Program & Exhibit Guide

NINETEENTH NATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE (AAAI-)

SIXTEENTH CONFERENCE ON INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE (IAAI-)

July –, 

San Jose McEnery Convention Center and San Jose Marriott San Jose, California

SPONSORED BY THE AMERICAN ASSOCIATION FOR ARTIFICIAL INTELLIGENCE

Cosponsored by ACM/SIGART, Defense Advanced Research Projects Agency (DARPA), Google, IBM Research, Intelligent Information Systems Institute (IISI), Cornell University , Microsoft Research, NASA Ames Research Center, National Science Foundation, and the Naval Research Laboratory ■ Student Abstract and Poster Cochairs: Avi Contents Pfeffer, Harvard University and Mark Craven, University of Wisconsin, Madison Acknowledgments /  Awards /  ■ Tutorial Forum Chair: Matthew Stone, Conference at a Glance /  Rutgers University Doctoral Consortium /  ■ Workshop Program Chair and Cochair: Exhibition /  Milos Hauskrecht, University of Pittsburgh General Information /  and Dieter Fox, University of Washington    IAAI- Program / – A complete listing of the AAAI- and IAAI-   Intelligent Systems Demonstrations / –  Program Committee members appears in   Invited Talks / – the conference proceedings. Maps / ,  National Botball Tournament /  NCER /  Sponsoring Registration /  Robot Competition and Exhibition / – Organizations Special Events and Awards / – Special Meetings /  AAAI gratefully acknowledges the generous Sponsoring Organizations /  contributions of the following organizations  Technical Program / – to AAAI- : Tutorial Forum / – ■ American Association for Artificial Intelli- Workshop Program / – gence ■ ACM/SIGART Acknowledgments ■ Defense Advanced Research Projects Agen- cy (DARPA) The American Association for Artificial Intel- ■ Google ligence wishes to acknowledge and thank the ■ IBM Research following individuals for their generous con- ■ Intelligent Information Systems Institute tributions of time and energy to the successful (IISI), Cornell University creation and planning of the Nineteenth Na- ■ Microsoft Research tional Conference on Artificial Intelligence and the Sixteenth Conference on Innovative ■ NASA Ames Research Center Applications of Artificial Intelligence. ■ National Science Foundation ■ AAAI Conference Committee Chair: James ■ Naval Research Laboratory A. Hendler, University of Maryland ■ AAAI- Program Cochairs: George Fergu- son, Rochester University and Deborah L. McGuinness, Stanford University ■ IAAI- Conference Chair: Randall Hill, Institute for Creative Technologies, Universi- ty of Southern California Contents / Acknowledgments Acknowledgments / Contents ■ IAAI- Conference Cochair: Neil Jacob- stein, Teknowledge Corporation ■ Intelligent Systems Demonstrations Chair: Christopher Welty, IBM Research ■ Mobile Robot Competition and Exhibition General Cochairs: Bill Smart, Washington University in St. Louis, and Sheila Tejada, University of New Orleans ■ AAAI/SIGART  Doctoral Consortium Chair: Robert St. Amant, North Carolina State University

2 Special Events & Awards : , in 3   :  r Dr.   , . iversity;   :   – to hono   AI Magazine. :   :  anford University; anford zona State Univer- zona State iversity.  iversity of Southern St ri Harvard Un AI Magazine ogram Cochair, and Un iversity of Southern Cal- ,from ,from Pr  Rice Un Un  n Dinner llows Recognition Dinner will Recognition llows Fe Daphne Koller, ynote Address,ynote “Intelligent Systems Chair Randall Hill and Cochair Neil and Cochair Chair Randall Hill  Stuart M. Shieber, Ke aig A. Knoblock,  Cr  bstein, IAAI- bstein will the four winnersbstein announce (see plications Awards gazine,AI applications community, and the The bbarao Kambhampati, A onsored by IAAI-04 and by IAAI-04 onsored cture was established in was established cture ntributions the field of to artificial intelli- co co ward will be presented to Edward Feigen- Edward to willward be presented the Pagoda Restaurant of Restaurant the Pagoda Hotel the Fairmont AAAI Fellows AAAI Fellows Recognitio Each year, the American Association for Artifi- of a small number recognizes cial Intelligence made significant sustained members who have co gence, distinc- attained unusual have and who tion in the profession. an- AAAI is pleased to for Fellows the six newly elected nounce sity; Ja for paper titles).schedule Certificates will be during paper sessions. presented Robert S. Englemore Memorial Award and Lecture Sp Su IAAI Deployed IAAI Deployed Ap IAAI- baum, of Professor Kumagai Sci- Computer Neil by Emeritus University ence at Stanford Ja Engelmore’s extraordinaryAAAI, service to Engelmore’s AI Ma a and his contributions to applied AI. applied and his contributions to The California; Y. and Moshe Vardi, The lecture will be held Wednesday, willThe lecture be held July Ballroom,San Jose Marriott at G.Gérard Medioni, The Robert S. and Award Memorial Engelmore Le Leake,David Editor-in-Chief, ifornia; July be held Tuesday,  .   :  sity of Mas- ft Research, ft Research, Univer Microso Microso Theory of for Action sity of Texas, Austin

ynote Address ynote in the San Jose Ballroom of Ballroom Jose in the San the Marriott. ong Kong University of University ong Kong H iversity of Massachusetts  and the Nations Vision for Space Exploration,” Space for Vision and the Nations on Tuesday, July Ke Daniel J. Clancy, Chief, Division, Sciences Computational Re- Ames NASA Center,search AAAI- will the give Univer Un .  aul Viola, Lin Liao, Fox, Dieter and Henry P iversity of Washington AAAI Classic Paper Award will be Award AAAI Classic Paper AAAI Distinguished ServiceAward Un ble Mention active Informationactive Extraction with tz, nstrained Random Fields. Conditional utines. on Culotta, austi Kristjansson, ohyung Lee, ohyung   mherst, an Jose Ballroom,an Jose Marriott, at hairs Deborah L. and George McGuinness aordinaryAAAI and his out- service to esque for “The Tractability of“The Tractability esque for Subsumption and Lin Fangzhen, ScienceTechnology and sachusetts Amherst A and Andrew McCallum, Loop Formulas for Circumscription. Loop Formulas Jo Inter Co Tr Learning and Inferring Transportation Ro Ar Kau c ll be presented to Bruce Buchanan for his Buchanan Bruce to ll be presented rguson. tr ,S emplify high standards in technical contri- emplify high in technical standards ulti-Agent Planning.”ulti-Agent pre- All papers were mplicit and Explicit Belief.”mplicit and Explicit Men- Honorable ■ Honora ■ Fe ■ Outstanding Paper Award Outstanding Paper Co Outstanding Paper Awards Outstanding Paper This year,Ar- on Conference National AAAI’s papers that honors three tificial Intelligence ex bution and exposition.AAAI- by Presented The Distinguished ServiceAward Classic Paper Award Classic Paper The Logic of “A for Levesque Hector to presented I and Hector Brachman Ronald to tion is given Lev Descriptionin Frame-based Languages” and Michael Georgeoff“A for AAAI- at sented AAAI Awards will Tuesday,be presented Awards AAAI July  M standing contributions to the AI community. the standing contributions to ex wi San Jose. A reception will begin at : ,fol- AAAI Business Meeting lowed by dinner at :  (by invitation only). The AAAI Annual Business Meeting will be held Thursday, July . : – : ,Meeting Opening Reception Room A,San Jose McEnery Convention Cen- ter. The AAAI- Opening Reception will be held Monday, July , : – :  in the San Jose Ballroom of the Marriott. This event will pro- AAAI Conference vide the traditional opportunity for attendees Committee Meeting to socialize in a unique setting prior to the be- ginning of the first day of technical sessions. A The Conference Committee Meeting will be variety of hors d’oeuvres and a no-host bar held Thursday, July , : – : , will be available. Admittance to the reception Guadalupe Room, San Jose Marriott. is free to AAAI- registrants. A $. per person fee ($. for children) will be charged for spouses and other nontechnical AAAI Publications conference registrants. Committee Meeting

The Publications Committee Meeting will be

ngs,& Posters Student Abstract held Tuesday, July , : – : , Poster Program Guadalupe Room, San Jose Marriott. Students whose abstracts were chosen for in- clusion in the conference proceedings will dis- AI Journal Editorial play their work at the Student Abstract Poster Session in the Exhibit Concourse on Tuesday, Board Meeting July  from : – : .In addition, par- The AI Journal Editorial Board lunch will be ticipants in the AAAI/SIGART Doctoral Con- held Wednesday, July , : – : , sortium will display their poster presentations Guadalupe Room, San Jose Marriott. during this session. All students will be avail- able for questions. Refreshments during the afternoon coffee break accompanying this ses- AI Magazine Editorial sion are sponsored by IBM Research. Board Meeting

Doctoral Consortium (DC-04) The AI Magazine Editorial Board Meeting will be held Monday, July , : – : ,Sa- The Ninth AAAI/SIGART Doctoral Consor- lon II, San Jose Marriott. tium program will be held on Sunday and Monday, July - from :  – :  in the Almaden Room of the San Jose Marriott. Executive Council Meeting The Doctoral Consortium provides an oppor- tunity for a group of Ph.D. students to discuss The AAAI Executive Council Meeting will be        and explore their research interests and career held Monday, July , : – : ,Sa- objectives in an interdisciplinary workshop lon I, San Jose Marriott. Continental breakfast will be available at : .

