Semantic Technologies for Big Data   THE  8TH  INTERNATIONAL  CONFERENCE   ON  SEMANTIC  TECHNOLOGIES   NOVEMBER  12-­‐15,  2013   Mason  Inn  Conference  Center   George  Mason  University   Fairfax,  Virginia  Campus    Conference Proceedings Kathryn  B.  Laskey   Ian  Emmons   Paulo  C.  G.  Costa   (Eds.)             Preface         The  8th  International  Conference  on  Semantic  Technologies  for  Intelligence,  Defense,  and   Security   (STIDS   2013)   provides   a   forum   for   academia,   government   and   industry   to   share   the   latest   research   on   semantic   technology   for   defense,   intelligence   and   security   applications.     Semantic   technology   is   a   fundamental   enabler   to   achieve   greater   flexibility,   precision,   timeliness  and  automation  of  analysis  and  response  to  rapidly  evolving  threats.       The  STIDS  2013  theme  is  Semantic  Technologies  for  Big  Data.     Topics  of  general  interest  for  STIDS  include:   • Creating  an  interoperable  suite  of  public-­‐domain  ontologies  relevant  to   intelligence  analysis  covering  diverse  areas   • Ontologies  and  reasoning  under  conditions  of  uncertainty   • Semantic  technology  and  ontological  issues  related  to:   o Source  credibility  and  evidential  pedigree     o Use  of  sensing  devices  including  security,  e.g.  global   infrastructure  grid  (GIG),  images  and  intelligence  collection  in   general     • Usability  issues  relating  to  semantic  technology   • Best  practices  in  ontological  engineering     Fairfax,  VA   November  2013     Ian  Emmons  and  Kathryn  Laskey   STIDS  2013  Technical  Chairs     Paulo  Costa   STIDS  2013  General  Chair       STIDS 2013 Proceedings Page i       STIDS  2013  Committees       STIDS 2013 Program Committee Stephen Allen Semantic Research, Inc. Carl Andersen Raytheon BBN Technologies Robert Battle Amazon.com Rom m el George Mason University Carvalho W erner Ceusters University at Buffalo Suzanne Collier Raytheon BBN Technologies Paulo Costa George Mason University M ike Dean Raytheon BBN Technologies Jody Des Roches Joint Warfare Analysis Center Ian Em m ons Raytheon BBN Technologies M att Fisher Progeny Systems Corporation Katherine Xcelerate Solutions Goodier M ark Greaves Pacific Northwest National Lab Richard George Mason University H aberlin Brian H augh Institute for Defense Analyses John H ebeler University of Maryland - BC Terry Janssen SAIC, Inc. Greg Joiner Raytheon BBN Technologies Kenneth Kisiel Office of Naval Research M ieczylaw Northeastern University Kokar Dave Kolas Raytheon BBN Technologies Kathryn Laskey George Mason University N ancy Lawler US Department of Defense M ike Letsky Office of Naval Research W illiam Data Tactics, Inc. M andrick Dan M axwell KaDSci, Inc. Dave M ireles Raytheon BBN Technologies Ranjeev M ittu US Navy Research Laboratory Jeffrey M orrison Office of Naval Research Leo Obrst MITRE Corporation M ary Parm elee MITRE Corporation STIDS 2013 Proceedings Page ii   Andrew Perez- Opower, Inc. Lopez Plam en Petrov Raytheon BBN Technologies Setareh George Mason University Rafatirad Doug Reid Google, Inc. Cyladian Technology Joe Rockm ore Consulting Dorene Ryder Raytheon BBN Technologies Ciara Sibley US Navy Research Laboratory Barry Sm ith NCOR, University at Buffalo Tony Stein Raytheon BBN Technologies Gheorghe Tecuci George Mason University Andreas Tolk Old Dominion University Brian U licny Vistology, Inc. Andrea SAIC W esterinen Dum inda George Mason University W ijesekera STIDS 2013 Proceedings Page iii   STIDS Steering Committee Paulo Costa George Mason University M ike Dean Raytheon BBN Technologies Ian Em m ons Raytheon BBN Technologies Katherine Goodier NIC, Inc. Terry Janssen SAIC Kathryn Laskey George Mason University W illiam M andrick Data Tactics Leo Obrst MITRE Corporation Barry Sm ith NCOR, University at Buffalo STIDS 2013 Organizing Committee General Chair Paulo Costa Technical Chairs Ian Em m ons Kathryn Laskey Publicity Chair W illiam M andrick Classified Session Chair Brian H augh Local Team (GM U ) Debra Schenaker (Adm inistrative Chair) Priscilla M cAndrews Shou M atsum oto Felipe Bom barda Karen Tai STIDS 2013 Proceedings Page iv   STIDS  2012  Platinum  Sponsor     Data  Tactics  Corporation  is  a  minority-­‐ owned  small  business  that  specializes  in   Data  Management,  Data  Architecture,  Data   Engineering,  Semantic  Data   Representations,  and  Big  Data.    