Workshop on Reasoning in Adversarial and Noncooperative Environments (WRANE)
Duke University, November 18-19


The Workshop on Reasoning in Adversarial and Noncooperative Environments (WRANE) is sponsored by ARO and will be held at the Millennium Hotel next to Duke University. The main meeting room is Greenbriar AB, which is half of the second floor ballroom.

The workshop will be dedicated to the topic of reasoning in strategic settings, in which there are parties with opposite or nonaligned interests, with a focus on how these techniques can be used by DoD. The technical directions to be pursued include game theory, modeling and representing strategic situations, algorithms and combinatorial optimization, and inference and learning. The workshop will lead to a report on the state of the field with recommendations for topics to be researched and funded.

Please contact the organizers, Vincent Conitzer and Ronald Parr, with any questions.

Travel directions from RDU airport: the most convenient way is to take a taxi from RDU airport to the Millennium Hotel (2800 Campus Walk Avenue). This will take approximately 23 minutes according to the Google Maps directions.

List of participants

External academic participants
Itai Ashlagi (MIT Sloan)
Lance Fortnow (Northwestern EECS)
Drew Fudenberg (Harvard Economics)
Geoff Gordon (CMU Machine Learning Department)
Rachel Greenstadt (Drexel CS)
Seth Guikema (Johns Hopkins Geography and Environmental Engineering)
Richard John (USC Psychology)
Chris Kiekintveld (U Texas El Paso CS)
Georgios Kotsalis (Georgia Tech ECE)
Rajiv Maheswaran (USC Information Sciences Institute)
Asu Ozdaglar (MIT EECS)
Jesus Rios (IBM Research)
Tuomas Sandholm (CMU CS)
Bart Selman (Cornell CS)
Jeff Shamma (Georgia Tech ECE)
Satinder Singh (Michigan EECS)
VS Subrahmanian (Maryland CS)
Katia Sycara (CMU Robotics Institute)
Milind Tambe (USC CS & Industrial and Systems Engineering)
Matthew Taylor (Lafayette CS)
Yevgeniy Vorobeychik (Sandia National Labs)
Michael Wellman (Michigan EECS)
Martin Zinkevich (Yahoo! Inc.)

Local academic participants
Vincent Conitzer (Duke CS & Economics)
Brendan Daley (Duke Fuqua)
Sylvia Ferrari (Duke Mechanical Engineering and Materials Science)
Dmytro Korzhyk (Duke CS)
R. Vijay Krishna (UNC Economics)
Joshua Letchford (Duke CS)
Aleksandar (Sasa) Pekec (Duke Fuqua)
Ronald Parr (Duke CS)
Munindar Singh (NCSU CS)
Peng Sun (Duke Fuqua)
Lirong Xia (Duke CS)

Government participants
Harry Chang (ARO)
Purush Iyer (ARO)
Sven Koenig (NSF)
Nino Srour (ARL)
Cliff Wang (ARO)
Jun Zhang (AFOSR)


Schedule
Note: panelists may give a brief introductory presentation at the beginning of the panel; panelists who are also giving an invited talk should keep their intro presentation short.

THURSDAY Topic Materials
7:30am-8am Continental breakfast (2nd floor balcony)
8am-8:05am Welcome
8:05am-8:42am Invited talk: VS Subrahmanian.
Some Research on Adversarial Behavioral Models and Games.
8:42am-9:19am Invited talk: Tuomas Sandholm.
Automatically Generating Game-Theoretic Strategies for Huge Imperfect-Information Games.
slides
9:19am-9:56am Invited talk: Mike Wellman.
Empirical Game-Theoretic Analysis.
slides
9:56am-10:10am Break
10:10am-11:10am Panel: Computing game-theoretic solutions.
Fortnow, Selman, Vorobeychik.
Fortnow slides, Vorobeychik slides
11:10am-11:57am Invited team talk: Milind Tambe, Chris Kiekintveld, Richard John.
Game theory for security: Lessons learned from deployed applications.
slides
noon - 1pm Lunch
1pm-1:37pm Invited talk: Asu Ozdaglar.
Opinion Dynamics and Influence in Conflict Situations: The Role of Networks.
slides
1:37pm-2:40pm Panel: How do we model opponents?
Greenstadt, Rios; John, Sandholm, Subrahmanian.
Rios slides, Sandholm slides
2:40pm-2:55pm Break
2:55pm-4pm Panel: Learning and dynamics.
Fudenberg, Gordon, Shamma, Singh; Ozdaglar.
Gordon slides, Shamma slides
4pm-5pm Panel: How do we operationalize insights from behavioral/experimental game theory?
Maheswaran, Sycara; Tambe.
6pm Dinner
FRIDAY Topic Materials
7:30am-8am Continental breakfast (2nd floor balcony)
8am-9am Panel: Settings with many players.
Ashlagi, Guikema; Wellman.
Guikema slides, Wellman slides
9am-10am Panel: How do we evaluate our techniques?
Taylor, Zinkevich; Kiekintveld.
Taylor slides, Zinkevich slides
10am-10:15am Break
10:15am-11:55am Breakout groups to discuss main issues raised at workshop.
(1a) Modeling & acting in dynamic settings with many networked agents. Discussion leader: Subrahmanian.
(1b) What have we learned from testbeds? Discussion leader: Zinkevich.
(2a) How should we use behavioral/experimental game theory? Discussion leader: Kiekintveld.
(2b) How should we/our agents learn in environments with agents we do not control? Discussion leader: Gordon.
Gordon slides, Zinkevich slides
noon - 1pm Lunch.
1pm-3pm Breakout groups summary, final discussion, workshop conclusion.
3pm Final snacks, informal discussion, goodbyes; Duke campus tour.