Receptions, Meeti together with a panel of established re- searchers. The twelve students accepted to participate in this program will also partici- pate in the Student Poster program on Tues- Program Committee Luncheon day, July ,from : – :  in the Exhib- The AAAI- / IAAI- Program Committee it Concourse. All interested AAAI- student Luncheon will be held Tuesday, July , : – registrants are invited to observe the presenta- : , in the Willow Glen Room, San Jose tions and participate in discussions at the Marriott, to honor the contributions of all the workshop. AAAI and ACM/SIGART gratefully members of the AAAI- and IAAI- Pro- acknowledge a grant from the National Sci- gram Committees (by invitation only). ence Foundation and Microsoft that provides partial funding for this event. 4 Conference at a Glance 5 28 29 27 26 ULY 25 ULY ,J ULY ,J ULY kshops Workshops kshops Workshops ial Forum Forum Tutorial NCER Reception ial Forum Forum Tutorial r r ULY istration AAAI-04 / IAAI-04 istration Registration Workshop Robot istration Registration NCER NCER r r ,J ,J g g MORNING AFTERNOON EVENING ,J d Presentations Exhibition / IS Demos d Presentations Exhibition / IS Demos to to Wo Wo ynote Address ynote Session Poster Student Collegiate Botball Re Re Registration Registration Dinner AAAI Fellows Registration Registration Opening Reception Reg te te Tu Tu Ke vi vi AAAI/SIGART DCAAAI/SIGART DC AAAI/SIGART AAAI/SIGART DC AAAI/SIGART DC AAAI-04 / IAAI-04 Meeting AAAI Business AAAI-04 / IAAI-04 AAAI-04 / IAAI-04 Collegiate Botball AAAI-04 / IAAI-04 AAAI-04 / IAAI-04 bots / Botball / NCER bots / Botball / NCER In In EDNESDAY ONDAY Exhibition / IS Demos / Botball / NCER Robots Exhibition / IS Demos / Botball / NCER Robots HURSDAY UESDAY Ro Ro UNDAY T W T M S Tutorial Forum the proceedings room. distribution Tutorial attendees may redeem theirtutorial syllabitickets at be taken dueto schedules. parallel AAAI- Forum Tutorial 6 Foundations and Applications Stochastic LocalSearch: MA2: MA1 Market Clearing Algorithms :  July 26 Monday, Session III: Preferences SP1: : July 25 Sunday, Session II: Programming Genetic Automated Invention by Means of SA2: Collaborative Multiagent Systems SA1: :  July 25 Sunday, Session I: co A:High Level Vision MA3: Exactand Approximate Inference inProbabilistic Models Graphical SP2: Building Intelligent Agents Using Soar(FULLDAY)SA3: A:AITechniques forPersonalized Recommendation (FULLDAY)MA4: mprehensive AAAI- A Ho A T A Jo A Jo A Ba A Dav A Kev B Jo A Jo uomas Sandholm –          nR Koza andLee Spector hn R. nE Laird hn E. nRel nhn aeo,andJoseph Konstan Anthony Jameson, hn Riedl, n Doyle and Ulrich Junkern Doyle andUlrich ,S / / ,S br rs,CaleOtz andMilind Tambe Ortiz, Charlie rbara Grosz, ,S ,S ,S ,S ,S grH Hoos and Thomas Stützle lger H. :     in Murphy dA Forsyth id A. ,S ,S an Jose Convention McEnery Center – – an Jose Convention McEnery Center an Jose Convention McEnery Center an Jose Convention McEnery Center an Jose Convention McEnery Center an Jose Convention McEnery Center an Jose Convention McEnery Center :  :  an Jose Convention McEnery Center an Jose Convention McEnery Center te c nclrgsrto nldsamsint pt orttrasadoecp fthe includesadmissionto upto fourtutorials andonecopyhnical registration of  Tu to r a ou oe.Amxmmo fourconsecutive tutorials may Amaximum of ial Forum Notes. Tutorial Forum & Marriott Map 7

M O O V R N L O L L A M A I B O

S eting Space O V R N L O L L A A Marriott San Jose Me B S M O O IV R N L O L L A A B S

M O III O R N L O L L A A W B S O I L M N HILL I IL E BLOSSOM O L I W O G R N L O L L A A S M W B II O O II L O N HILL II R N IL E BLOSSOM L O W L L L G A A B S

W O III L N BLOSSOMHILL III IL E W L

G

R

E

T ALMADEN

N

E Third Floor

C

N S IO N

O

T A L

N S

E S R

S V A

E N

N C C O O Y

AC C E D C I L

L R

Y I L M E

R S A A

O P

E V O

O

N B U

R E L C A

D unsberger M A

H U an Jose McEnery Convention Center McEnery Convention an Jose an Jose McEnery Convention Center McEnery Convention an Jose an Jose McEnery Convention Center McEnery Convention an Jose

id M. P. and Michael Pennock Wellman G :  v ,S ,S ke ,S aneesh Agrawala, Heiser, Julie and Tversky Barbara    – Da A M Lu A A : MP1: in Uncertainty: Markets Risk, Gambling,Aggregation and Information Session IV:Session Monday, 26 July MP3: the Design of Automating Visualizations MP2: Problems in Planning Reasoning Temporal Quantitative Workshop Program

Attendance at the workshops is limited, and participation is by invitation only. All workshop par- ticipants must register for the AAAI- technical program.

Sunday, July 25

W2: Agent Organizations: Theory and Practice Organizers: Virginia Dignum, Daniel Corkill, Catholijn Jonker, and Frank Dignum Guadalupe, San Jose Marriott :  – :  W4: Challenges in Game AI (Two-Day) Organizers: Dan Fu and Jeff Orkin B/,San Jose McEnery Convention Center :  – :  W6: Fielding Applications of Artificial Intelligence Organizer: William E. Cheetham San Carlos, San Jose Marriott :  – :  W13: Semantic Web Personalization Organizers: Bamshad Mobasher, Sarabjot Singh Anand, Bettina Berendt, and Andreas Hotho Willow Glen II, San Jose Marriott :  – :  W16: Supervisory Control of Learning and Adaptive Systems Organizers: Michael Rosenstein and Mohammad Ghavamzadeh Willow Glen III, San Jose Marriott

Workshop Program :  – : 

Monday, July 26

W1: Adaptive Text Extraction and Mining Organizer: Ion Muslea Guadalupe, San Jose Marriott :  – :  W3: Anchoring Symbols to Sensor Data Organizers: Silvia Coradeschi and Alessandro Saffiotti Blossom Hill I, San Jose Marriott :  – :  W4: Challenges in Game AI (Two-Day) Organizers: Dan Fu and Jeff Orkin B/,San Jose McEnery Convention Center :  – :  W7: Forming and Maintaining Coalitions in Adaptive Multiagent Systems Organizers: Leen-Kiat Soh and John E. Anderson Willow Glen III, San Jose Marriott :  – : 

8 W8: Intelligent Agent Architectures: Program Workshop Combining the Strengths of Software Engineering and Cognitive Systems Organizers: Randolph M. Jones, Robert E. Wray, and Matthias Scheutz Blossom Hill II, San Jose Marriott :  – :  W9: Learning and Planning in Markov Processes—Advances and Challenges Organizers: Daniela P. de Farias, Shie Mannor, Doina Precup, and Georgios Theocharous Willow Glen I/II, San Jose Marriott :  – :  W14: Sensor Networks Organizers: Gaurav S. Sukhatme, Adnan Darwiche, and Deborah Estrin Blossom Hill III, San Jose Marriott :  – :  W15: Spatial and Temporal Reasoning Organizers: Hans W. Guesgen and Gerard Ligozat San Carlos, San Jose Marriott :  – : 

Wednesday, July 28

W17: Mobile Robotic Competition and Exhibition Workshop Organizer: Frederick Crabbe A,San Jose McEnery Convention Center :  – : 

9 AAAI- invited talks will be held in the San broad application. The vision is still sound and Jose Ballroom of the San Jose Marriott and will be reinvigorated by two new develop- IAAI- panels will be held in Meeting Room ments: semantic web languages, which provide B/ of the San Jose Convention Center. stronger knowledge sharing technology; and the emergence of pervasive computing envi- ronments, whose requirements match the Tuesday, July 27 strengths of multiagent systems.

: – :  :  – :  AAAI-04 Keynote Address: AAAI-04 Invited Talk: Intelligent Systems and the Nation’s Real Robots for the Real World Vision for Space Exploration Sebastian Thrun, Stanford University Daniel J. Clancy, NASA Ames Research Center In , nine coal mines were almost buried Over the past few months, NASA has been in alive in Somerset, Pennsylvania, when acciden- the news numerous times both for the exciting tally breaching a neighboring mine submersed discoveries of the pair of Mars Rovers as well in water, and thought to be several hundred as the country's vision for space exploration. yards away. According to a  survey “tens of This vision focuses on the joint human and thousands, perhaps even hundreds of thou-

