Open Innovation and Semantic Web :
Problem Solver Search on Linked Data
Milan Stankovic
hypios & STIH – Université Paris-S...
Challanges for OI on Semantic Web
• Specifics of OI:
– we seek innovative and disruptive solutions, that
might come form m...
Expert Finding before Linked Data
Content User Activities Reputation and
Acheivements
user-generated content
publications,...
A hidden assumption: Experties
hypothesis
Expert
Candidate
Expertise
Evidence
Expertise
Topic
hypothesis
If the user
wrote...
Expert Search on Linked Data
selection and
ranking of
experts
expertise
hypothesis
How to Choose an Expertise Hypothesis
• Look at the structure of data:
– global data or local data store
– dataset caracte...
Linked Data Metrics
• Metrics based on topic distribution
• Metrics based on topic proximity
• What has been done so far
– pilot study
• What’s been keeping us busy
– qualitative experiment: is there a correlation
b...
Hypothesis Recommendation and
Expert Finding system
• Hy.SemEx system
• Next Challange: Provide a way to explore
relevant ...
Questions Please?
Milan Stankovic
milan.stankovic@hypios.com
Upcoming SlideShare
Loading in...5
×

Open Innovation and Semantic Web

1,220

Published on

presentation given at the Doctoral Consortium of International Semantic Web Conference (ISWC) 2010

Published in: Technology
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total Views
1,220
On Slideshare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
22
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide

Open Innovation and Semantic Web

  1. 1. Open Innovation and Semantic Web : Problem Solver Search on Linked Data Milan Stankovic hypios & STIH – Université Paris-Sorbonne
  2. 2. Challanges for OI on Semantic Web • Specifics of OI: – we seek innovative and disruptive solutions, that might come form many places not necesairly best experts • Challanges for SW: – find experts using existing Linked Data sources – Find related domains where the solver might come from
  3. 3. Expert Finding before Linked Data Content User Activities Reputation and Acheivements user-generated content publications, e-mails, blogs, Wikipedia pages… Buitelaar, P., &Eigner, T. (2008) ;; Kolari, P., Finin, T., Lyons, K., &Yesha, Y. (2008) …. content owned by users Semantic desktop Demartini, G., &Niederée, C. (2008) online activities question answering, bookmarking Adamic et al. (2008) ; Zhang et al.. (2007) … offline activities obtaining research grants, participating in projects endorsment of user’s content Noll et al.(2009). .. replies Jurczyk, P., &Agichtein, E. (2007). data structured data selection and ranking of experts
  4. 4. A hidden assumption: Experties hypothesis Expert Candidate Expertise Evidence Expertise Topic hypothesis If the user wrote a paper saved a bookmark saved a bookmark before the others was retweeted on TopicX then he/she is an expert then he/she is a better ranked expert on TopicX
  5. 5. Expert Search on Linked Data selection and ranking of experts expertise hypothesis
  6. 6. How to Choose an Expertise Hypothesis • Look at the structure of data: – global data or local data store – dataset caracteristics already published with VoID and SCOVO – Tools that index data summeries: Khatchadourian, S., & Consens, M. (2010); Harth et al. (2010). • We propose Linked Data metrics based on: – data quantity – topic distribution – topic proximity
  7. 7. Linked Data Metrics • Metrics based on topic distribution • Metrics based on topic proximity
  8. 8. • What has been done so far – pilot study • What’s been keeping us busy – qualitative experiment: is there a correlation between the values of the metrics and the precsion and recall expectation of a hypothesis
  9. 9. Hypothesis Recommendation and Expert Finding system • Hy.SemEx system • Next Challange: Provide a way to explore relevant domains of knowledge and include them in the expert search. – considered work in: Recommender Systems based on semantic proximity; Serendipity; problem topic 1 topic 2 Recommend hypothesis VoID + SCOVO Find Experts Invite Experts Recommend Problems
  10. 10. Questions Please? Milan Stankovic milan.stankovic@hypios.com
  1. A particular slide catching your eye?

    Clipping is a handy way to collect important slides you want to go back to later.

×