Jordi Bieger · Ben Goertzel Alexey Potapov (Eds.) 5 Artificial 0 2 9 I General Intelligence A N L 8th International Conference, AGI 2015 Berlin, Germany, July 22–25, 2015 Proceedings 123 fi Lecture Notes in Arti cial Intelligence 9205 Subseries of Lecture Notes in Computer Science LNAI Series Editors Randy Goebel University of Alberta, Edmonton, Canada Yuzuru Tanaka Hokkaido University, Sapporo, Japan Wolfgang Wahlster DFKI and Saarland University, Saarbrücken, Germany LNAI Founding Series Editor Joerg Siekmann DFKI and Saarland University, Saarbrücken, Germany More information about this series at http://www.springer.com/series/1244 Jordi Bieger Ben Goertzel (cid:129) Alexey Potapov (Eds.) fi Arti cial General Intelligence 8th International Conference, AGI 2015 – Berlin, Germany, July 22 25, 2015 Proceedings 123 Editors Jordi Bieger Alexey Potapov ReykjavikUniversity Saint PetersburgState University Reykjavik of Information Technologies, Iceland MechanicsandOptics St.Petersburg BenGoertzel Russia Hong KongPolytechnic University Hong KongSAR ISSN 0302-9743 ISSN 1611-3349 (electronic) Lecture Notesin Artificial Intelligence ISBN 978-3-319-21364-4 ISBN978-3-319-21365-1 (eBook) DOI 10.1007/978-3-319-21365-1 LibraryofCongressControlNumber:2015943355 LNCSSublibrary:SL7–ArtificialIntelligence SpringerChamHeidelbergNewYorkDordrechtLondon ©SpringerInternationalPublishingSwitzerland2015 Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpartofthe material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilarmethodologynow knownorhereafterdeveloped. Theuseofgeneraldescriptivenames,registerednames,trademarks,servicemarks,etc.inthispublication doesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevant protectivelawsandregulationsandthereforefreeforgeneraluse. Thepublisher,theauthorsandtheeditorsaresafetoassumethattheadviceandinformationinthisbookare believedtobetrueandaccurateatthedateofpublication.Neitherthepublishernortheauthorsortheeditors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissionsthatmayhavebeenmade. Printedonacid-freepaper SpringerInternationalPublishingAGSwitzerlandispartofSpringerScience+BusinessMedia (www.springer.com) Preface Almost exactly 60 years ago, in the summer of 1955, John McCarthy coined the term “artificial intelligence” (AI) to refer to “the science and engineering of making intelli- gentmachines”inaproposalforasummerresearchprojectatDarthmouthCollege.The subsequentDarthmouthConferencesof1956areoftencreditedwiththecreationofthe field of AI. But as the problem proved much more difficult than anticipated, disillu- sionmentsetin.Thegoalofcreatingmachinesthatcouldthinkatalevelcomparableto humanswassetasidebymany,infavorofthecreationof“smart”applicationsthatwere highly successful in specialized domains. Since then “AI” and “narrow AI” have become almost synonymous and the development of systems showing more general intelligencein a wide variety ofdomainswas seenas unattainable. Butafterhavingbeenlargelyignoredformanydecades,thelasttenyearshaveseen a small resurgence in the pursuit of what we now call artificial general intelligence (AGI). While the contributions of narrow AI to science and society are undeniable, manyresearcherswerefrustratedwiththelackofprogresstowardthelargergoalofAI. Armed with novel technology and ideas, a new optimism has taken hold of the community. Creating thinking machines may be a daunting task, but many people todaybelievethatitisnotimpossible,andthatwecantakestepstowardthatgoalifwe keep our eye on the ball. TheAGIconferenceseries,organizedbytheAGISociety,hasbeenthemainvenue forbringingtogetherresearchersinthisre-emergingfield.Forthepasteightyearsithas facilitated the exchange of knowledge and ideas by providing an accessible platform for communication and collaboration. This volume contains the research papers accepted for presentation at the Eighth Conference on Artificial General Intelligence (AGI-15),held duringJuly 22–25, 2015,inBerlin. Atotalof72research papers were submitted to the conference, and after desk rejecting 14 (19 %), each paper was reviewed by at least two, and on average 2.93, Program Committee members. We accepted23papersfororalpresentation(32%)aswellas19posters(26%),ofwhich one was withdrawn. Inaddition to these contributed talks, theconference featured JürgenSchmidhuber, directoroftheSwissAIlabIDSIAinLugano,andFrankWood,associateprofessorat the University of Oxford, who gave invited keynote speeches on “The Deep Learning RNNaissance” and probabilistic programming with the Anglican language. José Hernández-Orallo, professor at the Polytechnic University of Valencia, gave a tutorial ontheevaluationofintelligentsystems.AnothertutorialwasgivenbyAlexeyPotapov, professorattheITMOUniversityandSt.PetersburgStateUniversity,ontheminimum descriptionlengthprinciple.Athirdtutorial,givenbyNilGeisweiler,CosmoHarrigan and Ben Goertzel, described how tocombine program learning and probabilistic logic inOpenCog.MartinBalekandDusanFedorcakpresentedavisualeditorfordesigning the architectureof artificial