Toward Representation Independent Similarity Search Over Graphs - PowerPoint PPT Presentation

Toward Representation Independent Similarity Search Over Graphs Yodsawalai Chodpathumwan , University of Illinois at Urbana-Champaign Arash Termehchy , Oregon State University Yizhou Sun , Northeastern University Amirhossein Aleyasin , University of Illinois at Urbana-Champaign Jose Picado , Oregon State University

1 Similarity Search over Graph Databases Which movies are similar to “Casino Royale” in IMDb? Daniel Craig Lord Asriel actor character James Bond Casino Royale Golden Compass character movie movie Query Martin Campbell Casino Royale director Answers Mask of Zorro Green Lantern Golden Compass movie movie Green Lantern Mask of Zorro

2 Same information is represented in many ways Daniel Craig actor Lord Asriel character James Bond Casino Royale character Golden Compass movie movie Lord Asriel Martin Campbell character Daniel Craig director actor starring Mask of Zorro Green Lantern starring movie movie Golden Compass James Bond movie Casino Royale character movie directed-by Martin Campbell director Green Lantern Mask of Zorro directed directed movie movie by by

3 Same information is represented in many ways EDBT KDD conference conference 2009 RankClus Truth Discovery 2007 year paper paper year J. Han P. S. Yu Y. Sun, PVLDB’11 author author P. Zhao, CIKM’09 EDBT KDD conference conference 2009 2007 J. Han P. S. Yu year year author author RankClus Truth Discovery paper paper

4 Which movies are similar to “ Casino Royale ” ? Daniel Craig actor Lord Asriel character James Bond Casino Royale character movie Golden Compass movie Martin Campbell director Mask of Zorro Green Lantern SimRank [Jeh , KDD’02] movie movie Golden Compass Mask of Zorro, Green Lantern

5 Which movies are similar to “ Casino Royale ” ? Daniel Craig actor Lord Asriel character starring starring James Bond Casino Royale character Golden Compass movie movie directed-by Martin Campbell director directed-by directed-by SimRank Mask of Zorro Green Lantern Mask of Zorro, movie movie Green Lantern Golden Compass

6 Generality of Similarity Search over Graphs •

7 Invertible Transformation • Daniel Craig Daniel Craig actor actor James Bond James Bond starring character character T – 1 T Casino Royale Casino Royale movie film

8 Information Preserving Transformation • Daniel Craig Daniel Craig actor actor James Bond James Bond starring character character T – 1 T Casino Royale Casino Royale movie film

9 Information Equivalent Graphs Databases D 1 and D 2 are information equivalent iff there exists an information preserving transformation from D 1 to D 2 and vice versa. Daniel Craig Daniel Craig actor actor James Bond James Bond starring character character Casino Royale Casino Royale movie film

10 SimRank is not general Daniel Craig actor Lord Asriel character James Bond Casino Royale character Golden Compass movie movie Lord Asriel Martin Campbell character director Daniel Craig actor starring Mask of Zorro Green Lantern movie movie starring Golden Compass SimRank James Bond movie Casino Royale character Golden Compass movie Mask of Zorro, directed-by Green Lantern SimRank Martin Campbell director Mask of Zorro, Green Lantern Mask of Zorro Green Lantern directed directed movie movie by by Golden Compass

11 Evaluation on Generality of Similarity Search Algorithms Ranked DB answers (Movies,Bib) Algorithms Q T SimRank, 30 RWR, Kendall’s Tau queries/DB PathSim Ranked T(DB) answers (Movies,Bib)

12 Average Ranking Differences • Top 10 answers Movie DBs Bib DBs Algorithms RWR [Tong, KDD’06] 0.375 0.773 SimRank [Jeh , KDD’02] 0.418 0.626 PathSim [Sun, VLDB’11] 0.375 0.953 • Top 50 answers Movie DBs Bib DBs Algorithms RWR [Tong, KDD’06] 0.204 0.718 SimRank [Jeh , KDD’02] 0.264 0.673 PathSim [Sun, VLDB’11] 0.110 0.688

13 Conclusion • Introduced the problem of representation independent similarity search over graph data. • Proposed a formal framework to measure generality • Performed empirical study and showed that some of the well-known similarity search algorithms are not general.

Unused Slides

When PathSim fails?

17 Transformations Movie Databases IMDb Freebase Bibliography Databases [P. Zhao, CIKM’09] [Y. Sun, PVLDB’11]