Diese Präsentation wurde erfolgreich gemeldet.
Die SlideShare-Präsentation wird heruntergeladen. ×

Semantic technologies at work - 2007

Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Anzeige
Nächste SlideShare
Semantic Technologies - 2007
Semantic Technologies - 2007
Wird geladen in …3
×

Hier ansehen

1 von 22 Anzeige

Weitere Verwandte Inhalte

Diashows für Sie (20)

Anzeige

Ähnlich wie Semantic technologies at work - 2007 (20)

Anzeige

Aktuellste (20)

Semantic technologies at work - 2007

  1. 1. Semantic Technologies at work Experiences and issues A talk by Yannis Kalfoglou at Date of presentation here
  2. 2. Overview <ul><li>Emergence ( and maturity) of semantic technologies </li></ul><ul><ul><li>from the academic silos to industrial settings </li></ul></ul><ul><ul><li>Business incentives </li></ul></ul><ul><ul><li>Application areas </li></ul></ul><ul><ul><li>Semantic Web and Web 2.0 drivers </li></ul></ul><ul><li>Experiences with Semantic Technologies </li></ul><ul><ul><li>MyPlanet – web-based organisational newsletter </li></ul></ul><ul><ul><ul><li>Enhancing information navigation and discovery </li></ul></ul></ul><ul><ul><li>CS AKTive Space – information portal </li></ul></ul><ul><ul><ul><li>“ health and state of a discipline” exploration – information integration and aggregation </li></ul></ul></ul><ul><ul><li>e-Response - sense making aid for decision makers </li></ul></ul><ul><ul><ul><li>Full fledged knowledge and information management using STs </li></ul></ul></ul><ul><li>Issues </li></ul><ul><ul><li>Web 2.0 impact </li></ul></ul><ul><ul><li>The role of interoperability </li></ul></ul><ul><ul><li>Trust and privacy </li></ul></ul><ul><ul><li>Revenue streams </li></ul></ul><ul><ul><li>STs adoption strategies </li></ul></ul>Semantic technologies Experiences - MyPlanet Experiences – CAS Experiences – e-Response Issues 2/14
  3. 3. Semantic Technologies why do we need them? <ul><li>business drivers </li></ul><ul><ul><li>unprecedented exposure of information in digitised form  too much information available: “ infosmog ” </li></ul></ul><ul><ul><li>turn that information into usable knowledge </li></ul></ul><ul><ul><li>collaborative, inter-organisational business systems </li></ul></ul><ul><ul><li>globalisation of business processes – need for increased interoperability </li></ul></ul><ul><ul><li>empowerment of the individual </li></ul></ul><ul><ul><li>retention of knowledge </li></ul></ul><ul><ul><li>reduction of implementation time and cost </li></ul></ul><ul><li>societal drivers </li></ul><ul><ul><li>trends in employment: “up to 30% of EU’s working population will be directly employed in the production and diffusion of knowledge” [EU 2004] </li></ul></ul><ul><ul><li>knowledge based economy </li></ul></ul><ul><ul><li>ICTs affordable and available to everyone – participation is cheaper and global </li></ul></ul><ul><ul><li>(1.1bn web users, 17bn devices on the Internet by 2012 – [IDC 2005/6] ) </li></ul></ul><ul><li>technology triggers </li></ul><ul><ul><li>syntax based systems can’t cope with information explosion ( semantic heterogeneity ) </li></ul></ul><ul><ul><li>semantics can be codified and represented in computational form </li></ul></ul><ul><ul><li>maturity of highly distributed open-end environments, e.g., (Semantic) Web; and Web 2.0 approaches (social networking) </li></ul></ul>3/14 Semantic technologies Experiences - MyPlanet Experiences – CAS Experiences – e-Response Issues