Invited Talks robotic exploration of the solar system starting sands, of abandoned mines exist today in the with a return to the moon and then the human United States. Not even the U.S. Bureau of exploration of Mars. In addition, the vision in- Mines knows the exact number, because feder- cludes a continued array of exciting robotic al recording of mining claims was not required missions to Mars as well as other destinations. until .” An improved understanding of the There are numerous challenges that must be sizes, locations, and conditions of abandoned addressed to accomplish this vision. One criti- mines could substantially reduce the many cal area of technology is the broad area of in- risks to society that arise from subsidence, formation technology with a particular em- contamination, and other factors characteris- phasis on intelligent systems. Future human tic of many abandoned mines today. exploration will require a new level of technol- This talk describes the Groundhog robot ogy enabling NASA to deploy smarter, more system, developed to acquire volumetric maps adaptive systems. These systems must be de- abandoned mines. Thrun will discuss some of signed to complement their human partners. the challenges that arise when sending a robot In this talk, Clancy will present a brief into mines, where it has to operate truly au- overview of the new vision and then will talk tonomously, outside the communication in depth about a range of technologies that range of its creators. Specifically, he will pre- have been developed and demonstrated by sent some recent innovations on large-scale NASA over the past few years. In particular, he robotic mapping, exploration, and navigation will focus on how advanced artificial intelli- in unknown and hazardous environments. gence technologies have been developed and These techniques leverage some of the most demonstrated within a robust environment advanced work in large-scale graphical mod- and then in some cases deployed on NASA els, probabilistic inference, and decision mak- missions. Technologies that will be discussed ing under uncertainty—topics that have been include: automated planning and scheduling, at the heart of AI research for more than a human-centered computing, remote collabo- decade. ration tools, machine learning and multi- The Groundhog project was jointly con- model interaction using EEG and EMG sig- ducted with Red Whittaker and Scott Thayer, nals. who led the robot development, and the Carnegie Mellon University Mobile robot de- : – :  sign class. Some of the material presented was contributed by researchers from the Fraun- AAAI-04 Invited Talk: hofer Institute in Germany. Agents Meet the Semantic Web in the Aether Tim Finin, University of Maryland, Baltimore County The software agents paradigm has received considerable research but has not yet found 10 Invited TalksInvited : – :  : – :  AAAI-04 Invited Talk: AAAI-04 Invited Talk: Strength of Character: If Not the Turing Test, Then What? AI Personalities in Video Games Paul Cohen, University of Massachusetts Amherst Ian Lane Davis, Mad Doc Software and USC Information Sciences Institute Davis will examine the upcoming great fron- Turing proposed his indistinguishability test tier in AI, which is the construction of believ- more than half a century ago, and although it able characters in video games. He will touch is never far from the thoughts of AI re- on all the aspects of making believable charac- searchers, it has done us little good. While oth- ters from animations to behaviors to internal er sciences seem motivated by their grand models of AI. challenges, AI does not strive to pass Turing's test. It is too remote a prospect, and the test it- : – :  self provides no compass or route markers. It is mired in philosophical problems, it ignores AAAI-04 Invited Talk: Human Dynamics physical and perceptual aspects of intelligence, and it has too many prerequisites. Yet Turing's Alex (Sandy) Pentland, MIT Media Laboratory test has virtues we should preserve in more ef- Important and complex discussions are almost fective grand challenges: it emphasizes com- always face-to-face, with little or no technolo- mon sense, universal scope, and understand- gy intervention. The result is that current ing; success is judged by ordinary people; and computer and communication tools are large- the test provides ample rope to hang oneself. ly restricted to mechanizing back-office and These are good criteria for grand challenge support functions. Pentland believes that we problems in general. can radically improve the functioning of the But there are other criteria that Turing's test organizations through active analysis of face- does not satisfy. We need progressions of chal- to-face interactions. By using wearable ma- lenges, spaced at two-to-five year intervals, chine perception to characterize, annotate, which will get us to Turing's goal within twen- and influence face-to-face interactions we can ty years, and parallel series of challenges in the augment peoples' memory, initiate productive perceptual/physical realm. We need “come as interactions, improve group decision making, you are” challenges that do not require some- and better understand and manage the func- thing else to be achieved before we can start tioning of distributed workforce. Examples of work.These problems should drive the devel- each type of augmentation will be shown, and opment of new technologies but discourage privacy concerns addressed. “self-inflicted” research. Ph.D. students should be able to trace very short paths from their dis- sertation topics to one or more of these prob- Wednesday, July 28 lems, and should see clearly the progression from it to longer-term challenges. The grand : – :  challenge problems should be updated regu- Joint AAAI-04 / IAAI-04 Robert S. Engelmore larly to deal with unforeseen impediments. Memorial Lecture: Cohen will give several possible examples The Power of “Clear and Demonstrative of progressions of grand challenge problems. Knowledge”: In Honor of a “True Son of It is time to put aside the Turing Test and re- Science” place it with challenges that work for us. Edward Feigenbaum, Stanford University This talk, given to honor the memory and : – :  work of Robert S. Engelmore, revisits some of AAAI-04 Invited Talk: the themes that were central to his work by AI and Biomedicine: Helping Scientists giving them a contemporary context. What is Reason about Genomes, Drugs, Diseases an up-to-date view of the “knowledge is pow- Russ B. Altman, Stanford University Medical Center er” hypothesis? What about knowledge webs, new applications of knowledge-based systems, Biology and medicine are now generating data and particularly the blackboard-architectures at a pace far exceeding the ability of scientists that RSE helped to develop? completely to analyze it. Bioinformatics has re- sponded with a number of powertools for re- ducing large data sets into manageable chunks. These tools are limited in two ways: they can't 11 read the published literature (or textbooks) Thursday, July 29 and they don't have biological common sense, so they sometimes propose things that are al- : – :  ready well-known or don't make sense (and thus can lose credibility). We are trying to AAAI-04 Invited Talk: ameliorate these problems with technology Applications of Artificial Intelligence to squarely within, or closely related to AI. Web Search Peter Norvig, Google : – :  Web search engines have accumulated billions of pages of text. A corpus of this size provides IAAI Robotics Panel an interesting opportunity for techniques in machine learning, natural language process- : :  – ing, and other areas of artificial intelligence. AAAI-04 Invited Talk: This talk will examine how these techniques Intelligent Technology for can be used in a large, practical web search en- Adaptive Aging gine. Martha E. Pollack, University of Michigan The world's population is rapidly aging, and :  – :  some are predicting a coming “crisis in care- IAAI-04 Panel Session: Invited Talks giving.” Intelligent technologies have the po- The Broader Role of Artificial Intelli- tential to help avoid this crisis, by providing gence in Large-Scale Scientific Research people with ways to adapt to the physical and Panel Organizer: Yolanda Gil, USC Information Sci- cognitive changes that can accompany aging. ences Institute This talk will survey the state-of-the-art in Panelists: David De Roure, University of Southamp- such technology, highlight the important role ton, Steven Meacham, National Science Foundation, of AI techniques in its design, and describe and Joel Saltz, The Ohio State University open research challenges. Many ambitious scientific endeavors are being pursued by shifting from an emphasis on indi- : :  – vidual efforts to cross-disciplinary and large- AAAI-04 Invited Talk: scale experimentation and analysis. This panel Building a Life Science Company will cover new directions and opportunities in around Statistical Machine Learning this area that will challenge the field of AI in Astro Teller, BodyMedia, Inc. the years to come. Current distributed com- A tour through five years of challenges and puting infrastructure supports sharing, man- discoveries building a wearable body monitor- aging, and accessing diverse resources such as ing business using statistical machine learning peta-scale data repositories, high performance techniques. The talk will cover challenges computing, sophisticated instruments, remote gathering data, building body state models, sensors and data collection facilities, docu- validating the models with the medical com- ment management and digital libraries, and munity, and place AI within the larger context people with complementary expertise about a of the company and the industry. phenomenon under study. The panelists will discuss how the U.S. National Science Founda- tion and the U.K. e-Science program are plan- ning for this future, and present some ongoing work under those and other programs using AI techniques to integrate data and scientific applications.

12 Almaden Terrace Almaden 24 hr Room Control Security G H J4 J Kitchen Rooms F Meeting Almaden Concourse VIP Lounge VIP Elevator Parking ers ilton F1 J3 J2 J1 F2 To & Tow San José To H

Staff Offices Staff

p am Food

Concourse 3 Loading R Loading (50,000 Square Feet) (50,000 Square Exhibit Hall 3 (above) Show Mgr’s Office To Dock Access Stair Ele- vator Parking

Food

Boulevard Almaden AED Concourse 2 Loading Docks (50,000 Square Feet) (50,000 Square

Exhibit Hall 2

FORTUNE 2 Dumpster Dumpster

(above) Compactor Lobby Show Mgr’s Office Almaden Security Starbucks To Offices To Control Room Control Entrance VIP Administrative Ele- Stair Stair vator To Almaden Blvd. Lounge Dock Access Ele- vator AED lton & owers T San José To Hi Street Level Street Office Food & Beverage Garage Entrance Food Lower Level ista Garage Entrance V Point Concourse 1 (43,000 Square Feet) (43,000 Square Exhibit Hall 1 Parking Show Office Mgr’s (above) Stair Ele- vator Stair Show Mgr’s Office To Ele- vator AED A4 A5 A3 A6 A2 A7 A1 A8 to level evel lower Ramp parking Ballroom Concourse Ballroom Ballroom A Ballroom evel (22,000 Square Feet) (22,000 Square l ATM l Hotel Arcade San José TTY Marriott To errace Market T C3 C4 C2 C1 B3 B4 B2 B1 XHIBIt treet C B s E Kitchen Meeting Rooms

Stair ce D Parking

E erra arket T Ele- M vator 700 on site parking spaces Entrance Main to Entrance San Carlos Street Registration Area level lower Ramp parking San Carlos Street ATM isitor Information & Business Center V Stair Ele- vator TTY Parking Hotel San José M Marriott To L Rooms Meeting Entrance KN

Market Street

Entrance Lower Level Garage

Market Street

13 7/27 8:30 – 10:00 AM 10:30 – 11:30 AM 11:40 AM – 12:40 PM

Welcome and Opening Remarks/Outstanding Invited Talk Invited Talk Award Presentations George Ferguson and Deborah McGuinness, Agents Meet the Semantic Web in the Aether Real Robots for the Real World Program Cochairs Tim Finin, University of Maryland, Baltimore Sebastian Thrun, Stanford University County IAAI-04 Deployed Application

iottt Award Announcements Randall Hill and Neil Jacobstein, IAAI Program rr Cochairs Ma AAAI Special Award Presentations Ron Brachman, AAAI President, and Tom

San Jose Ballroom, San Jose Mitchell, AAAI Past President

: – :  Extracting Opinions Information Extraction San Jose Ballroom, Marriott Mining Opinion Features in Customer Methods for Domain-Independent Reviews Information Extraction from the Web: AAAI-04 Keynote Address Minqing Hu and Bing Liu An Experimental Comparison Oren Etzioni, Michael Cafarella, Doug Downey, Intelligent Systems and the Nations Vision for Just How Mad Are You? Finding Strong and Ana-Maria Popescu, Tal Shaked, Stephen Soder- Space Exploration land, Daniel S. Weld, and Alexander Yates Daniel J. Clancy, NASA Ames Research Center Weak Opinion Clauses Theresa Wilson, Janyce Wiebe, and Rebecca Hwa

Center Outstanding Paper Honorable Mention: Interactive Information Extraction with eting Room A1, eting Room Constrained Conditional Random Fields Trausti Kristjansson, Aron Culotta, Paul Viola, Me and Andrew McCallum San Jose Convention San Jose

Temporal Planning Branching and Pruning: An Optimal Ontologies Temporal POCL Planner Based on Evaluating Ontology Cleaning Constraint Programming Christopher Welty, Ruchi Mahindru, and Vincent Vidal and Héctor Geffner Jennifer Chu-Carroll

Continuous Time in a SAT-based Planner Mereological Semantics for Bio-Ontologies

g Room A2, g Room Ji-Ae Shin and Ernest Davis Udo Hahn, Stefan Schulz, and Kornél Markó Center Meetin San Jose Convention San Jose

CP Nets mCP Nets: Representing and Reasoning with Preferences Preferences of Multiple Agents Low-cost Addition of Preferences to DTPs and Francesca Rossi, K. Brent Venable, and T. Walsh TCSPs Bart Peintner and Martha E. Pollack Extending CP-Nets with Stronger Conditional Preference Statements A Computational Study of the Kemeny Rule

g Room A7, g Room Nic Wilson for Preference Aggregation

Center Andrew Davenport and Jayant Kalagnanam Meetin San Jose Convention San Jose

Robot Learning Skill Acquisition and Use for a Dynamically- Robot Mission Planning Balancing Soccer Robot Advice Generation from Observed Execution: uesday, 27 July Brett Browning, 10:00 – 10:30 Ling Xu, and Break Coffee Manuela Veloso Abstract Markov Decision Process Learning

T Patrick Riley and Manuela Veloso AAAI / IAAI Machine Learning for Fast Quadrupedal Locomotion Interleaving Temporal Planning and

g Room A8, g Room Nate Kohl and Peter Stone Execution in Robotics Domains Center Tuesday Solange Lemai and Félix Ingrand Meetin San Jose Convention San Jose

Technical IAAI Deployed Applications Tenth Anniversary of the Plastics Color IAAI Emerging Applications Formulation Tool VModel: A Visual Qualitative Modeling William Cheetham Environment for Middle-School Students Kenneth D. Forbus, Karen Carney, Bruce L. Sessions Sherin, and Leo C. Ureel II Making Better Recommendations with Online Profiling Agents Danny Oh and Chew Lim Tan Branching Storylines in Virtual Reality

Center Environments for Leadership Development Andrew Gordon, Michael van Lent, Martin van Velsen, Paul Carpenter, and Arnav Jhala eting Room B2/3, eting Room Me San Jose Convention San Jose 14 2:00 – 3:00 PM 3:10 – 4:10 PM 4:30 – 5:30 PM

Invited Talk Invited Talk Strength of Character: AI Personalities in Human Dynamics Video Games Alex (Sandy) Pentland, MIT Media Laboratory Ian Lane Davis, Mad Doc Software