Since  2005,   our  engineers  have  been  on  the  forefront  of   large,  multi-­‐domain,  data  systems   supporting  Government  and  commercial  organizations.    Our  engineering  staff  is  over  90%   TS/SCI  cleared  (many  with  polygraphs)  with  over  25%  having  advanced  degrees  and   doctorates.     We  offer  a  suite  of  solutions  to  help  customers  handling  very  large,  “Big  Data”  problem  sets.   Our  team  of  senior  engineers  and  data  scientists  excel  at  the  most  intractable  problems  for   customers  such  as  AIR  FORCE,  ARMY,  DARPA,  DHS,  DNI,  NSA  and  many  others.    From   tactical  to  strategic  efforts,  our  team  has  led  the  creation,  integration,  and  implementation   of  innovative  and  proven  solutions  in  the  world  of  Data  Alignment,  Modeling,  and  Analytics.   We  are  also  very  active  in  standards  development  including  the  NIST  Cloud  Computing  and   Big  Data  standards  along  with  Semantic  Standards  (e.g.  BFO,  SUMO,  DOLCE,  etc)  and   actively  contribute  to  the  open  source  communities  (e.g.  Apache,  Source  Forge,  GIT,  etc).         Data  Tactics  is  highly  invested  in  fostering  and/or  leading  collaborations  with  academia   and  national  labs  in  advanced  research  &  development  initiatives  that  support  disruptive   technologies.      Our  team  brings  a  rich  history  of  supporting  prototyping,  experimental   technology  integration,  mission  oriented  demonstrations,  and  specifically  cloud   development  and  integration.       DATA  TACTICS    –  WHAT  WE  DO   CLOUD/DISTRIBUTED  COMPUTING   • Information  Artifact  Ontology   REFERENCE  ARCHITECTURES   Development   • IC  ITE  DNI  Enterprise  Strategy   • Advanced  Machine  Learning   • Army  Red  Disk/NSA  Ghost   (i.e.  NLP)  integration   Machine   • Advanced  Video  and  Image   • DCGS-­‐A  Standard  Cloud  (DSC)   Entity  extraction     • Air  Force  TENCAP   SECURE  DATABASE  ARCHITECTURES     TURNKEY  BIG  DATA  IMPLEMENTATIONS   • Secure  Entity  Database  (SED)   • Secure  Enterprise  Hadoop   • Defense  Cross-­‐Domain  Analytic   Capability  (DCAC)   • Elastic  Ingest  and  Semantic     Markup   CYBERSECURITY   • Distributed  Analytics   • Information  Assurance       ADVANCED  ANALYTICS   • Security  Architecture,  Design,   and  Configuration   • Multiple  Algorithm   Development  Programs     • Policies,  Process  Development,   and  Validation   STIDS 2013 Proceedings Page v         STIDS  2012  Keynote  Speaker:  November  13         Dr.  Benjamin  Grosof   Highly  Expressive  yet  Scalable  Knowledge  for   Intelligence,  Defense,  and  Security   We  present  recent  results  on  semantic  web  knowledge  representation  &  reasoning,   and   knowledge   acquisition,   that   tightly   combine   highly   expressive   rules   and   ontologies   specified   semi-­‐automatically   -­‐-­‐   yet   rapidly   -­‐-­‐   by   starting   from   effectively   unrestricted   English  text.  The  knowledge  employs  a  new  logic,  Rulelog,  that  has  strong  capabilities  to   express  meta  knowledge.  Rulelog  permits  higher-­‐order  logic  formulas  that  are  defeasible   (i.e.,  can  have  exceptions).  It  is  rich  enough  to  serve  as  a  relatively  direct  target  for  natural   language   processing,   using   Textual   Logic,   a   new   method   that   employs   logic-­‐based   mappings  in  natural  language  (NL)  text  interpretation  and  text  generation.     Rulelog  also  leverages  its  meta  capabilities  to  achieve  computational  tractability  via   restraint,   a   new   form   of   bounded   rationality.   Rulelog,   and   the   Textual   Logic   that   leverages   it,  constitute  a  pretty  radical  step  forward  in  terms  of  fundamental  capabilities  in  semantic   tech,   with   a   number   of   advantages   including   in   the   social   scalability   of   knowledge   authoring  and  reuse.  Yet  this  step  is  incremental  relative  to  legacy  technology,  in  that  the   new   knowledge   representation   transforms   into   the   same   fundamental   logic   as   used   in   relational  and  RDF  databases  and  in  commercially  predominant  business  rule  systems.  