15
A personal look at the future of Seman2c Technologies Valen2n Zacharias Riga, 7.7.2011

STI Summit 2011 - A personal look at the future of Semantic Technologies

Embed Size (px)

DESCRIPTION

 

Citation preview

Page 1: STI Summit 2011 - A personal look at the future of Semantic Technologies

A  personal  look  at  the  future  of  Seman2c  Technologies  

Valen2n  Zacharias  Riga,  7.7.2011  

Page 2: STI Summit 2011 - A personal look at the future of Semantic Technologies

Summary  

1.  The  future  of  Seman2c  Technologies  does  not  belong  to  the  current  Seman2c  Web  technology  stack  

2.  It  may  belong  to  the  principles  underlying  it    3.  For  that  there  are  s2ll  many  relevant  and  

exci2ng  topics    

Page 3: STI Summit 2011 - A personal look at the future of Semantic Technologies

A  word  about  myself  

•  Working  with  Seman2c  Technologies  since  2001  in  public  and  industry  funded  project  in  different  roles  in  Rudis  FZI  group  (and  at  Ontoprise)  

•  Since  2010  head  of  interdisciplinary  research  division  IPE  at  FZI  

Page 4: STI Summit 2011 - A personal look at the future of Semantic Technologies

…  framework  is  proposed  ..  …  report  on  implementa2on  in  

progress…  

…  func2onal  prototype…  

…  outline  architecture…  realis2c  example  ...  

…  prototypic  implementa2on    ...  …  use  case  is  described  …  

…  prototype  on  the  web…  

…  proof  of  concept  …  

…  approach  with  preliminary  evalua2on  …  

Page 5: STI Summit 2011 - A personal look at the future of Semantic Technologies

Summary  

1.  The  future  of  Seman2c  Technologies  does  not  belong  to  the  current  Seman2c  Web  technology  stack  

2.  It  may  belong  to  the  principles  underlying  it    3.  For  that  there  are  s2ll  many  relevant  and  

exci2ng  topics    

Page 6: STI Summit 2011 - A personal look at the future of Semantic Technologies

Use  Case  1-­‐  Social  Seman2c  Tagging  &  Collabora2ve  Ontology  Engineering  

•  Nega2ve  Experience:  –  Social  /  Quan2ta2ve  data  (it  maZers  whether  a  statement  has  been  made  once  or  ten  2mes)  not  well  representable.    

– World  is  not  ternary  (e.g.  <user,tag,resource,2me>),  hence  o]en  awkward  to  represent  and  query  

–  Statement  metadata  (creator,    visibility,  2mestamp)  awkward  to  manage    

– Open  World  schema  seman2cs  hard  to  grasp  for  users  (SKOS  rescued  as  here)  and  frequent  source  of  programming  errors  

Page 7: STI Summit 2011 - A personal look at the future of Semantic Technologies

Use  Case  2  –  Seman2c  Context  Management  

•  Managing  a  complex  representa2on  of  a  device’s  or  user’s  context  –  E.g.  for  Ambient  Assisted  Living  &  Adaptable  user  interfaces  

–  Challenge  is  to  integrate  data  from  many  sensors,  to  deduce  informa2on  from  that  and  to  detect  and  act  on  predefined  situa2ons  

•  Nega2ve  Experience:  – Missing  (classical)  schema  language  to  specify  contracts  between  different  components  involved  in  managing  the  context  informa2on  (and  between  triple  store  and  object  oriented  world)  

–  Awkward  to  deal  with  uncertainty  –  OWL  not  well  suited  to  formalize  the  needed  reasoning  

Page 8: STI Summit 2011 - A personal look at the future of Semantic Technologies

Use  Case  3  –  Enterprise  Informa2on  Integra2on  

•  Crea2ng  a  long  term  plan  to  tackle  an  enterprise’s  eternal  informa2on  integra2on  challenges  

•  Nega2ve  Experience:  –  (Currently  SW  tools  &  concepts  not  mature  enough  to  seriously  propose  this  to  companies)    

– Mapping  arbitrary  RDBs  and  XMLs  to  RDF  o]en  results  in  awkward  representa2ons  and  unclear  benefit  

– OWL  does  not  help  in  any  way  I  can  see  

Page 9: STI Summit 2011 - A personal look at the future of Semantic Technologies

Why  

Stuff  within  the  Seman2c  Technologies  and  Seman2c  Web  visions  

*:  here  meant:  RDF,  RDFS,  OWL  

Data  web  

Stuff  „Seman2c  

technologies“*  are  actually  good  for  

Page 10: STI Summit 2011 - A personal look at the future of Semantic Technologies

Summary  

1.  The  future  of  Seman2c  Technologies  does  not  belong  to  the  current  Seman2c  Web  technology  stack  

2.  It  may  belong  to  the  principles  underlying  it    3.  For  that  there  are  s2ll  many  relevant  and  

exci2ng  topics    

Page 11: STI Summit 2011 - A personal look at the future of Semantic Technologies

‚Underlying  Principles‘  /  Core  Ideas  

•  Complex  Databases:  pushing  the  boundaries  complex  informa2on  processing  in  databases  

•  (Partly)  Domain  Knowledge  Agnos:c  Systems:  Taking  domain  knowledge  out  of  code  and  db  schema  and  make  it  changeable  at  run2me  

•  Collabora:ve  Domain  Knowledge:  Make  formal  domain  knowledge  models  editable,  shareable  and  linkable  

•  Using  Web  Like  Data  Integra:on  to  tackle  very  large  data  integra2on  problems  

Page 12: STI Summit 2011 - A personal look at the future of Semantic Technologies

Summary  

1.  The  future  of  Seman2c  Technologies  does  not  belong  to  the  current  Seman2c  Web  technology  stack  

2.  It  may  belong  to  the  principles  underlying  it    3.  For  that  there  are  s2ll  many  relevant  and  

exci2ng  topics    

Page 13: STI Summit 2011 - A personal look at the future of Semantic Technologies

Future  Topics?  

•  Complex  Databases:  pushing  the  boundaries  complex  informa2on  processing  in  databases  –  In  CEP  /  Stream  Processing  – Elas2c,  distributed,  in  memory,  mul2-­‐tenancy  ,  everything  in  (business)  real  2me,  big  data,  schemaless  and  No(t  only)  SQL  

Page 14: STI Summit 2011 - A personal look at the future of Semantic Technologies

Future  Topics?  

•  Using  Web  Like  Data  Integra2on  to  tackle  very  large  data  integra:on  problems  – An  En2re  Enterprise  – All  the  laws  of  a  state  (or  the  EU)  – An  en2re  scien2fic  field  

Page 15: STI Summit 2011 - A personal look at the future of Semantic Technologies

Summary  

1.  The  future  of  Seman2c  Technologie  does  not  belong  to  the  current  Seman2c  Web  technology  stack  

2.  It  may  belong  to  the  principles  underlying  it    3.  For  that  there  are  s2ll  many  relevant  and  

exci2ng  topics