|
Contact Information:
j e a n o h @ cs.cmu.edu
Language Technologies Institute
School of Computer Science
Carnegie Mellon University
5000 Forbes Ave.
Pittsburgh, PA 15213
NSH 1502F
(412) 268-5921
Fax: (412) 268-5569
Other stuff
"I use gmail and my id is damcci. Some people asked what it meant. Here's Damcci, the Brave, who once caught a snake, and who survived from the attack of a pit bull."
This is me in my old office with two windows. According to my architect friend in Chicago, windows (or Windows®?) decrease the efficiency at work in general. I think he's right about that.
Someday, I wish to visit here and here.
I'm thrilled to learn that
these boys
have been cheering for me since before I was even born.
|
|
|
Dissertation research
Thesis Committee:
Stephen F. Smith, Chair
Jaime Carbonell
Manuela Veloso
Sarit Kraus, Bar-Ilan University, Israel, and University of Maryland, College Park
I proposed my thesis "A few good agents: multi-agent social learning towards factored correlated equilibria" in December 2007, and hope to complete it by Spring 2009.
|
Teaching Assistant experience
- Artificial Intelligence: Representation and Problem Solving, 15-381, Spring 2008
- Advanced Artificial Intelligence: 15-780/16-731, Spring 2005
|
Current Projects @CMU
CMRADAR |
RADAR (2004 ~ present)
"Radar, Radar, what do you see? I see
CALO looking at me."
CMRadar is a multi-agent scheduling system in which a scheduling agent learns
its master user's scheduling preferences by observing a series of
meeting scheduling episodes.
This framework has been extented
to solve a highly constrained room finding problem under a crisis scenario
in which CMRadar efficiently selects room options which
minimizes the number of bumping (preemption) requests
by learning room owners' willingness to give up a room upon a request.
|
Interdisciplinary collaboration: AI in urban planning (2006 ~ )
In this project, we explore opportunities for applying AI techniques in urban design/planning problem domains.
This project is a joint work with my friend and collaborator,
Jie-Eun Hwang at Northeastern University in Boston.
(We used to play in the same orchestra many years ago.
She is a great timpanist, and I am a
"not so great yet trying hard" violinist.)
Heuristic Nolli Map (Demo)
|
|
Finding Main Streets
As an illustrative example of using machine learning techniques
in urban design problems we have developed an interactive design decision support
system that can assist designers to identify a certain type of urban setting,
a candidate Main Street, by utilizing active learning algorithms.
A Main Street
represents a commercial district in the center of a residential area
that has potential for revitalization, and its selection criteria is
based on complex architectural and socioeconomic features of its vicinity.
|
|
RAISE (Representativa Agents in Intelligent Survey Environment)
Recent catastrophic disasters such as hurricane Katrina have
brought increasing concerns on disaster management and
hazard mitigation.
RAISE is an agent-based urban planning decision support system
in aid of post-disaster recovery and reconstruction.
In particular, we are interested in resolving the NIMBY (not in my back yard) dilemma
raised in many urgent siting issues in post-disaster settings.
Related sites: FEMA,
HUD,
American Planning Association
|
Projects in the past @ISI, Marina del Rey, CA
Projects in the Stone Age
|
|
Jean Oh
Jean Oh