  4. 4. <ul><li>Emergence of “meaningful computing”: emphasis on codified representations of semantics </li></ul><ul><ul><li>represent meaning in computational forms; </li></ul></ul><ul><ul><li>enable automation and intelligent tasks; </li></ul></ul><ul><ul><li>better information management (search, discovery, integration); </li></ul></ul><ul><ul><li>enable re-use; </li></ul></ul><ul><ul><li>improved interoperability (semantic-based vs. syntax-based) </li></ul></ul><ul><li>Wide range of technologies: </li></ul><ul><ul><li>ontologies (OWL, KIF, Ontolingua, etc.) </li></ul></ul><ul><ul><li>data formats (RDF & RDF(S), ebXML ) </li></ul></ul><ul><ul><li>query languages (SPARQL) </li></ul></ul><ul><ul><li>APIs (HP Jena) </li></ul></ul><ul><ul><li>open standards (OG(eo)C, SUO, CL) </li></ul></ul><ul><ul><li>SOA </li></ul></ul><ul><ul><li>… </li></ul></ul><ul><li>Variety of application areas: </li></ul><ul><ul><li>knowledge representation & reasoning (ontologies) </li></ul></ul><ul><ul><li>semantic annotation (semantic tagging) </li></ul></ul><ul><ul><li>semantic querying (SPARQL, OWL-driven) </li></ul></ul><ul><ul><li>semantic interoperability and integration (ontology/schema mapping) </li></ul></ul><ul><ul><li>semantically enriched storage (RDF) </li></ul></ul>Semantic Technologies what are they? where do we apply them? 4/14 Semantic technologies Experiences - MyPlanet Experiences – CAS Experiences – e-Response Issues
  5. 5. <ul><li>Semantic Web </li></ul><ul><ul><li>(towards) an infrastructure made up of representation languages, communication protocols, access controls & authentication services for coordinated sharing of knowledge across particular domain-oriented applications </li></ul></ul><ul><ul><li>(vision) turning the web into a globally distributed knowledge base where software agents will assist in most, if not all, tasks of information management </li></ul></ul><ul><li>Web 2.0 </li></ul><ul><ul><li>enabling technologies for mass participation in social networking activities on the Web </li></ul></ul><ul><ul><li>range of supporting technologies (AJAX, wikis, tagging, bloging, etc.) </li></ul></ul><ul><ul><li>no centralised control, no agreed vocabularies, “wisdom of the clouds” (collective intelligence) </li></ul></ul><ul><li>Semantic technologies are a core component for both </li></ul>Semantic Technologies The Semantic Web and Web 2.0 impact 5/14 Semantic technologies Experiences - MyPlanet Experiences – CAS Experiences – e-Response Issues
  6. 6. 13/14 Semantic technologies  Experiences - MyPlanet Experiences – CAS Experiences – e-Response Issues MyPlanet: a personalised ontology-driven front-end to KMi Planet Java Applet run on any Java-enabled Web browser accessible from http://eldora.open.ac.uk/my-planet/
  7. 7. MyPlanet: the User Interface and email address Browse ontology-drawn hierarchy Display ontology-drawn information for the selected node Edit selections Control buttons Password-protected access Semantic technologies  Experiences - MyPlanet Experiences – CAS Experiences – e-Response Issues User enters account details
  8. 8. MyPlanet: ontology-drawn information Ontology-drawn hierarchy class: kmi-research-area class: kmi-research-theme class: organization class: project class: kmi-member class: kmi-technology class: application-domain User selects: Enrico Motta Instance of class: kmi-senior-research-fellow Local slots: has-job-title involved-in-projects Semantic technologies  Experiences - MyPlanet Experiences – CAS Experiences – e-Response Issues
  9. 9. MyPlanet: finding relevant e-Stories Standard “string-matching” Ontology-based reasoning: use of ontological relations that hold for the particular entity: kmi-member involved-in-projects Υ do string-matching on fire ontological relations that hold for Y: has-research-area Z do string-matching on … and so on. Y Z Z Semantic technologies  Experiences - MyPlanet Experiences – CAS Experiences – e-Response Issues
