yutao zhang jie tang zhilin yang jian pei and philip s yu
play

Yutao Zhang + , Jie Tang + , Zhilin Yang + , Jian Pei # , and Philip - PowerPoint PPT Presentation

COSNET: Connecting Heterogeneous Social Networks with Local and Global Consistency Yutao Zhang + , Jie Tang + , Zhilin Yang + , Jian Pei # , and Philip S. Yu* + Tsinghua University # Simon Fraser University * University of Illinois at Chicago 1


  1. COSNET: Connecting Heterogeneous Social Networks with Local and Global Consistency Yutao Zhang + , Jie Tang + , Zhilin Yang + , Jian Pei # , and Philip S. Yu* + Tsinghua University # Simon Fraser University * University of Illinois at Chicago 1

  2. AMiner II (ArnetMiner) p Academic Social Network Analysis and Mining system—AMiner ( http://aminer.org ) p Online since 2006 p >38 million researcher profiles p >100 million publications p >241 million requests p >12.35 Terabyte data p 100K IP access from 170 countries per month p 10% increase of visits per month p Deep analysis, mining, and search 2

  3. Knowledge Acquisition from the Web (ACM TKDD, WWW’12, ISWC’06, ICDM’07, ACL’07) Ruud Bolle Office: 1S-D58 video database indexing Letters: IBM T.J. Watson Research Center Contact Information Ruud Bolle video processing P.O. Box 704 visual human-computer interaction Yorktown Heights, NY 10598 USA Office: 1S-D58 biometrics applications Packages: IBM T.J. Watson Research Center IBM T.J. Watson Letters: IBM T.J. Watson Research Center 19 Skyline Drive Research Center P.O. Box 704 IBM T.J. Watson Research Hawthorne, NY 10532 USA Yorktown Heights, NY 10598 USA Center Research_Interest Email: bolle@us.ibm.com Packages: IBM T.J. Watson Research Center 19 Skyline Drive Research Staff 19 Skyline Drive Affiliation Hawthorne, NY 10532 USA Ruud M. Bolle was born in Voorburg, The Netherlands. He received the Bachelor's Hawthorne, NY 10532 USA Degree in Analog Electronics in 1977 and the Master's Degree in Electrical IBM T.J. Watson Research Email: bolle@us.ibm.com Address http://researchweb.watson.ibm.com/ Engineering in 1980, both from Delft University of Technology, Delft, The Center Position ecvg/people/bolle.html Educational history Netherlands. In 1983 he received the Master's Degree in Applied Mathematics and in P.O. Box 704 Ruud M. Bolle was born in Voorburg, The Netherlands. He received the Bachelor's 1984 the Ph.D. in Electrical Engineering from Brown University, Providence, Rhode Yorktown Heights, Degree in Analog Electronics in 1977 and the Master's Degree in Electrical Address Homepage Island. In 1984 he became a Research Staff Member at the IBM Thomas J. Watson Engineering in 1980, both from Delft University of Technology, Delft, The NY 10598 USA Research Center in the Artificial Intelligence Department of the Computer Science Netherlands. In 1983 he received the Master's Degree in Applied Mathematics and in Photo Ruud Bolle Department. In 1988 he became manager of the newly formed Exploratory Computer 1984 the Ph.D. in Electrical Engineering from Brown University, Providence, Rhode bolle@us.ibm.com Email Name Vision Group which is part of the Math Sciences Department. Island. In 1984 he became a Research Staff Member at the IBM Thomas J. Watson Research Center in the Artificial Intelligence Department of the Computer Science Ruud Bolle Bsdate 1984 Phddate 1977 Currently, his research interests are focused on video database indexing, video Department. In 1988 he became manager of the newly formed Exploratory Computer Bsuniv Phduniv Brown University processing, visual human-computer interaction and biometrics applications. Vision Group which is part of the Math Sciences Department. Delft University of Technology Phdmajor Bsmajor Electrical Engineering Msdate Ruud M. Bolle is a Fellow of the IEEE and the AIPR. He is Area Editor of Computer Currently, his research interests are focused on video database indexing, video Analog Electronics Vision and Image Understanding and Associate Editor of Pattern Recognition. Ruud processing, visual human-computer interaction and biometrics applications. 1980 Academic services Msmajor Msuniv M. Bolle is a Member of the IBM Academy of Technology. Msmajor Delft University of Technology Ruud M. Bolle is a Fellow of the IEEE and the AIPR. He is Area Editor of Computer Electrical Engineering Vision and Image Understanding and Associate Editor of Pattern Recognition. Ruud Co-author Applied Mathematics Co-author M. Bolle is a Member of the IBM Academy of Technology. Publication 2# Publication 1# DBLP: Ruud Bolle Publications Title Title 2006 . . . Fingerprint Nalini K. Ratha, Jonathan Connell, Ruud M. Bolle, Sharat Chikkerur: Cancelable Biometrics: Representation Using 50 EE Date A Case Study in Fingerprints. ICPR (4) 2006: 370-373 Cancelable Biometrics: Date Localized Texture Venue A Case Study in End_page Features Venue Sharat Chikkerur, Sharath Pankanti, Alan Jea, Nalini K. Ratha, Ruud M. Bolle: Fingerprint End_page Fingerprints 49 EE Start_page 2006 Representation Using Localized Texture Features. ICPR (4) 2006: 521-524 Start_page 2006 ICPR 521 524 Andrew Senior, Arun Hampapur, Ying-li Tian, Lisa Brown, Sharath Pankanti, Ruud M. Bolle: 370 373 ICPR 48 EE Appearance models for occlusion handling. Image Vision Comput. 24(11): 1233-1243 (2006) 2005 UIUC coauthor affiliation Ruud M. Bolle, Jonathan H. Connell, Sharath Pankanti, Nalini K. Ratha, Andrew W. Senior: Publication #3 47 EE Ruud Bolle The Relation between the ROC Curve and the CMC. AutoID 2005: 15-20 Professor position Sharat Chikkerur, Venu Govindaraju, Sharath Pankanti, Ruud M. Bolle, Nalini K. Ratha: 46 EE Novel Approaches for Minutiae Verification in Fingerprint Images. WACV. 2005: 111-116 coauthor ... Publication #5 3