CSP Consistency Overconstrained Problems CSP Tractability The Practice of Approximated Consistency for Leap Before You Look:An Effective Strategy in Tractable Tree Convex Constraint Networks Knapsack Constraints an Oversubscribed Scheduling Problem Yuanlin Zhang and Eugene C. Freuder Meinolf Sellmann Laura Barbulescu, L. Darrell Whitley, and Adele E. Howe Callapsibility and Consistency in Quantified Domain Transmutation in Constraint Constraint Satisfaction Satisfaction Problems QuickXplain: Preferred Explanations and Hubie Chen James Bowen and Chavalit Likitvivatanavong Relaxations for Over-Constrained Problems Ulrich Junker

Machine Learning Learning Search Bayesian Network Classifiers Versus k-NN Error Detection and Impact-Sensitive Instance Temperature Discovery Search Classifier Using Sequential Feature Selection Ranking in Noisy Datasets Martin Müller, Markus Enzenberger, and Franz Pernkopf Xingquan Zhu, Xindong Wu, and Ying Yang Jonathan Schaeffer

Fibring Neural Networks On the Optimality of Probability Estimation Best-First Frontier Search with Delayed Artur S. d’Avila Garcez and Dov M. Gabbay by Random Decision Trees Duplicate Detection Wei Fan Richard E. Korf

Combinatorial Auctions POMDP’s Perception: Sketch Understanding Dinner,llows Hotel Fairmont

Methods for Boosting Revenue in Stochastic Local Search for POMDP Perceptually Based Learning of Shape Fe Combinatorial Auctions Controllers Descriptions for Sketch Recognition

Anton Likhodedov and Tuomas Sandholm Darius Braziunas and Craig Boutilier Olya Veselova and Randall Davis PM

Combinatorial Auctions with Structured Item Dynamic Programming for Partially Automatically Translating Symbolic Shape Graphs Observable Stochastic Games Descriptions for Use in Sketch Recognition Vincent Conitzer, Jonathan Derryberry, and Eric A. Hansen, Daniel S. Bernstein, and Shlomo Tracy Hammond and Randall Davis Tuomas Sandholm Zilberstein : ‒ :

Robot Navigation Resolution/Agents Partially Observable Planning Analogical Path Planning Implementing a Generalized Version of Regression with Respect to Sensing Actions Saul Simhon and Gregory Dudek Resolution and Partial States uesday, 27 July 12:40 – 2:00 Break Lunch Heidi E. Dixon, Matthew L. Ginsberg, David K. Le-Chi Tuan, Chitta Baral, Xin Zhang, and Tran T A Multi-Resolution Pyramid for Outdoor Hofer, Eugene M. Luks, and Andrew J. Parkes Cao Son Robot Terrain Perception Michael Montemerlo and Sebastian Thrun Useful Roles of Emotions in Artificial Agents: Distance Estimates for Planning in the A Case Study from Artificial Life Discrete Belief Space Matthias Scheutz Jussi Rintanen

IAAI Emerging Applications IAAI Emerging Applications IAAI Emerging Applications A Robotic Wayfinding System for the Visually AI Characters and Directors for Interactive Identifying Terrorist Activity with AI Plan Impaired Computer Games Recognition Technology Vladimir Kulyukin, Chaitanya Gharpure, Prad- Brian Magerko, John E. Laird, Mazin Assanie, Peter A. Jarvis, Teresa F. Lunt, and nya Sute, Nathan De Graw, and John Nicholson Alex Kerfoot, and Devvan Stokes Karen L. Myers The Independent LifeStyle Assistant™ Towards Autonomic Computing: Adaptive Job Agent-based Simulation of Geo-Political (I.L.S.A.): AI Lessons Learned Routing and Scheduling Conflict uesday, 27 July Karen Zita Haigh, Liana 4:10 – 4:30 M. Kiff, Janet Break Coffee Myers, Posters Student Shimon Whiteson and Peter Stone Glenn Taylor, Richard Frederiksen, Russell R. T Valerie Guralnik, Christopher W. Geib, John Vane III, and Edward Waltz Phelps, and Tom Wagner

15 7/28 8:45 – 10:00 AM 10:30 – 11:30 AM 11:40 AM – 12:40 PM

CRA Outstanding Undergraduate Awards Invited Talk Presented by Tim Finin, AAAI CRA Liaison If Not the Turing Test, then What? Paul Cohen, University of Massachusetts : – :  Amherst and USC Information Sciences Institute Joint AAAI-04 / IAAI-04 Robert S. Engelmore Memorial Lecture iottt The Power of “Clear and Demonstrative rr Knowledge”: In Honor of a “True Son of Science” Ma Edward Feigenbaum, Stanford University San Jose Ballroom, San Jose

Negotiation/Argumentation Answer Set Programming Answer Set Programming II Negotiation as Mutual Belief Revision Logic Programs with Abstract Constraint Adding Time and Intervals to Procedural and Dongmo Zhang, Norman Foo, Thomas Meyer, Atoms Hierarchical Control Specifications and Rex Kwok Victor W. Marek and Miroslaw Truszczynski Tran Cao Son, Chitta Baral, and Le-Chi Tuan

Making Argumentation More Believable On Odd and Even Cycles in Normal Logic SAT-Based Answer Set Programming Anthony Hunter Programs Enrico Giunchiglia, Yuliya Lierler, and Marco Fangzhen Lin and Xishun Zhao Maratea Center eting Room A1, eting Room Me San Jose Convention San Jose

Syntax and Semantics NLP / Collaborative Filtering Knowledge Acquisition On the Relationship between Lexical Identification and Tracing of Ambiguous Interpreting Loosely Encoded Questions Semantics and Syntax for the Inference of Names: Discriminative and James Fan and Bruce Porter Context-Free Grammars Generative Approaches Tim Oates, Tom Armstrong, Justin Harris, and Xin Li, Paul Morie, and Dan Roth Common Sense Data Acquisition for Indoor Mark Nejman Mobile Robots Exploring More Realistic Evaluation Measures Rakesh Gupta and Mykel J. Kochenderfer Distributed Representation of Syntactic g Room A2, g Room for Collaborative Filtering

Center Structure by Tensor Product Representation Giuseppe Carenini and Rita Sharma and Non-linear Compression Heidi H. T. Yeung and Peter W. M. Tsang Meetin San Jose Convention San Jose

Planning Temporal Reasoning Learning Preferences Transport Logistics Planning with A Polynomial-Time Algorithm for Simple Eliciting Bid Taker Non-price Preferences in Service-Level Constraints Temporal Problems with Piecewise Constant (Combinatorial) Auctions Hoong Chuin Lau, Kien Ming Ng, and Xiaotao Domain Preference Functions Craig Boutilier, Tuomas Sandholm, and Rob Wu T. K. Satish Kumar Shields

Effective Approaches for Partial Satisfaction Model Checking Temporal Logics of Learning Social Preferences in Games (Over-Subscription) Planning g Room A7, g Room Knowledge in Distributed Systems Ya’akov Gal, Avi Pfeffer, Francesca Marzo, and

Center Menkes van den Briel, Romeo Sanchez, Minh B. Kaile Su Barbara J. Grosz Do, and Subbarao Kambhampati Meetin San Jose Convention San Jose

Search Search Qualitative Modeling Simple Search Methods for Finding a Nash dnesday, 28 July A General 10:00 – 10:30 Solution to Break Coffee the Graph History A Qualitative-Quantitative Methods-Based Equilibrium Interaction Problem e-Learning Support System in Economic Ryan Porter, Eugene Nudelman, and Yoav Akihiro Kishimoto and Martin Müller Education

Shoham We Tokuro Matsuo, Takayuki Ito, and Toramatsu Structured Duplicate Detection in Shintani Towards Efficient Sampling: Exploiting External-Memory Graph Search Random Walk Strategies g Room A8, g Room Rong Zhou and Eric A. Hansen Spatial Aggregation for Qualitative

Center Wei Wei, Jordan Erenrich, and Bart Selman Assessment of Scientific Computations Chris Bailey-Kellogg and Naren Ramakrishnan Meetin San Jose Convention San Jose

IAAI Deployed Applications IAAI Emerging Applications The General Motors Variation-Reduction An Explainable Artificial Intelligence System Adviser: Deployment Issues for an for Small-unit Tactical Behavior AI Application Michael van Lent, William Fisher, and Michael Alexander P. Morgan, John A. Cafeo, Kurt God- Mancuso den, Ronald M. Lesperance, Andrea M. Simon, Deborah L. McGuinness, and James L. Benedict The KOJAK Group Finder: Connecting the Dots via Integrated Knowledge-Based and

Center Ergonomics Analysis for Vehicle Assembly Statistical Reasoning Using Artificial Intelligence Jafar Adibi, Hans Chalupsky, Eric Melz, and Nestor Rychtyckyj Andre Valente eting Room B2/3, eting Room Me San Jose Convention San Jose 16 2:00 – 3:00 PM 3:10 – 4:10 PM 4:30 – 5:30 PM

Invited Talk Invited Talk Invited Talk AI and Biomedicine: Helping Scientists Intelligent Technology for Adaptive Aging Building a Life Science Company around Reason about Genomes, Drugs, Diseases Martha E. Pollack, University of Michigan Statistical Machine Learning Russ B. Altman, Stanford University Medical Astro Teller, BodyMedia, Inc. Center

Nonmonotonic Reasoning SAT and CSP Hardness and Complexity Outstanding Paper, Honorable Mention: Complete Local Search for Propositional The Backdoor Key: A Path to Understanding Loop Formulas for Circumscription Satisfiability Problem Hardness Joohyung Lee and Fangzhen Lin Hai Fang and Wheeler Ruml Yongshao Ruan, Henry Kautz, and Eric Horvitz

Conservative Belief Revision Modeling Choices in Quasigroup Completion: The Complexity of Global Constraints James P. Delgrande, Abhaya C. Nayak, and SAT vs. CSP Christian Bessiere, Emmanuel Hebrard, Brahim Maurice Pagnucco Carlos Ansótegui, Alvaro del Val, Iván Dotú, Hnich, and Toby Walsh Cèsar Fernández, and Felip Manyà

Plan Recognition Ensembles Learning: Graphical Models Assessing the Complexity of Plan Recognition An Ensemble Technique for Stable Learners Bayesian Inference on Principal Component Christopher W. Geib with Performance Bounds Analysis using Reversible Jump Markov Chain Ian Davidson Monte Carlo High-Level Goal Recognition in a Wireless Zhihua Zhang, Kap Luk Chan, James T. Kwok, LAN Online Parallel Boosting and Dit-Yan Yeung Jie Yin, Xiaoyong Chai, and Qiang Yang Jesse A. Reichler, Harlan D. Harris, and Michael A. Savchenko Hierarchical Hidden Markov Models with General State Hierarchy Hung H. Bui, Dinh Q. Phung, and Svetha Venkatesh

MDPs MDPs Solving Concurrent Markov Solving Generalized Semi-Markov Decision Decision Processes Processes Using Continuous Phase-Type Mausam and Daniel S. Weld Distributions Håkan L. S. Younes and Reid G. Simmons Reinforcement Learning for CPG-Driven Biped Robot An Instance-based State Representation for Takeshi Mori, Yutaka Nakamura, Masa-aki Sato, Network Repair and Shin Ishii Michael L. Littman, Nishkam Ravi, Eitan Fenson, and Rich Howard

Deliberation and Control of Reasoning Multi-Robot Task Allocation Perceptual Map Making dnesday, 28 July 4:10 – 4:30 Break Coffee dnesday, 28 July Low-Knowledge 12:40 – 2:00 Algorithm Break Lunch Control Affective Recruitment of Distributed Large-Scale Map-Making Tom Carchrae and J. Christopher Beck Heterogeneous Agents Kurt Konolige Aaron Gage and Robin R. Murphy We We Using Performance Profile Trees to Improve Self-Organizing Visual Maps Deliberation Control Task Allocation via Self-Organizing Swarm Robert Sim and Gregory Dudek Kate Larson and Tuomas Sandholm Coalitions in Distributed Mobile Sensor Network Kian Hsiang Low, Wee Kheng Leow, and Marcelo H. Ang, Jr.