brains.Finally, theconference also featured workshops on Socioeconomic Implications of AGI and on Synthetic Cognitive Development and VI Preface Integrated-Distributed Agency, organized in collaboration with the Global Brain Institute at the Free University of Brussels. Finally, many thanks are due to those who helped organize the conference, includingJanKlauck,JoschaBach,andmanyothers;andtotheconference’ssponsors including Kurzweil AI, Keen Software House, and the OpenCog Foundation. May 2015 Jordi Bieger Ben Goertzel Alexey Potapov Organization Organizing Committee Ben Goertzel AGI Society, USA (Conference Chair) Joscha Bach MIT and Harvard University, USA Matthew Iklé Adams State University, USA Jan Klauck Austrian Space Forum, Austria (Local Chair) Program Chairs Jordi Bieger Reykjavik University, Iceland Alexey Potapov AIDEUS and ITMO University, Russia Program Committee Bo An Nanyang Technological University, China Itamar Arel University of Tennessee, USA Joscha Bach MIT and Harvard University, USA Tarek Besold University of Osnabrück, Germany Cristiano Castelfranchi Institute of Cognitive Sciences and Technologies, Italy Antonio Chella University of Palermo, Italy Blerim Emruli Luleå University of Technology, Sweden Stan Franklin University of Memphis, USA Deon Garrett Icelandic Institute for Intelligent Machines, Iceland Nil Geisweiller Novamente LLC, USA Helmar Gust University of Osnabrück, Germany José Hernández-Orallo Polytechnic University of Valencia, Spain Bill Hibbard University of Wisconsin–Madison, USA Marcus Hutter Australian National University, Australia Matthew Iklé Adams State University, USA Benjamin Johnston University of Sydney, Australia Cliff Joslyn Pacific Northwest National Laboratory, USA Randal Koene Carboncopies.org, USA Kai-Uwe Kühnberger University of Osnabrück, Germany Shane Legg Google Inc., USA Moshe Looks Google Inc., USA Maricarmen Martinez University of Los Andes, Colombia Amedeo Napoli LORIA Nancy, France Eric Nivel Icelandic Institute for Intelligent Machines, Iceland VIII Organization Laurent Orseau Google Inc., USA Guenter Palm Ulm University, Germany Maxim Peterson ITMO University, Russia Paul Rosenbloom University of Southern California, USA Rafal Rzepka Hokkaido University, Japan Samer Schaat Vienna Technical University, Austria Ute Schmid University of Bamberg, Germany Jürgen Schmidhuber IDSIA, Switzerland Javier Snaider Google Inc., USA Bas Steunebrink IDSIA, Switzerland Claes Strannegård University of Gothenburg, Sweden Kristinn Thórisson Reykjavik University, Iceland Julian Togelius IT University of Copenhagen, Denmark Mario Verdicchio University of Bergamo, Italy Pei Wang Temple University, USA Roman Yampolskiy University of Louisville, USA Byoung-Tak Zhang Seoul National University, South Korea Additional Reviewers Mayank Daswani Australian National University, Australia Tom Everitt Stockholm University, Sweden Matthias Jakubec Vienna Technical University, Austria Jan Leike Australian National University, Australia Lydia Chaido Siafara Vienna Technical University, Austria Qiong Wu Nanyang Technological University, China Steering Committee Ben Goertzel AGI Society, USA (Chair) Marcus Hutter Australian National University, Australia Contents Papers Presented Orally Modeling Motivation in MicroPsi 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Joscha Bach Genetic Programming on Program Traces as an Inference Engine for Probabilistic Languages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Vita Batishcheva and Alexey Potapov Scene Based Reasoning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Frank Bergmann and Brian Fenton Anchoring Knowledge in Interaction: Towards a Harmonic Subsymbolic/Symbolic Framework and Architecture of Computational Cognition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Tarek R. Besold, Kai-Uwe Kühnberger, Artur d’Avila Garcez, Alessandro Saffiotti, Martin H. Fischer, and Alan Bundy Safe Baby AGI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Jordi Bieger, Kristinn R. Thórisson, and Pei Wang Observation, Communication and Intelligence in Agent-Based Systems . . . . 50 Nader Chmait, David L. Dowe, David G. Green, and Yuan-Fang Li Reflective Variants of Solomonoff Induction and AIXI. . . . . . . . . . . . . . . . 60 Benja Fallenstein, Nate Soares, and Jessica Taylor Are There Deep Reasons Underlying the Pathologies of Today’s Deep Learning Algorithms? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 Ben Goertzel Speculative Scientific Inference via Synergetic Combination of Probabilistic Logic and Evolutionary Pattern Recognition . . . . . . . . . . . . . . . . . . . . . . . 80 Ben Goertzel, Nil Geisweiller, Eddie Monroe, Mike Duncan, Selamawit Yilma, Meseret Dastaw, Misgana Bayetta, Amen Belayneh, Matthew Ikle’, and Gino Yu Stochastic Tasks: Difficulty and Levin Search . . . . . . . . . . . . . . . . . . . . . . 90 José Hernández-Orallo Instrumental Properties of Social Testbeds. . . . . . . . . . . . . . . . . . . . . . . . . 101 Javier Insa-Cabrera and José Hernández-Orallo