Keynote ACIS/AAI2014 conference

• 1. Aug 31-Sep 4, 2014@ACIS AAI 2014,Kitakyushu International Conference CenterFacilitatingLarge-scale Collective DiscussionTakayuki ItoNagoya Institute of Technology,[email protected]

2. Takayuki ItoProfessor,Nagoya Institute of TechnologyE-mail : [email protected]://www.itolab.nitech.ac.jp/~ito/2000 Dr. of Engineering, Nagoya Institute of Technology1999 Research fellow, Japan Society for the Promotion of Science (JSPS).2000 Visiting Researcher, USC/ISI.2001 Assoc. Professor, Japan Advanced Insti. of Sci. & Tech. (JAIST).2003 Assoc. Professor, Dept. of CSE, Nagoya Institute of Technology.2005 Visiting Scholar, Computer Science, Harvard University.2005 Visiting Researcher, MIT Sloan School of Management.2006 Assoc. Prof., School of Techno-Business Admin., Nagoya Institute of Technology.2008 Visiting Scientist, Center for Collective Intelligence, MIT Sloan School ofManagement.2010 JST PREST researcher (super challenging type)2011 Japanese Cabinet Offices NEXT fund Principal Investigator2013 AAMAS 2013 Program Chair2014 Professor, School of Techno-Business Admin., Nagoya Institute of Technology. 3. [email protected] :Computer Science, Artificial Intelligence, Multi-Agent Systems,Collective Intelligence, Automated Negotiations, Auction Theory,Mechanism Design, Smart City, Smart Grid, etc.Main achievements: 2014 JSSST Fundamental Research Award 2013 JSPS Award ( Prizes for Science and Technology, The Commendation for Science andTechnology by the Minister of Education, Culture, Sports, Science and Technology,2013. The NEXT Funding Program from the Japanese Cabinet Office The Young Scientists' Prize, The Commendation for Science and Technology bythe Minister of Education, Culture, Sports, Science and Technology, 2007. Nagao Special Researcher Award, IPSJ 2007 Best Paper Award, The Fifth International Joint Conference on AutonomousAgents and Multi-Agent Systems (AAMAS2006, 1/553), 2006. 2005 Best Paper Award, Japan Society for Softoware Science and Technoglogy. 2004 IPA Exploratory Software Creation Project, Super Creator Award. 4. Background and GoalNeed for a large-scale discussion and gathering opinionsCity planning DesignDirect democratic discussion [email protected] focused very much.New products was developed usingmuch user voice.Some fields need for a large-scale consensusWe develop an open web-based forum systemcalled COLLAGREE (COLLective, COLLaborateive, AGREEment) 5. Background and GoalSome management-level people are checking the newlycreated contents in open-source projects.From viewpoint of collective intelligence, some level ofmanagement or control are required. Some management-level people avoid flamingand check some contents.Facilitator: who adequately coordinate, lead, integrate,classify, and summarize [email protected] workshopsWikipedia and Linux projectsReal workshop in collaborative groupCity planning divisionWorkshop with about 15 personsDesign divisionBrain storming with 4 persons The real-world workshops are coordinated by a facilitator.We introduce facilitators and support functionsinto an large-scale online discussion. 6. Supporting FacilitatorsIn our system, some human facilitators coordinatethe large-scale discussion.Facilitator lead discussion process. maintain human relationsCurrent Web-based discussion Large-scale discussionFlaming Conflict between usersGroupthink The group results in an irrational.Monitoring cost They must check enormous opinions.Scale out problem Previous methods is for [email protected] discussion City planning DesignWe have to support facilitators so that they can manage the large-scalediscussion.We develop the support functionsto support facilitators for coordinating discussion. 7. Related WorkClimate CoLab is one of the mostfamous web-based collectiveintelligence projects. The goal ofClimate CoLab is to harness thecollective intelligence of thousands ofpeople from all around the world toaddress global climate change.This is a crowdsourcing platformwhere citizens work with experts andeach other to create, analyze, andselect detailed proposals for what todo about climate change.They do not have facilitators.Joshua Introne, Robert Laubacher, Gary Olson, and Thomas Malone. 2011. The Climate CoLab: Large scalemodel-based collaborative planning. In Proceedings of International Conference on CollaborationTechnologies and Systems (CTS 2011)Thomas W. Malone, Robert Laubacher, Josh Introne, Mark Klein, Hal Abelson, John Sterman, and Gary Olson.2009. The Climate Collaboratorium: Project Overview. CCI Working Paper 2009-003 (Sept. 2009)[email protected] 8. Related WorkDeliveratorium is another projectwhere people submit ideas byfollowing the argumentation map.The argumentation map is a kind ofthe discussion structure.Deliveratoriummakes it possible toshow the entire relations amongideas and opinions very clearly usingthe structure.Ali Gurkan, Luca Iandoli, Mark Klein, and Giuseppe Zollo. 2010. Mediating debatethrough on-line large-scale argumentation: Evidence from the field. InformationSciences 180, 19 (2010), 3686 [email protected] 9. COLLAGREE User InterfaceTopPeople can submit their opinions as a chunk of texts as like theordinary forums and electronic bulletin [email protected] categoriesWe can discuss some theme at the same time. 10. User InterfaceArgumentationWe develop several supporting functions for facilitators.Support functions Agreement or disagreement analysisfor a comment is shown. Keywords highlighted. Facilitators canunderstand what kind of keywords arefocused and important. Issue tags by which participants canadd a tag each opinion and commentso that they can search it afterwards. Facilitation tab from which a facilitatorcan input his/her instructions tor theparticipants. Reminders for participants. It sendsemails as well as reminders when somerelated events [email protected] developed several supporting functions for facilitators. 