IAAI Session IAAI Session IAAI Emerging Applications Robotics Panel Robotics Panel A Comparison of Techniques for Scheduling (: – 4: ) (: – 4: ) Earth Observing Satellites Al Globus, James Crawford, Jason Lohn, and Anna Pryor

SOFIA’s Choice: An AI Approach to Scheduling Airborne Astronomy Observations Jeremy Frank, Michael A. K. Gross, and Elif Kürklü

17 7/29 9:00 – 10:00 AM 10:30 – 11:30 AM 11:40 AM – 12:40 PM

Invited Talk Applications of Artificial Intelligence to Web Search Peter Norvig, Google iottt rr Ma San Jose Ballroom, San Jose

Text Classification SAT Max – SAT Text Classification by Labeling Words Additive versus Multiplicative Clause MAX-2-SAT: How Good Is Tabu Search in the Bing Liu, Xiaoli Li, Wee Sun Lee, and Philip S. Yu Weighting for SAT Worst-Case? John Thornton, Duc Nghia Pham, Stuart Bain, Monaldo Mastrolilli and Luca Maria Gam- Learning Indexing Patterns from One and Valnir Ferreira Jr. bardella Language for the Benefit of Others Udo Hahn, Kornél Markó, and Stefan Schulz Hiding Satisfying Assignments: Two Are Study of Lower Bound Functions for Better than One MAX-2-SAT

Center Dimitris Achlioptas, Haixia Jia, and Cristopher Haiou Shen and Hantao Zhang Moore eting Room A1, eting Room Me San Jose Convention San Jose

Scheduling Multiagent Systems Planning in Nondeterministic Domains Regrets Only! Online Stochastic Optimization Performance Bounded Reinforcement Forward-Chaining Planning in under Time Constraints Learning in Strategic Interactions Nondeterministic Domains Russell Bent and Pascal Van Hentenryck Bikramjit Banerjee and Jing Peng Ugur Kuter and Dana Nau

An Effective Algorithm for Project Scheduling Visibility-Based Pursuit-Evasion with Limited Generating Safe Assumption-Based Plans for with Arbitrary Temporal Constraints Field of View Partially Observable, Nondeterministic Tristan B. Smith and John M. Pyle g Room A2, g Room Brian P. Gerkey, Sebastian Thrun, and Geoff Domains

Center Gordon Alexandre Albore and Piergiorgio Bertoli Meetin San Jose Convention San Jose

Recognition and Reconstruction Cognitive Modeling Reconstruction of 3D Models from Intensity On the Integration of Grounding Language Images and Partial Depth and Learning Objects Luz A. Torres-Méndez and Gregory Dudek Chen Yu and Dana Ballard

Rapid Object Recognition from Comparing Cognitive and Computational Discriminative Regions of Interest Models of Narrative Structure Gerald Fritz, Christin Seifert, Lucas Paletta, g Room A7, g Room David B. Christian and R. Michael Young

Center and Horst Bischof Meetin San Jose Convention San Jose

Probabilistic Reasoning/Inference Commonsense Reasoning Argumentation and Negotiation PROBCONS: Probabilistic Consistency-Based Scaling Up Reasoning about Actions Using Towards Higher Impact Argumentation Multiple Alignment of Amino Acid Sequences Relational Database Technology Anthony Hunter Chuong B. Do, Michael Brudno, and Serafim Giuseppe De Giacomo and Toni Mancini Batzoglou

Thursday, 29 July 10:00 – 10:30 Break Coffee Logical Foundations of Negotiation: Encoding Probabilistic Causal Model in Outcome, Concession, and Adaptation Complexity of Contextual Reasoning Probabilistic Action Language Thomas Meyer, Norman Foo, Rex Kwok, and Floris Roelofsen and Luciano Serafini g Room A8, g Room Nam Tran and Chitta Baral Dongmo Zhang Center Meetin San Jose Convention San Jose

IAAI Emerging Applications IAAI Panel: The Broader Role of IAAI Panel: The Broader Role of Synthetic Adversaries for Urban Combat Artificial Intelligence in Large-Scale Artificial Intelligence in Large-Scale Training Scientific Research Scientific Research Robert E. Wray, John E. Laird, Andrew Nuxoll, (:  – : ) (:  – : ) Devvan Stokes, and Alex Kerfoot Panel Organizer: Yolanda Gil, USC Information Panel Organizer: Yolanda Gil, USC Information Artemis: Integrating Scientific Data on the Sciences Institute Sciences Institute Grid Panelists: David De Roure, University of Panelists: David De Roure, University of

Center Rattapoom Tuchinda, Snehal Thakkar, Yolanda Gil, and Ewa Deelman Southampton, Steven Meacham, National Sci- Southampton, Steven Meacham, National Sci- ence Foundation, and Joel Saltz, Ohio State ence Foundation, and Joel Saltz, Ohio State

eting Room B2/3, eting Room University University Me San Jose Convention San Jose 18 2:00 – 3:00 PM 3:10 – 4:10 PM 4:30 – 5:30 PM

Learning Behaviors Economic Models Learning and Applying Competitive Strategies Computing Shapley Values, Manipulating Esther Lock and Susan L. Epstein Value Division Schemes, and Checking Core Membership in Multi-Issue Domains Outstanding Paper Award: Vincent Conitzer and Tuomas Sandholm Learning and Inferring Transportation Routines Using Contracts to Influence the Outcome Lin Liao, Dieter Fox, and Henry Kautz of a Game Robert McGrew and Yoav Shoham

Mechanism Design Planning/Temporal Reasoning Searching for Stable Mechanisms: Automated Shortest Path Discovery Problems: A Design for Imperfect Players Framework, Algorithms and Andrew J. Blumberg and Abhi Shelat Experimental Results Csaba Szepesvári GROWRANGE: Anytime VCG-Based Mechanisms Identifying Linear Causal Effects David C. Parkes and Grant Schoenebeck Jin Tian

Search Event Recognition Space-Efficient Memory-Based Heuristics CASEE:A Hierarchical Event Representation Rong Zhou and Eric A. Hansen for the Analysis of Videos Asaad Hakeem, Yaser Sheikh, and Mubarak Shah

Thursday, 29 July Compressing 12:40 – 2:00 Pattern Databases Break Lunch Thursday, 29 July 4:10 – 4:30 Break Coffee Ariel Felner, Ram Meshulam, Robert C. Holte, Repeated Observation Models and Richard E. Korf Avi Pfeffer

Emerging Applications IAAI Emerging Applications IAAI Emerging Applications Machine Learning for Adaptive Image An Application View of COORDINATORS: Automatic Generation of Artistic Chinese Interpretation Coordination Managers for First Responders Calligraphy Ilya Levner and Vadim Bulitko Thomas Wagner, John Phelps, Valerie Guralnik, Songhua Xu, Francis C.M. Lau, and Ryan VanRiper Kwok-Wai Cheung, and Yunhe Pan CaBMA: Case-Based Project Management Assistant The U.S. National Football League Scheduling Detecting and Eliminating the Cascade Ke Xu and Héctor Muñoz-Avila Problem Vulnerability Problem from Multilevel Bistra N. Dilkina and William S. Havens Security Networks using Soft Constraints Stefano Bistarelli, Simon N. Foley, and Barry O’Sullivan

19 The exhibition will be held in Exhibit Hall  on Booth # the exhibit level of the San Jose McEnery Con- ActivMedia Robotics vention Center, Tuesday, July  and Wednes-  Columbia Drive day, July .Admittance is restricted to badged Amherst, NH  conference attendees. Further information re- -- garding access to the Exhibition can be ob- ActivMedia Robots provide autonomous mo- tained from the Exhibitor Registration Desk. bile platforms on a variety of scales. Our high- ly useable ActivMedia Robotics Interface for Exhibit Hours Applications (ARIA) development API is Tuesday, July : :  – :  open-sourced under the GNU license for use Wednesday, July : :  – :  with C++. Our Laser Mapping and Navigation software enables our robots to know exactly where they are, and how to get to their desti- Exhibitors nation. Powerful software, reliable, extensible ActivMedia Robotics bases and our integrated ■ AAAI Press accessories are what make ActivMedia Ro- ■ Acroname Inc. botics platforms the most-used in robotics ■ ActivMedia Robotics R&D today. Come see a demonstration of our ■ Franz Inc. new PatrolBot in action! ■ Google ■ The Intelligent Information Systems Insti- Booth # tute at Cornell University ■ The MIT Press Franz Inc. ■ Morgan Kaufmann Publishers, an imprint  th Street, Suite   of Elsevier Oakland, CA Tel: --