It's   applicable  in  several  intelligence,  defense,  and  security  (IDS)  areas  including:  info  access   policies   (e.g.,   confidentiality,   compliance);   info   integration,   flow   and   ontology   mapping   (e.g.,  in  situation  awareness);  modeling  of  causal  events  and  risk;  intelligence  analysis  and   debate;   e-­‐learning   (e.g.,   just-­‐in-­‐time   training);   contracts   (e.g.,   compliance);   question-­‐ answering  (QA);  and  NL-­‐based  human-­‐computer  interaction  (HCI).     We  discuss  some  of  the  exciting  opportunities  and  challenges.     STIDS 2013 Proceedings Page vi       Biography:  Dr.  Benjamin  Grosof   Benjamin   Grosof   is   an   industry   leader   in   knowledge   representation,   reasoning,   and   acquisition.   He   has   pioneered   semantic   technology   and   industry   standards   for   rules,   the   combination  of  rules  with  ontologies,  the  applications  of  rules  in  e-­‐commerce  and  policies,   and  the  acquisition  of  rules  and  ontologies  from  natural  language  (NL).  He  has  had  driving   roles   in   RuleML,   W3C   RIF   (Rule   Interchange   Format),   and   W3C   OWL-­‐RL   (rule-­‐based   ontologies).  He  led  the  invention  of  several  fundamental  technical  advances  in  knowledge   representation,   including   courteous   defeasibility,   restraint   bounded   rationality,   and   the   rule-­‐based   technique,   which   rapidly   became   the   currently   dominant   approach   to   commercial   implementation   of   OWL.   He   has   extensive   experience   in   machine   learning,   probabilistic  reasoning,  and  user  interaction  design.     Dr.  Grosof  has  experience  applying  core  technology  for  knowledge,  reasoning,  and   related   HCI   in   a   wide   variety   of   application   areas,   including:   trust/privacy/security,   contracts,   compliance,   legal,   and   services   engineering;   financial/   insurance   services,   risk   management,and  regulations;  defense  and  national  intelligence;  biomedical  research;  and   data/  decision  analytics.  From  fall  2007  to  early  2013,  he  led  a  large  research  program  in   Artificial  Intelligence  (AI)  and  rule-­‐based  semantic  technologies  at  Vulcan  Inc.  for  Paul  G.   Allen;   this   centered   around   the   SILK   system   for   highly   expressive,   yet   scalable,   rules.   Previously   he   was   an   IT   professor   at   MIT   Sloan   (2000-­‐2007)   and   a   senior   software   scientist   at   IBM   Research   (1988-­‐2000).   He   is   president   of   the   expert   consulting   firm   Benjamin  Grosof  &  Associates  founded  while  he  was  at  MIT,  and  co-­‐founder  of  the  recent   start-­‐up  Coherent  Knowledge  Systems.     His   background   includes   4   major   industry   software   releases,   2   years   in   software   startups,   a   Stanford   PhD   (Computer   Science),   a   Harvard   BA   (Applied   Mathematics),   2   patents,  and  over  50  refereed  publications.         STIDS 2013 Proceedings Page vii       STIDS  2013  Keynote  Speaker:  November  14     Dr.  Jeffrey  Morrison   Exploring  the  role  of  Context     in  Applied  Decision  Making   Decision   makers   in   operational   environments   are   often   surprised   by   emerging   events   and   have   little   time   to   give   deep   consideration   to   alternative   courses   of   action   before   being   forced   to   make   a   decision.   Decision   support   has   evolved   over   the   last   20   years   but   even   today,   decision   support   tools   do   not   dynamically   adapt   to   a   decision   maker’s   context.     This   often   results   in   less   than   optimal   decision   making.   Recent   advances   in   the   fields   of   cognitive   science,   the   mathematics   of   decision   science,   human   behavioral   modeling,   team   decision   making,   knowledge   creation   and   transfer,   mental   model   processes,   semantic   techniques   and   human   factors   present   new   opportunities   to   create   decision  support  that  is  context  sensitive,  and  potentially,  proactive.  To  accomplish  this,  a   systematic   exploration   of   the   role   of   context   needs   to   be   studied   in   decision   support   systems  that  enable  operational  decision  making.   Decision   making   is   challenging   for   a   number   of   reasons.   Finding   and   integrating   decision-­‐relevant  information  is  hard.  