freya an interactive way of querying linked data using natural language n.
Download
Skip this Video
Loading SlideShow in 5 Seconds..
FREyA : an Interactive Way of Querying Linked Data using Natural Language PowerPoint Presentation
Download Presentation
FREyA : an Interactive Way of Querying Linked Data using Natural Language

Loading in 2 Seconds...

play fullscreen
1 / 18
jade-gamble

FREyA : an Interactive Way of Querying Linked Data using Natural Language - PowerPoint PPT Presentation

107 Views
Download Presentation
FREyA : an Interactive Way of Querying Linked Data using Natural Language
An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.

- - - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript

  1. FREyA:an Interactive Way of Querying LinkedData using Natural Language Danica Damljanović, Milan Agatonović, Hamish Cunningham contact: danica@dcs.shef.ac.uk

  2. Natural Language Interfaces and Linked Open Data • What are NLIs? • Challenges: • NL understanding/grammar • ambiguity/expressiveness • knowledge structure/portability • small ontologies, large ontologies, multiple ontologies, Linked Open Data? QALD-1

  3. Shift in Challenges • Portability 5 years ago vs. today? • Heterogeneity, incompleteness, redundancy • Can one system support both? QALD-1

  4. FREyA workflow QALD-1

  5. Finding POCs ESWC 2010

  6. Finding OCs ESWC 2010

  7. Mapping POC to OCs POC POC population geo:State geo:State new york geo:City ESWC 2010 geo:cityPopulation

  8. New York is a city ESWC 2010

  9. New York is a state ESWC 2010

  10. Learning IF THEN

  11. Querying Linked Data with FREyA:the usual cycle for i=1 to n { Initialise the system using dataset Ai(forceDialog) Train the system by asking questions Save learningModeli } Intialise the system (automatic mode) by loading learningModeli, i=1,n connect to the repository containing all Aidatasets, where i=1, n QALD-1

  12. evaluation • Initialisation • System performance • Precision, recall, f-measure using MusicBrainz and DBPedia datasets • Mean Reciprocal Rank to assess the effect of learning mechanism • Analysis of failures QALD-1

  13. Initialisation and the datasets size QALD-1

  14. Results: F-measure statistics QALD-1

  15. Learning QALD-1

  16. Conclusion • Output: • Correct answer OR • Identifying the flaws in the data? • Ranking/disambiguation algorithms to improve MRR QALD-1

  17. thank you for your attention! questions? Thanks to Ivan Peikov from Ontotext who helped with the configuration of OWLIM which was necessary for performing the experiments reported in this paper. Contact: danica@dcs.shef.ac.uk

  18. Demo United States geography: • http://gate.ac.uk/sale/dd/movies/mooney/html/freya.html MusicBrainz: • http://gate.ac.uk/sale/dd/movies/mb/html/musicbrainz.html • DBpedia: • http://gate.ac.uk/sale/dd/movies/dbpedia/html/small/freya.html