Exhibit Program Exhibit ■ NASA Ames Research Center Fax: -- ■ Springer-Verlag New York, LLC. Franz Inc. is the leading vendor of Common Lisp (CL) and Common Lisp Object System (CLOS) development environment — Allegro AAAI Press CL. True to its roots in AI, Franz has released  Burgess Drive a fast SAX-based XML parser and a CLOS- Menlo Park, CA - based Prolog compiler for accessing and rea- Tel: -- soning over complex data. Franz also provides Fax: -- a SOAP and WSDL web service API for Lisp Email: [email protected] applications to access external information Online Catalog: and knowledge easily over the web. Along with www.aaai.org/Publications/Catalog/ high-performance direct interfaces to Oracle and MySQL, Allegro CL makes commercial Booth # deployment of AI and expert systems easy and Acroname Inc. practical. Stop by for an entertaining demon- stration of software robot—Pandorabots.  Sterling Driove Boulder, CO   Tel: -- Booth # Acroname has spent over a decade helping Google nearly , customers worldwide in  Ampitheater Parkway robotics sensing, control, and information. Mountain View, CA  This experience has been incorporated into Google’s innovative search technologies con- many of Acroname’s in-house products, such nect millions of people around the world with as the BrainStem line of intelligent controllers information every day. Founded in  by and motion control solutions as well as the Stanford Ph.D. students Larry Page and Sergey Garcia robot. Stop by the Acroname exhibit to Brin, Google today is a top web property in all see demonstrations of Garcia robots, discuss major global markets. Google’s targeted ad- your robotics needs, and pick up materials vertising program, which is the largest and outlining the hundreds of robotics solutions fastest growing in the industry, provides busi- offered by Acroname. Acroname can help you nesses of all sizes with measurable results, accelerate your research with open, inexpen- while enhancing the overall web experience sive, proven, and flexible robot platforms. for users. Google is headquartered in Silicon Valley with offices throughout North America, 20 Exhibitors Europe, and Asia. For more information on Booth # joining the engineering team, visit www. Morgan Kaufmann google.com/jobs/eng.html. An imprint of Elsevier  Park Avenue South Booth # New York, NY  The Intelligent Information Systems Tel: -- Fax: -- Institute, Cornell University Morgan Kaufmann, an imprint of Elsevier, be-  Upson Hall gan publishing its prestigious list of AI publi-  Ithaca, NY cations  years ago this year. New titles this -- year include three major text references: The mandate of the Intelligent Information Brachman & Levesque: Knowledge Representa- Systems Institute is threefold: to perform and tion and Reasoning; Hoos & Stützle: Stochastic stimulate research in compute- and data-in- Local Search Foundations and Applications; tensive methods for intelligent decision mak- and Ghallab, Nau & Traverso: Automated ing systems; to foster collaborations between Planning: Theory and Practice Cornell researchers, the Institute’s sponsors, and the scientific community; and to play a Booth # leadership role in the research and dissemina- tion of the core areas of the Institute. The In- NASA Ames Research Center stitute promotes research collaborations with NASA Computational Sciences Division   our sponsors and the research community at Mail Stop - Moffett Field CA  large. Activities supported by the Institute in- -- clude research collaborations and projects, The Computational Sciences Division at visiting scientists, working groups, confer- NASA’s Ames Research Center performs mis- ences and workshops, special programs on sion-driven research in computer science and specific topics and challenge problems, techni- information technology. Areas of focus in- cal reports, and other publications. clude autonomous systems and robotics; intel- ligent system health management; dependable  Booth # software systems; and collaborative and assis- The MIT Press tant systems. Please visit the division’s web site  Cambridge Center at http://compsci.arc.nasa.gov for the latest in- Cambridge, MA  formation on the wide range of research pro-    Tel: - - jects currently underway. Fax: -- The MIT Press publishes books and journals Booth # in artificial intelligence, robotics and related computer sciences. Please visit our booth to Springer-Verlag New York, LLC. receive a % discount on our newest and  Meadowlands Parkway  classic titles including: Baum / What Is Secaucus, NJ --- Thought?, Dorigo / Ant Colony Optimization and Siegwart / Introduction to Autonomous Mobile Robots. Please visit the MIT Press web- site at http://mitpress.mit.edu

21 The Intelligent Systems Demonstrations will be CAMEO is an omnidirectional camera system held in Exhibit Hall  of the San Jose McEnery designed to explore the research issues in- Convention Center and will be open to regis- volved in understanding the activities and in- tered conference attendees during exhibit teractions of people in meeting settings. hours. Continuing advances in AI research are CAMEO monitors the positions of people in a making it possible to develop intelligent sys- meeting, acknowledges specific people it has tems in a wide range of application areas. The been trained to recognize, and attempts to AAAI- Intelligent Systems Demonstrations classify their activities. program showcases state-of-the-art AI imple- Table # mentations and provides AI researchers with an Centibots: Very Large Scale Distributed opportunity to show applications of their re- Robotic Teams search in action. The program is intended to Charlie Ortiz, Kurt Konolige, Regis Vincent, highlight innovative contributions to the sci- Benoit Morisset, Andrew Agno, Michael Eriksen, ence of AI with an emphasis on the benefits to SRI International; Dieter Fox, Benson Limketkai, be gained from developing and using imple- Jonathon Ko, Benjamin Steward, Dirk Schulz, Uni- mented systems in AI research. System builders versity of Washington will be on hand to present their work, and au- We will demonstrate Centibots, our distribut- dience interaction with the systems is encour- ed robotic team, in a coordinated deployment aged as much as possible. Demonstrations are to explore and map an unknown area, to ex- scheduled throughout the AAAI Exhibition. haustively search the area for an object of in- Table # terest, and to guard the found object. Agent-Based Modeling with Social Net- Table # works for Terrorist Recruitment CMRadar: A Personal Assistant Agent for Teresa Ko and Nina Berry, Sandia National Labs Calendar Management Seldon combines concepts from agent-based Pragnesh Jay Modi, Manuela Veloso, Stephen F. modeling and social network analysis to create Smith, Jean Oh, Carnegie Mellon University a computation model of social dynamics for Software agents that automate tasks such as cal- terrorist recruitment. The underlying recruit- endar management promise to increase human ment model is based on a unique hybrid productivity. CMRadar is a calendar manage- agent-based architecture that contains simple ment system in which distributed agents, each agents (individuals) and abstract agents (con- serving a different user, can negotiate to effec- ceptual entities) where the interactions of the tively schedule meetings. Our demo consists of agents are defined through the multiple social a graphical depiction of the distributed negoti- networks that form and dissipate according to ation algorithm that underlies CMRadar. an individual’s actions. Table #  Table # Domain-Independent Reason-Enhanced The Autonomous Sciencecraft Experiment Controller for Task-Oriented Systems –  Onboard the EO- Spacecraft DIRECTOR Daniel Tran, Steve Chien, Rob Sherwood, Rebecca Darsana P. Josyula, Michael L. Anderson and Don Castano, Benjamin Cichy, Ashley Davies, and Perlis, University of Maryland Gregg Rabideau, Jet Propulsion Laboratory, Califor- nia Institute of Technology An agent may control different devices by translating user requests to device instruc- The Autonomous Sciencecraft Experiment tions, tracking the effect of the instructions (ASE), currently flying onboard the Earth Ob- and detecting any perturbations. We discuss serving-1 (EO-) spacecraft, integrates several the architecture for such a perturbation-toler- autonomy software technologies enabling au- ant agent that can interface users with differ- tonomous science analysis and mission plan- ent devices. The agent has been implemented ning. The demonstration will consist of two sec- by modeling its beliefs, desires, intentions, ex- tions: a real-time display of an ASE-commanded pectations and achievements. ground contact from EO-, and a simulation of the full ASE science-response scenario. Table #

Intelligent Demonstrations Systems Engineering Open Multi-Agent Systems as Table # Electronic Institutions CAMEO: Modeling Human Activity in Formal Meeting Situations M. Esteva, D. de la Cruz, B. Rosell, J. Ll. Arcos, J. A., Rodríguez-Aguilar, and G. Cuní, Artificial Intel- Paul E. Rybski and Brett Browning, Carnegie Mel- ligence Research Institute (IIIA), Spain lon University 22 Intelligent Systems Demonstrations Intelligent Intelligent Systems Demonstrations Schedule

Tuesday, July 27 Wednesday, July 28 :  :  Online Semantic Extraction by Backpropagation Neural Responsive Information Architect: A Context-Sensitive Network with Various Syntactic Structure Representa- Multimedia Conversation Framework for Information tions. Table # Seeking Table # Intelligent Agents for Coalition Search and Rescue Task The Secure Wireless Agent Testbed (SWAT) Table #  Support Table # :  :  WordNet::Similarity—Measuring the Relatedness of Centibots: Very Large Scale Distributed Robotic Teams Concepts Table #  Table # PRECISE on ATIS: Semantic Tractability and Experi- :  mental Results Table # iBundler: An Agent-Based Decision Support Service for :   Combinatorial Negotiations Table # SCoT: A Spoken Conversational Tutor Table #  SenseClusters—Finding Clusters that Represent Word SEM-Ether: Semantic Web based Pervasive Computing  Senses Table # Framework—Integrating Web, Devices and People Engineering Open Multi-Agent Systems as Electronic Table #  Institutions Table # Mobile Emergency Triage Support System Table # :  :  Agent-Based Modeling with Social Networks for Terror- Visual Odometry Using Commodity Optical Flow  ist Recruitment Table # Table # Multi-Agent System Development: Design, Runtime, A Robotic Model of Human Reference Resolution  and Analysis Table # Table # CAMEO: Modeling Human Activity in Formal Meeting Agent-Based Modeling with Social Networks for Terror-  Situations Table # ist Recruitment Table # :  :  CMRadar: A Personal Assistant Agent for Calendar CMRadar: A Personal Assistant Agent for Calendar  Management Table # Management Table # Domain-Independent Reason-Enhanced Controller for Domain-Independent Reason-Enhanced Controller for  Task-ORiented systems – DIRECTOR Table # Task-ORiented systems—DIRECTOR Table # The Autonomous Sciencecraft Experiment Onboard the The Autonomous Sciencecraft Experiment Onboard the   EO- Spacecraft Table # EO- Spacecraft Table #  :  :  Mobile Emergency Triage Support System iBundler: An Agent-Based Decision Support Service for  Table # Combinatorial Negotiations Table # Visual Odometry Using Commodity Optical Flow Multi-Agent System Development: Design, Runtime,  Table # and Analysis Table # A Robotic Model of Human Reference Resolution CAMEO: Modeling Human Activity in Formal Meeting  Table # Situations Table # :  :  PRECISE on ATIS: Semantic Tractability and Experi- Centibots: Very Large Scale Distributed Robotic Teams  mental Results Table # Table #  SCoT: A Spoken Conversational Tutor Table # SenseClusters—Finding Clusters that Represent Word SEM-Ether: Semantic Web Based Pervasive Computing Senses Table # Framework – Integrating Web, Devices and People Engineering Open Multi-Agent Systems as Electronic  Table # Institutions Table # :  :  Responsive Information Architect: A Context-Sensitive Online Semantic Extraction by Backpropagation Neural Multimedia Conversation Framework for Information Network with Various Syntactic Structure Representa-  Seeking Table # tions Table #  The Secure Wireless Agent Testbed (SWAT) Table # Intelligent Agents for Coalition Search and Rescue Task WordNet::Similarity—Measuring the Relatedness of Support Table # Concepts Table #