11. Demo MovieUser-Interface1. Top / Theme2. DiscussionSupport function1. Agreement or disagreement2. Facilitation tab3. Keywords highlighted4. Issue tag5. [email protected] 12. System ArchitectureLight weight architectureCoffeeScriptJavaScriptSpine.js Ruby on [email protected] analyticsJudgement ofagree / disagreeWeb server(Nginx)Opinion and CommentdataKeyword dataDatabase(mongoDB)BrawserData managementOpinion dataUser dataKeyword dataKeywordAgree / DisagreeCommentPostUbuntuComputer environmentsOS Ubuntu12.04ComputerLanguageRuby, JavaScriptCoffeeScriptDatabase mongoDBWeb server NgnixFramework Ruby on Rails3, Spine.js 13. A Large-scale Experiment in Nagoya CityWe deployed our system for internet-based town meeting inNagoya city, as a city project led by Nagoya city mayorAn internet-based town meeting about Town based on theNagoya Next Generation Total City Planning 2014-2018(4 discussion categories)Press conference [email protected] 19, 2013 ~Dec 3, 2013 [15 days]Nov18, Press conference with the mayor of Nagoya Nov 13, 30 Chunichi Shinbun (Japanese newspaper)The civilian participants 264 peopleThe professional facilitators 9 peopleQuestionnaires online Operability and effect The importance of facilitatorsWe evaluated COLLAGREE by the large-scale experiment. 14. [email protected] 15. [email protected] 16. A Large-scale Experiment in Nagoya CityWe announced in our experiment on Nagoya city website.Explanation of COLLAGREE and Introductionof an internet-based town meetingWe evaluated COLLAGREE by the large-scale [email protected] Nagoya city websiteA link to COLLAGREE 17. [email protected] 18. Awarding Users A Prize from MayorNagoya city mayor awarded a prize to the good users3,072 visits1,151 opinions 18,446 viewAwarding 3 good participants10s, maleCOLLAGREE succeeded gathering good [email protected] 29 Chunichi Shinbun(Japanese newspaper)30s, female40s, femaleCity planning which matched featureof NagoyaArt and barrier-freeTraining for region disaster preventionMayor There are predominantly many youngpeople in the site, and opinions different fromthe town meeting are useful. 19. [email protected] 20. We compared our system with real town meeting.Real town meeting(Nagoya) COLLAGREECOLLAGREEAges of Participants1020304010-19 20-29 30-39 40-49 5500-59 6O0ver 60 No answ erArgument by the participant of the wide generation unlike real-worldtown [email protected]%30-3915%50-5917%40-4927%Over 6057%No Answer10-191%Over 607%12areas678 peopleParticipants264 people 21. Questionnaires about Facilitation [User]We asked users about the importance of facilitation 83%97%Q.1 Was facilitation necessary?Agree Strongly AgreeQ.2 Could facilitator harness discussion?Agree Strongly Agree0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%CommentsAgree strongly Agree Indifferent Disagree Disagree strongly Facilitators and Facilitations are essential. In specially, there is no direction in the discussionParticipants understood the importance of facilitatorsto coordinate discussion and gather many [email protected] many and unspecified persons. It is easy to post opinions because facilitator comments have much humanity. It is better that facilitator summarize discussion some times.0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 22. Questionnaires about Facilitation [Facilitator]We asked faciliators about the importance of facilitation Agree Strongly AgreeAgree Strongly0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%Agree strongly Agree Indifferent Disagree Disagree strongly Facilitators understood the importance of facilitatorsto coordinate discussion and gather many [email protected]%88%Q.1 Was facilitation necessary?Q.2 Could facilitator harness discussion?Agree0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%Comments I thought that facilitators was necessary not to keep on giving an opinion. Facilitator is needed to keep the discussion safely. It is difficult to understand opinions of Silent majority.0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 23. Questionnaires about Facilitator support functionsWe ask facilitators about the support functionsQ. Was each support functions useful for discussion? 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Agree / Disagree Activity Agree strongly Agree Indifferent Disagree Disagree strongly We confirm usefulness of the support about the discussionanalysis, some difficulty to manage the large-scale [email protected] tabIssue tagKeyword highlightReminder 24. [email protected] for large-scale discussion and gathering opinions.We develop the functions to support the facilitatorcoordinating discussion.Analysis real-world workshopDevelopment of systemLarge-scale experimentAn internet-based town meeting in Nagoya cityabout Town based on the Nagoya NextGeneration Total City Planning 2014-2018.Participants (80%) and facilitators (75%)realized the importance of facilitator.We confirmed usefulness offacilitatorCOLLAGREE succeeded coordinating a large-scalediscussion(266 people) and gatheringmany opinions compared with the real-worldtown meeting.Usefulness of facilitatorFuture WorkMany directions that include creating more effectivefacilitation supports. 25. Ongoing work & Future direction1. Online discussion for the design of new products ina company Furniture design with Toyo-set inc.2. Large scale online & offline simultaneous discussion Project with Aichi prefecture3. Automated facilitation & Automated [email protected] 26. Future [email protected] 27. Automated facilitation & Automated agreementDone [email protected] facilitatorAgentDiscussionAgent supports human facilitatorHuman facilitatorAgents & human participateAgents discuss on behalf of usersAgent facilitatorDiscussionAutomated facilitation by agentAgentDiscussionAutomated agreementAgentNextUser UserNext 28. Special ThanksProf. Eizo Hideshima (NIT) :For arranging research collaborationProf. Takanori Ito (NIT) :For designing and planningStudents : For implementing and actual workMr. Kawamura (Mayor of Nagoya) : For announcement and awardsNagoya city officers : For experiment planningJapan Facilitation Association : For facilitationParticipants : For their valuable opinions and participation [email protected] family : For patient supportsDr. Mark Klein (MIT) :For wonderful comments