USC
• Camara --- Take 5, UR
‘04
• Atwood ---EMC
• Cragg --- UR ‘04
• Isman --- UR ‘04
Then What Happened
• At this point we saw the Edmonton Video,
• Then a few seconds of the Acapulco Video,
• Then two spreadsheets detailing the joint
faculty and undergraduate research projects
that have happened since 1991. A huge
number, and growing every year to 52 in
2003.
Finally, the Mandatory
“Advice to Graduates”
• In your future, research skills are survival
skills.
• Pain is the feeling of weakness leaving the
body AND mind.
• Don’t be too quick to dismiss problems as
impossible -- you may be selling someone
else short. Or worse, selling yourself short.
Photo Credits
•
•
•
•
•
•
•
•
Marty Guenther
Jenine Turner
Tom Kollar
Michael Scott
Weilie Yi
Various Roboteers
Currents
UR Jnl. Of Undergraduate Research