23 We focus on the engineering of open multi- erates in weak-connectivity conditions. The agent systems as electronic institutions. Elec- system facilitates structured data collection, tronic institutions are a formalism to define thus it can be used as an electronic patient’s the rules that structure agent interactions, es- chart, and it supports triage directly at the tablishing what agents are permitted and for- point of care. bidden to do. We present a set of tools that Table # support the specification, analysis and execu- Multi-Agent System Development: tion of institutions, as well as the implementa- Design, Runtime, and Analysis tion of agents. Our methodology allows for a K. S. Barber, J. Ahn, K. Fullam, T. Graser, N. Gu- successive refinement approach to multiagent jral, D. C. Han, D. N. Lam, R. McKay, J. Park, and systems engineering. M. Vanzin, The University of Texas at Austin Table # This demonstration addresses the entire devel- iBundler: An Agent-Based Decision Support opment process for multiagent systems; illus- Service for Combinatorial Negotiations trating tools for the initial design of the agent Andrea Giovannucci and Juan Antonio Rodríguez- system, the capabilities encoded in the individ- Aguilar, IIIA-CSIC Campus UAB, Bellaterra, Spain; ual agents (including novel work in belief revi- Jesús Cerquides, Universitat de Barcelona; A. Reyes sion, agent organizations, and action selection), and F. X. Noria, Intelligent Software Components, S.A. and analysis tools that enhance developer com- We will demonstrate the use of iBundler, an prehension of the behavior of the system. agent-aware decision support service acting as Table # a combinatorial negotiation solver (solving Online Semantic Extraction by Backpropa- the winner determination problem) for both gation Neural Network with Various Syntac- multi-item, multi-unit negotiations and auc- tic Structure Representations tions. The demonstration GUI will show a ne- gotiation involving three providing agents and Heidi H. T. Yeung, City University of Hong Kong one buyer agent. The main goal for the buyer The goal of this research project is to develop agent will be to buy the necessary parts to a hybrid architecture that can extract semantic manufacture  cars. information from English sentences. This is an online intelligent system built on top of high  Table # automation and interaction with Java Server Intelligent Agents for Coalition Search and Page (JSP) and Macromedia Flash MX. By Rescue Task Support training a robust -layer back propagation Austin Tate, Jeff Dalton, Clauirton de Siebra, and neural network with numerical syntactic Stuart Aitken, AIAI, University of Edinburgh, UK; structure representation, the expected themat- Jeffrey M. Bradshaw and Andrzej Uszok, IHMC, ic frame roles are classified faithfully. Pensacola, FL The Coalition Search and Rescue Task Sup- Table # port (CoSAR-TS) demonstration showcases PRECISE on ATIS: Semantic Tractability and intelligent agents and artificial intelligence Experimental Results planning and plan execution systems working Ana-Maria Popescu, Alex Armanasu, Oren Etzioni, over a computational grid in which different David Ko, Alexander Yates, University of Washington policies govern who can do what. The agents The Precise system is a natural language inter- use semantic web services to dynamically dis- face to databases that uses a set of database-in- cover medical information and to find local dependent semantic constraints and a modern rescue resources. statistical parser to help answer user queries. Our AAAI demonstration shows Precise in ac- Table # Mobile Emergency Triage Support System tion on the ATIS travel domain, where it out- performs many domain-dependent systems. Wojtek Michalowski, University of Ottawa; Roman Slowinski and Szymon Wilk, Poznan University of Table # Technology, Poland Responsive Information Architect: A Con- The Mobile Emergency Triage (MET) system text-Sensitive Multimedia Conversation is a clinical decision support system for emer- Framework for Information Seeking

Intelligent Demonstrations Systems gency triage of different types of acute pain Michelle Zhou, Keith Houck, Rosario Uceda-Sosa, presentations. It uses clinical knowledge in Shimei Pan, Min Chen, Vikram Aggarwal, and form of decision rules acquired from data. James Shaw, IBM T. J. Watson Research Center MET interacts with a hospital information RealHunter is an automatically generated system, runs on handheld computers and op- multimedia conversation system that under- 24 stands imprecise natural language input and SeMEther is a pervasive computing frame- Systems Demonstrations Intelligent generates customized context-sensitive graph- work which harnesses the power of Semantic ics and speech output. By guiding users navi- Web and service-oriented architecture, over gating a large and complex information space which effective and intelligent pervasive envi- intelligently based their preferences and un- ronments can be built. This demo aims to derlying data, this real-estate application helps show how simple pervasive services like Bud- users search for residential properties effi- dy service, Music service can be built on this ciently. framework.

Table # Table # A Robotic Model of Human Reference SenseClusters—Finding Clusters that Resolution Represent Word Senses Matthias Scheutz, Virgil Andronache, and Kath- Amruta Purandare, University of Minnesota, Du- leen Eberhard, University of Notre Dame luth and Ted Pedersen, University of Minnesota, Evidence from psychology suggests that hu- Duluth mans resolve reference incrementally in the SenseClusters is free software that clusters presence of constraining visual context with- units of text that are contextually similar. It out the need for the explicit construction of supports the identification of lexical features parse trees. We present an embodied dis- from large corpora, the use of first and second tributed real-time system that demonstrates order context representations, dimensionality various interactions of auditory, visual, and reduction via SVD, and both agglomerative semantic processing components hypothe- and partitional clustering. It also allows for the sized to underlie human processes. evaluation of discovered clusters.

Table # Table # SCoT: a Spoken Conversational Tutor Visual Odometry Using Commodity Optical Karl Schultz, Brady Clark, Heather Pon-Barry, Flow Elizabeth Owen Bratt, and Stanley Peters, Stanford Jason Campbell and Rahul Sukthankar Intel Re- University search and CMU Robotics Institute; Illah Nour- SCoT is a Spoken Conversational Tutor imple- bakhsh, CMU Robotics Institute and NASA Ames mented to investigate the advantages of natu- Research Center ral language in tutoring. SCoT combines a We demonstrate a mobile robot visual odom- generic dialogue architecture, a set of natural etry system based on open source optical-flow language tools, and a set of tutorial rules to software and consumer grade hardware. We conduct a reflective dialogue with a student also illustrate the effect of various vision sys- about a problem they have worked on. tem design parameters on odometric perfor- mance. Table # The Secure Wireless Agent Testbed (SWAT) Table # G. Anderson, A. Burnheimer, V. Cicirello, D. WordNet::Similarity—Measuring the Dorsey, S. Garcia, M. Kam, J. Kopena, K. Malfet- Relatedness of Concepts tone, A. Mroczkowski, G. Naik, M. Peysakhov, Ted Pedersen, University of Minnesota, Duluth, W.Regli, J. Shaffer, E. Sultanik, K. Tsang, L. Ur- Siddharth Patwardhan, University of Utah, and Ja- bano, K. Usbeck, and J. Warren, Drexel University son Michelizzi, University of Minnesota, Duluth We will demonstrate the Secure Wireless WordNet::Similarity is free software that mea- Agent Testbed (SWAT), a unique facility devel- sures the semantic similarity or relatedness oped at Drexel University to study integration, between a pair of concepts. It provides nine networking and information assurance for measures, all of which are based on the lexical next-generation wireless mobile agent sys- database WordNet. These are implemented as tems. SWAT is an implemented system that Perl modules which input two concepts, and fully integrates: () mobile agents, () wireless return a numeric value that represents their ad hoc multi-hop networks, and () security. similarity or relatedness.

Table # SEM-Ether: Semantic Web Based Pervasive Computing Framework—Integrating Web, Devices and People Sushil Puradkar, Sachin Singh, Chintan Patel, Kar- tik Vishwanath, Rahul Gupta, and Yugyung Lee, University of Missouri – Kansas City 25 The Thirteenth Robot Competition and Exhi- The Robot Exhibition bition will be held in Exhibit Hall  of the San Jose Convention Center, and will be open to The mission of the Robot Exhibition is to registered conference attendees during exhibit demonstrate state of the art research in a less hours. The Competition brings together more structured environment than the competition  than teams from universities, colleges, and events. The exhibition gives researchers an op- research laboratories to compete and to portunity to showcase current robotics and demonstrate cutting edge, state of the art re- embodied-AI research that does not fit into search in robotics and artificial intelligence. the other competition events.

The Robot Challenge Workshop

In short, the robot is supposed to attend the The robot events culminate with a workshop conference. A breakdown of the tasks includes: where participants describe the research be- starting at the entrance to the conference cen- hind their entries. ter and finding the registration desk, register- ing for the conference, performing volunteer General Cochairs: duties as required, interacting with conference Bill Smart, Washington University in St. Louis attendees, and finally reporting at a prescribed Sheila Tejada, University of New Orleans time to a conference hall to give a talk. Some Advisor and Past Cochair: robots will be doing a portion of this chal- Bruce Maxwell, Swarthmore College lenge. Challenge Chair: Ashley Stroupe, Jet Propulsion Laboratory n and Exhibition The Rescue Robot Competition Exhibition Chair: The objective of this contest is to give partici- Magda Bugajska, Naval Research Lab pants the opportunity to work in a domain of critical practical importance. Robots must en- Open Interaction Event Chair: ter a simulated fallen structure, find human Bruce A. Maxwell, Swarthmore College victims, and direct human rescuers to the vic- Robot Rescue Cochairs: tims. The rules for this event provide for fully Jenn Casper, American Standard Robotics autonomous robots, teleoperated robots, and Adam Jacoff, NIST sliding autonomy. The event is being devel- oped in close coordination with experienced Mobile Robot Workshop Chair: rescue professionals. This event will use the Frederick Crabbe, U.S. Naval Academy joint AAAI/RoboCup Robot Rescue rules.

Open Interaction Task

In addition to the above two events, there will be a third event involving human-robot inter- action. This event will take the place of the Robot Host event in past years and will prob- ably involve interacting with conference atten- dees to achieve a particular task in an unstruc- tured environment. Robot Competitio

26 Robot Teams Robot

Robot Teams

Team: Academic Autonomy Team: MU Mites Swarthmore College University of Missouri-Columbia Contact: Bruce Maxwell Contact: Marge Skubic Event: Rescue Robot Competition Events: Open Interaction Task & Exhibition Team: ALCOR/DIA Team: ND Universitá di Roma University of Notre Dame Contact: Andrea Carbone Contact: Matthias Scheutz Event: Rescue Robot Competition Events: Open Interaction Task & Exhibition Team: B-Dog Team: Scrambled Eggs University of New Orleans Institute for Educational Contact: Sheila Tejada Advancement/PARC Event: Open Interaction Task Contact: Mark Yim Event: Rescue Robot Competition Team: Blue Swarm 3 Utah State University Team: Stony Brook Robot Design Contact: Dan Stormont Team Event: Rescue Robot Competition Stony Brook University Team: CMBalance Contact: Diana David Events: Open Interaction Task & Exhibition Carnegie Mellon University Contact: Paul E. Rybski Team: Team H.E.R. Event: Exhibition Human Emulation Robotics, LLC Team: GRACE and GEORGE Contact: David Hanson Events: Challenge & Exhibition CMU, NRL, Swarthmore Contact: Alan Schultz Team: The Scarabs Event: Open Interaction Task Contact: Michael Randall Team: INEEL Event: Rescue Robot Competition INEEL Team: UNO Robotics Team Contact: David Bruemmer University of New Orleans Event: Rescue Robot Competition Contact: Sheila Tejada Team: Keystone Rescue Event: Rescue Robot Competition University of Manitoba Team: Washington University Contact: Jacky Baltes Washington University in St. Louis Event: Rescue Robot Competition Contact: Bill Smart Team: MITRE Event(s): Challenge & Exhibition MITRE Contact: Dave Smith Event: Rescue Robot Competition

27 National Conference on Educational Robotics ❐ Botball Tournament ❐ Collegiate Botball Challenge ❐ Robotics Showcase

Each year teachers and students from middle school, high school and college gather with robotics professionals from across the country for the annual Na- tional Conference on Educational Robotics (NCER). Hosted by KISS Institute for Practical Robotics, this year’s National Conference on Educational Robotics is held in conjunction with AAAI-.All registered AAAI- partic- ipants are invited to attend the National Botball Tournament, the Collegiate Botball Challenge, the Robotics Showcase and all breakout sessions for no ad- ditional cost!