  10. 10. MyPlanet: using cue phrases to increase the answer set To increase the number of possible matches, we employ the notion of cue phrases which we associate with ontology instances: genetic-algorithms , instance of class: kmi-research-area is associated with cue phrases: evolutionary computing, evolutionary algorithms Hence, we do string-matching on those phrases: X has-cue-phrases Y do string-matching on hence, this story is related to X (genetic algorithms) Y Semantic technologies  Experiences - MyPlanet Experiences – CAS Experiences – e-Response Issues
  11. 11. <ul><li>CS AKTive Space </li></ul><ul><li> Award winning SW application </li></ul><ul><li> Exploring the domain of a discipline (CS) </li></ul><ul><li> Use of STs to locate, extract, model and use heterogeneous data in a meaningful manner </li></ul><ul><li> Integration and deployment of services </li></ul><ul><li> Time consuming engineering </li></ul><ul><li> Costly ontology building process </li></ul><ul><li> Difficult to re-purpose </li></ul><ul><li> Reliance on external data sources (reliability, responsiveness issues) </li></ul>Experiences 15/18 Semantic technologies  Experiences – MyPlanet  Experiences – CAS Experiences – e-Response Issues
  12. 12. Experiences 15/18 Semantic technologies  Experiences – MyPlanet  Experiences – CAS Experiences – e-Response Issues
  13. 13. <ul><li>e-Response demonstrator </li></ul><ul><li>Large scale AKT integrated feasibility demonstrator </li></ul><ul><li>9 technologies from 5 partners </li></ul><ul><li>realistic emergency response scenario </li></ul><ul><li>integrated view over heterogeneous sources and inference across them </li></ul><ul><li> Difficult to replicate </li></ul><ul><li> no integration framework </li></ul><ul><li> Reliance on external data sources </li></ul>Experiences 16/18 Semantic technologies  Experiences – MyPlanet  Experiences – CAS  Experiences – e-Response Issues
  14. 14. Scenarios and applications e-Response: term disambiguation and service enactment 13/14 <ul><li>Application of STs in an emergency response scenario </li></ul><ul><ul><li>aim to facilitate planners to make sense of a complex situation as it unfolds in real time </li></ul></ul><ul><ul><li>used a simulator and realistic scenario </li></ul></ul><ul><ul><li>focussed on real world resources and emergency services </li></ul></ul><ul><li>Technologies for: semantically enriched storage; semantic annotation, information extraction, sense making GUI, planning and scheduling, resource binding, communities of interest identification, service invocation and term disambiguation (modified CROSI CMS) </li></ul>Semantic technologies  Experiences – MyPlanet  Experiences – CAS  Experiences – e-Response Issues
  15. 15. <ul><li>Semantic technologies adoption </li></ul><ul><ul><li>“ start small with the big picture in mind” </li></ul></ul><ul><ul><li>(from lightweight semantic hypertext markup – microformats - to fully fledged ontologies) </li></ul></ul><ul><ul><li>Prepare for a major shift in business models </li></ul></ul><ul><ul><li>Long term investment </li></ul></ul><ul><ul><li>Keep it simple (semantics) </li></ul></ul><ul><ul><ul><li>Irreconcilable arguments among engineers at design time </li></ul></ul></ul><ul><ul><ul><li>Inappropriate interpretations and usage at deployment time </li></ul></ul></ul><ul><ul><ul><li>Hard to maintain – not stable over time – obsolete knowledge </li></ul></ul></ul><ul><li>Interoperability as a means to enable intelligent tasks in a variety of domains </li></ul><ul><ul><li>mobile web 2.0 trends (web 2.0 + mobile web) </li></ul></ul><ul><ul><li>more people with mobiles than PCs </li></ul></ul><ul><ul><li>Social networking (web 2.0) on the Web (single network) using a common driver (mobile device) </li></ul></ul><ul><ul><li>Global audience, global services, single marketplace </li></ul></ul><ul><li>Web access for “internet of things” </li></ul><ul><ul><li>Universal content, ubiquitous computing </li></ul></ul><ul><ul><li>Interoperability needs between: </li></ul></ul><ul><ul><ul><li>Global services, universal content, to enable social networking and ubiquitous computing </li></ul></ul></ul>Issues technology & socio-political 14/14 Semantic technologies  Experiences – MyPlanet  Experiences – CAS  Experiences – e-Response  Issues