Context  is  often  absent,  implicit,  sparsely  or  poorly   represented   in   task   environments   requiring   its   laborious   and   error-­‐prone   internal   reconstruction   by   decision   makers.   The   modern   pace   of   operations   often   means   that   warfighters   find   themselves   engaging   in   tasks   in   ways,   and   in   combinations,   for   which   they  hadn’t  planned,  and  for  which  they  may  not  be  prepared.  This  forces  decision  makers   to   multi-­‐task   amongst   many   competing   and   often   conflicting   mission   objectives   concurrently.     Next   generation   decision   support   will   not   just   “get   the   decision   maker   in   the   ball   park”   but   will   be   proactive   in   trying   to   “keep   the   decision   maker   in   the   ball   park”     STIDS 2013 Proceedings Page viii       throughout   the   process   despite   the   high   levels   of   uncertainty   and   highly   dynamic   environments.    At  the  center  of  this  new  research  initiative  is  the  idea  that  we  can  develop   technologies   that   are   contextually   aware   of   a   decision   makers’   missions   and   tasks.     It   is   asserted   that   algorithms   can   be   developed   that   effectively   anticipate   the   decision   and   information   needs   of   decision   makers,   in   many   kinds   of   task   environments.     Algorithms   would   then   enable   the   timely   presentation   of   information.   Enabling   machines   to   dynamically   model   and   share   context   with   the   human   decision   makers   will   be   key   to   enabling   Proactive   Decision   Support   (PDS).   Such   decision   support   will   enable   the   recognition  of  changes  in  the  environment  and  the  implications  for  shifting  priorities  for   decisions  that  could  address  operational  complexity  and  make  enable  decision  makers  to   make  more  optimal  decisions,  faster.       Biography:  Dr.  Jeffrey  Morrison   Dr.  Jeffrey  G.  Morrison  joined  ONR’s  Human  &  Bioengineered  Systems  Department   (341)  as  a  Program  Officer  in  January  2011  where  he  leads  the  Command  Decision  Making   (CDM)   program.     The   program   is   conducting   Basic   &   Applied   cognitive   science   research   for  application  to  individual  &  group  decision  making.    The  current  operational  focus  is  on   multi-­‐echelon   Command   &   Control.     The   science   focus   is   on   developing   Proactive   Decision   Support  tools  (PDS)  that  are  aware  of  mission  and  tasks  context  as  well  as  the  facilitating   the  development  of  a  science  of  Context-­‐Driven  Decision  Making  (CDDM).     Prior  to  coming  to  ONR,  Dr.  Morrison  was  an  Engineering  Psychologist  /  Cognitive   Scientist   with   the   Space   and   Naval   Warfare   Systems   Center   –   Pacific   (SSC   Pacific)   for   17   years.     He   was   most   recently   embedded   as   a   Navy   Scientist   with   the   Combating   Terrorism   Technical   Support   Office   (CTTSO)   where   he   served   as   Chief   Scientist   to   the   ASD   RDT&E   sponsored  Human  Social  Culture  and  Behavior  Modeling  Program  (HSCB).     During  2007-­‐ 2008,  Dr.  Morrison  was  detailed  to  the  Director  of  National  Intelligence  where  he  served   as   an   IARPA   Program   Manager   studying   the   analytic   process   and   the   potential   application   of  virtual  world  technologies  to  enable  it.     Dr.  Morrison  was  a  senior  scientist  supporting   several   DARPA   projects,   including   the   development   of   user-­‐composable   automation   for   Maritime   Domain   Awareness   (FastC2AP),   Predictive   Analysis   for   Naval   Deployment   Activity   (PANDA),   and   the   Augmented   Cognition   program.     He   also   was   principle   investigator   for   numerous   ONR   sponsored   projects,   including:   Knowledge   Web   (K-­‐Web),   and  Tactical  Decision  Making  Under  Stress  (TADMUS).   Dr   Morrison   has   been   the   recipient   of   numerous   professional   awards   including:   The   2005   Jerome   H   Ely   Award   for   Article   of   the   Year   in   the   Journal   of   Human   Factors;   the   2004   ONR   Arthur   E.   Bisson   Prize   for   Naval   Technology   Achievement;   and   the   American   Psychological  Association  -­‐  Division  21,  George  E.  Briggs  Award  for  Original  Research.     STIDS 2013 Proceedings Page ix