A Highlight of the National Conference on Educational Robotics includes the National Botball Tournament for middle- and high-school students, as well as the Collegiate Botball Challenge. Any middle- or high-school team who com- peted in one of the thirteen regional Botball programs is eligible to participate in the National Botball Tournament.

In addition to the Collegiate Botball Challenge and the National Botball Tour- nament, the National Conference on Educational Robotics features student and teacher presentations during breakout sessions designed to cover topics from curriculum integration to technical aspects of robotics. An evening robotics showcase will highlight demos and posters. Students and teachers are also given the opportunity to hear about current robotics research conducted by nationally recognized speakers from organizations such as NASA, the Naval Research Laboratory, and various universities.

Botball is an educational outreach program designed to engage students in learning the practical applications of science, technology, engineering and math through robot building and programming. In Botball, teams of students create autonomous mobile robots to play in a non-destructive tournament. You can find out more about Botball at the Botball website: www.botball.org. CER / National Botball Tournament CER / National N

28 Registration Information Registration Registration

Conference registration is located on the exhibit level of the San Jose Convention Center, beginning Sunday, July .Registration hours are: Sunday, July  :  – :  Monday, July  :  – :  Tuesday, July  :  – :  Wednesday, July  :  – :  Thursday, July  :  – : 

Only checks drawn on U.S. banks, VISA, MasterCard, American Express, government pur- chase orders, traveler’s checks, and U.S. currency will be accepted. We cannot accept for- eign currency or checks drawn on foreign banks. Registration Fees The AAAI- / IAAI- technical program registration fee includes admission to the tech- nical sessions, the Exhibition, the Student Abstract and Poster Session, the Workshop Pro- gram (by invitation only), the Opening Reception, and the AAAI- / IAAI- conference proceedings and CD. Students must present proof of full-time student status to qualify for the student rate. Onsite technical program fees are as follows: Technical Registration Fees Regular Member $ Regular Nonmember $ Student Member $ Student Nonmember $

AAAI Platinum Fees (Includes one year new or renewal membership in AAAI) Regular U.S./Canada $ Regular International $ Student U.S./Canada $ Student International $

Tutorial Forum Includes admittance to up to four consecutive tutorials and the accompanying Tutorial Fo- rum Notes. Regular $ Student $

Workshop Program Workshop registration is limited to those active participants determined by the organizer prior to the conference. All workshop participants must register for the AAAI- technical program.

Exhibition Admission to the exhibition hall programs is included in the AAAI- / IAAI- technical registration fee. For individuals interested in admittance to the exhibit hall only, an exhibits only registration is available. Exhibit hall programs include vendor exhibits, the Intelligent Systems Demonstrations, the Botball Tournament, and the Robot Competition and Exhi- bition.

All passes are good for Tuesday and Wednesday, July – Exhibits Only Adult $. Exhibits Only Students (-)$. Children (under )Free (Children must be accompanied by an adult conference registrant)

29 Admission Career Information

Each conference attendee will receive a name A bulletin board for job opportunities in the badge upon registration. This badge is re- artificial intelligence industry will be made quired for admittance to the technical, tutori- available in the registration area, on the exhib- al, exhibit, IAAI, and workshop programs. it level of the San Jose Convention Center. At- Workshop attendees will also be checked off a tendees are welcome to post job descriptions master registration list at individual rooms. of openings at their company or institution. Tutorial attendees must present syllabi tickets to receive syllabi volumes, and attendance tickets for admittance to the tutorial rooms. Child Care Service Smoking, drinking, and eating are not allowed in any of the technical, tutorial, workshop, or For information about childcare services, IAAI sessions. please contact Sitter’s Unlimited at -- .(This information is provided for your formation convenience and does not represent an en- Baggage Holding dorsement of this agency by AAAI. Responsi- bility for all childcare arrangements must be There is no baggage holding area at the San assumed by the parents.) Jose Convention Center. Please check your luggage with the bellman at your hotel after you have checked out. Neither AAAI, the San Coffee Breaks Jose Convention Center, the San Jose Mar- riott, nor San Jose State University accept lia- Coffee will be served in the meeting room foy- bility for the loss or theft of any suitcase, brief- er on the exhibit level of the San Jose McEnery case, or other personal belongings brought to Convention Center, Sunday, July  and Mon- the site of AAAI- / IAAI-. day, July , at : – :  and : – : .Coffee will also be served in the Willow Glen foyer of the San Jose Marriott during Banking these times. Coffee will be served only in the San Jose General In General An ATM machine is located on the Ballroom Ballroom Foyer of the San Jose Marriott on Concourse on the exhibit level of the San Jose Tuesday, July , :  – : .Coffee McEnery Convention Center. will be served in the meeting room foyer on Wells Fargo the exhibit level of the San Jose Convention        Park Center Plaza, San Jose CA  Center Tuesday, July , : – : , and     -- Wednesday and Thursday, July – , : – :  and : – : . Bank of America Corner Market Kitchen, located on the first  S. Market Street, San Jose CA  floor of the San Jose Marriott is open :  -- – :  daily. A Starbucks is also located in the San Jose McEnery Convention Center near Business Centers the Almaden entrance.

The following business centers are available in Handicapped Facilities the area: Visitor Information & Business Center The San Jose Convention Center and the San Main lobby, San Carlos entrance Jose Marriott are both equipped with handi- San Jose McEnery Convention Center capped facilities. San Jose Marriott Business Center Second floor Open  hours based on need Housing Kinko’s For information regarding hotel reservations,  East San Carlos Street please contact hotels directly. For student -- housing, please contact San Jose State Univer- sity at --. 30 Information/Messages/Robot Press General In Receptionist All members of the media are requested to  Please visit the information/message desk in register in the Press Room, Meeting Room B . the registration area. This year, you might be Press badges will only be issued to individuals assisted by either GRACE or GEORGE, the with approved credentials. The Press Room joint entries of Carnegie Mellon University, will be open during the following hours. the Naval Research Lab, and Swarthmore Col- Monday, July  :  – :  lege in the Open Interaction Task of the Robot Tuesday, July  :  – :         Competition. A volunteer will also be on hand Wednesday, July : – :        (in onsite registration) to receive and post Thursday, July : – : messages. The telephone number for leaving On Sunday, July ,press members should reg- messages only is --.Paging atten- ister in onsite registration. dees is not possible.

Printed Materials formation Internet Room & Access Display tables for the distribution of promo- AAAI will provide internet access in Meeting tional and informational materials of interest Room B of the San Jose Convention Center. to conference attendees will be located in the The internet room will be open Sunday, July  registration area on the exhibit level of the San – Thursday, July  during registration hours. Jose Convention Center. As a courtesy, please limit your access time to  minutes if others are waiting to use the ser- vice.Wireless access is also available in the ex- Proceedings / CD hibit hall. Each technical registrant will receive two tick- ets with the registration materials for one copy List of Attendees of the conference proceedings and one copy of the conference CD. Tickets can be redeemed at A list of preregistered attendees of the confer- the proceedings distribution center, located in ence will be available for review at the AAAI Meeting Room D on the exhibit level of the Desk in the registration area. Attendee lists San Jose Convention Center during registra- will not be distributed. tion hours. All tickets must be redeemed on- site by Thursday, July  at : . AAAI cannot mail proceedings or CDs to registrants Parking after the conference.

Parking is available at the San Jose McEnery Convention Center for $. per day, Mon- Shipping day-Friday, :  – : , and $. per day, Saturday and Sunday, :  – : . United Parcel Service Access to the garage is at  Almaden Boule-  Post Street vard between San Carlos Street and Balbach -- Street, or on Market Street between San Carlos United States Postal Service Street and Viola Avenue. Extensive parking is 200 South Third Street, available throughout downtown San Jose. For San Jose, CA  more information on parking , public trans- See also Visitor information & Business Cen- portation, and the DASH shuttle, visit www.sj- ter. downtownparking.com.

31 Transportation Visitor Information & Business Center Taxi Rainbow Cab: -- Located in the San Jose McEnery Convention Yellow Checker Cab Company: -- Center, the Visitor Information & Business Center (VIBC) offers a full array of business Bus services, including computer/internet access, Greyhound Bus—For information on fares faxing, and mailing options. In addition, visi- and scheduling, call --. tors and delegates will find San Jose souvenirs, reservation (restaurants and lodging) assis- City Transit System tance, calendar of events, and general infor- The Light Rail (VTA) runs seven days a week, mation. Open daily, Monday through Friday,    -hours a day. Trains operate every - :  – : , and Saturday and Sunday,  minutes weekdays, every minutes Saturdays :  – : . and Sundays, and every  minutes in the formation evening. For a complete listing of schedules, fares, and routes, please see www.vta.org. Volunteer Station Free Shuttles The volunteer station will be located in the VTA/SJC Airport Flyer registration area on the exhibit level of the San Service to San Jose International Airport, San- Jose McEnery Convention Center. All volun- ta Clara Caltrain station and Metro/Airport teers are required to sign in prior to their shift, Light Rail Station. Shuttles run weekdays ev- and sign out when it ends. ery  minutes from :  – : ,every  minutes until : , and every  min- utes until midnight. Weekend service is every Disclaimer  minutes. Downtown Area Shuttle (DASH) In offering the San Jose Marriott, San Jose Service connects Caltrain station with Light Rail McEnery Convention Center, San Jose State line, transit mall, San Jose McEnery Convention University, and all other service providers

General In General Center and San Jose State University. Shuttle (hereinafter referred to as “Supplier(s)” for the runs every  minutes, :  – : . National Conference on Artificial Intelligence Train and the Innovative Applications Conference), AAAI acts only in the capacity of agent for the BART Suppliers which are the providers of the ser- Bay Area Rapid Transit offers service within vice.Because AAAI has no control over the the Bay Area. For information call -- personnel, equipment or operations or  or visit www.bart.gov providers of accommodations or other servi- CalTrain ces included as part of the AAAI- / IAAI- CalTrain offers service between San Jose and program, AAAI assumes no responsibility San Francisco. For departure schedules, map for and will not be liable for any personal de- and fares, call -- or visit www.cal- lay, inconveniences or other damage suffered train.org by conference participants which may arise by reason of () any wrongful or negligent acts or omissions on the part of any Supplier or its Tutorial Forum Syllabi employees, () any defect in or failure of any vehicle, equipment or instrumentality owned,  Extra copies of AAAI- tutorial syllabi vol- operated or otherwise used by any Supplier, or ume will be available for purchase in AAAI on- () any wrongful or negligent acts or omis- site registration in the San Jose Convention sions on the part of any other party not under  Center, beginning Tuesday, July .Supplies are the control, direct or otherwise, of AAAI. limited. The cost is $. per volume (in- cludes all tutorials). Preregistration tutorial syllabi tickets may be redeemed at the Proceed- ings distribution center in Meeting Room D.

32