  16. 16. Semantic Technologies maturity The Gartner Hype Cycle (7/06) – time to “plateau of productivity” (Public) Semantic Web: trough of disillusionment  5-10y (Corporate) Semantic Web: peak of inflated expectations  5-10y Web 2.0: peak of inflated expectations  2-5y Web 2.0 More visible Large end user base (Public) Semantic Web Invisible to the end user Infrastructure (Corporate) Semantic Web Better information management Visibility to end user is not an issue 7/18 Semantic technologies  Experiences – MyPlanet  Experiences – CAS  Experiences – e-Response  Issues
  17. 17. <ul><li>Technology gaps </li></ul><ul><ul><li>Semantic interoperability vs. ontology building </li></ul></ul><ul><ul><li>Large scale semantic annotation vs. tagging </li></ul></ul><ul><ul><li>Ontology deployment architectures vs. SOA </li></ul></ul><ul><li>Non-uniform representation of communities knowledge </li></ul><ul><ul><li>Emergent semantics idea relies on a stable and uniform representation of interests </li></ul></ul><ul><ul><li>but: non-standard vocabulary </li></ul></ul><ul><ul><li>Possible conflicts of representation </li></ul></ul><ul><ul><li>Reach local agreement first, before achieving global consensus </li></ul></ul><ul><ul><li>Reliance on machine learning (supervised vs. unsupervised methods) – training sets? </li></ul></ul><ul><li>Mechanised trust </li></ul><ul><ul><li>Model and representation – no consensus </li></ul></ul><ul><ul><li>Sourcing – gather, broadcast in centralised units? </li></ul></ul><ul><ul><li>Measuring – no consensus, “who rates the raters”? </li></ul></ul><ul><ul><li>Context – local trust in a global environment </li></ul></ul><ul><li>Cost </li></ul><ul><ul><li>Conversion cost: convert legacy data to RDF, expose data to RDF </li></ul></ul><ul><ul><li>Maintenance cost: ontologies need regular updates </li></ul></ul><ul><ul><li>Organisational re-structuring costs: altering information gathering processes </li></ul></ul><ul><ul><li>Transaction costs: change of management structures </li></ul></ul><ul><ul><li>Reducing costs: increase the size of user base to offset the high cost of development </li></ul></ul><ul><ul><ul><li>Commodity: license the use of ontologies </li></ul></ul></ul><ul><li>BNW vs. BFC arguments </li></ul>Issues technology 13/18 Semantic technologies  Experiences – MyPlanet  Experiences – CAS  Experiences – e-Response  Issues
  18. 18. <ul><li>social software: Embedded observation </li></ul><ul><ul><li>Live the culture you are helping to frame </li></ul></ul><ul><ul><li>Designers love what they do and infuse their passion into the system </li></ul></ul><ul><ul><li>Take into account culture that cannot be systematically tested or modelled </li></ul></ul><ul><ul><li>Awareness of social problems – prompt reaction, bonding with the users (community) </li></ul></ul><ul><li>Respect local, regional and cultural aspects </li></ul><ul><ul><li>Local and regional affairs more important to local communities than national/international </li></ul></ul><ul><ul><li>Culture and customs need to be respected </li></ul></ul><ul><li>Informal gatherings influencing formal decisions (an oxymoron?) </li></ul><ul><ul><li>Online social networking and its decision making power </li></ul></ul><ul><ul><li>Boundaries of informal vs. formal </li></ul></ul><ul><li>Social cohesion and digital multiculturalism </li></ul><ul><ul><li>Voice your concerns in the digital world vs. real world </li></ul></ul><ul><ul><li>Same values, same person, two different personas? </li></ul></ul><ul><li>Blend of state-control vs. “no control” processes </li></ul><ul><ul><li>traditional formal processes vs. web 2.0 style informal modus operandi </li></ul></ul>Issues socio-political 14/18 Semantic technologies  Experiences – MyPlanet  Experiences – CAS  Experiences – e-Response  Issues
  19. 19. Knowledge society an example case Social media No moderation or censorship; two-way communication Individuals’ contributions clearly acknowledged; anonymity discouraged Spin and attempt to control are discouraged Pull system – let people bring to them the content and relationships they want Highly distributed, not centralised Adopted from Dion Hinchcliffe’s web site Participation powered by the network effect Low cost People in charge (use and control) Benefit of global scale syndication Democratisation and change of ground rules Shift from institutional control to consumer control 9/18 Semantic technologies  Experiences – MyPlanet  Experiences – CAS  Experiences – e-Response  Issues
  20. 20. Interoperability a key enabling technology <ul><li>business incentives </li></ul><ul><ul><li>disparate and open-ended environments for applications </li></ul></ul><ul><ul><li>complex tasks and cross sector applications </li></ul></ul><ul><ul><li>Inevitably, heterogeneity inherited in the system </li></ul></ul><ul><ul><ul><li>variety of protocols, domains, languages, processes, interpretations </li></ul></ul></ul><ul><ul><li>preserve the meaning of concepts used in everyday transactions of information sharing </li></ul></ul><ul><ul><ul><li>a key enabler for automated collaborative tasks </li></ul></ul></ul><ul><li>semantic interoperability </li></ul><ul><ul><li>syntax alone is not sufficient to capture the subtle differences in meaning </li></ul></ul><ul><ul><li>semantics allow to express meaning in structured computational forms </li></ul></ul><ul><ul><li>use these forms, alongside syntactic based approaches, to enable semantic interoperability </li></ul></ul><ul><li>application areas ( non-exhaustive list ) </li></ul><ul><ul><li>database schema integration ( construct global views ) </li></ul></ul><ul><ul><li>data warehouses ( data format transformations ) </li></ul></ul><ul><ul><li>e-commerce ( message exchange ) </li></ul></ul><ul><ul><li>semantic query processing ( query terms disambiguation ) </li></ul></ul><ul><ul><li>ontology integration/merging ( construct top-level/global ontologies ) </li></ul></ul><ul><ul><li>ontology alignment/mapping (concepts’ alignment) </li></ul></ul><ul><ul><li>SW agents interoperability ( agents vocabulary alignment ) </li></ul></ul>6/14 Semantic technologies  Experiences – MyPlanet  Experiences – CAS  Experiences – e-Response  Issues
  21. 21. A talk delivered by Yannis Kalfoglou at Semantic Technologies at work Experiences and issues Detica presentation date here Thank you for listening
  22. 22. <ul><li>W3C Mobile Web Initiative </li></ul><ul><ul><li>Best Practices – online checker </li></ul></ul><ul><ul><li>mobileOK conformance mark – machine testable </li></ul></ul><ul><li>dot mobi </li></ul><ul><ul><li>mTLD (top level domain for mobiles – ICANN) </li></ul></ul>Standards, early adopters, trends <ul><li>Early adopters </li></ul><ul><ul><li>Orange – Pikeo, bubbletop, Soundtribes </li></ul></ul><ul><ul><li>Vodafone – Betavine </li></ul></ul><ul><ul><li>NeoMedia/Qode, radar.net, ShoZu, TellMe </li></ul></ul><ul><ul><li>BBC + CNN mobile  web 2.0 features – UGC </li></ul></ul><ul><li>Trends </li></ul><ul><ul><li>think beyond numbers (mobile device not only a phone – URIs, tags, etc.) </li></ul></ul><ul><ul><li>Pushing out too much content leads to microcontent </li></ul></ul>

×