  4. Researcher Profile Database [1] Extracted more than 1,000,000 researcher profiles from the Web [1] J. Tang, L. Yao, D. Zhang, and J. Zhang. A Combination Approach to Web User Profiling. ACM Transactions on Knowledge Discovery from 4 Data (TKDD), (vol. 5 no. 1), Article 2 (December 2010), 44 pages.

  5. Is this Enough? 5

  6. Required semantics are distributed in multiple sources LinkedIn Videolectures 6

  7. Identity Linking • Identifying users from multiple heterogeneous networks and integrating semantics from the different networks together. 7

  8. COSNET : Connecting Social Networks with Local and Global Consistency • Input: G ={ G 1 , G 2 , …, G m }, with G k =( V k , E k , R k ) • Formalization: X ={ x i }, all possible pairwise y i ∈ {1,0} matchings and each corresponds to • COSNET: an energy-based model Y * = argmin E ( Y , X ) [1] Yutao Zhang, Jie Tang, Zhilin Yang, Jian Pei, and Philip Yu. COSNET: Connecting Heterogeneous Social Networks 8 with Local and Global Consistency. KDD’15.

  9. Local vs. Global consistency • Given three networks, 𝐻 1 Username: Ortiz_Brandy Nation: USA 1 𝑤 1 Gender: female 1 1 𝑤 2 𝑤 3 Username: 2 3 @ortizbrandy 𝑤 1 𝑤 1 Nation: USA Gender: female 𝐻 2 𝐻 3 2 2 3 3 𝑤 2 𝑤 3 𝑤 2 𝑤 3 9

  10. Local vs. Global consistency • Local matching: matching users by profiles Pairwise similarity features – Username similarity and 𝐻 1 Username: Ortiz_Brandy uniqueness Nation: USA 1 𝑤 1 Gender: female Local – Profile content similarity 局部一致性 consistency – Ego network similarity 1 1 𝑤 2 𝑤 3 – Social status Username: Energy function @ortizbrandy 2 3 𝑤 1 𝑤 1 Nation: USA Gender: female 𝐻 2 𝐻 3 2 2 3 3 𝑤 2 𝑤 3 𝑤 2 𝑤 3 10

  11. Local vs. Global consistency • Network matching: matching users’ ego networks 𝐻 1 Username: Ortiz_Brandy Nation: USA 1 𝑤 1 Gender: female Network Local 网络一致性 局部一致性 matching consistency Encourage “neighborhood 1 1 𝑤 2 𝑤 3 -preserving matching” Username: 2 3 @ortizbrandy 𝑤 1 𝑤 1 Nation: USA Gender: female 𝐻 2 𝐻 3 2 2 3 3 𝑤 2 𝑤 3 𝑤 2 𝑤 3 11

  12. Local vs. Global consistency • Network matching: matching users’ ego networks 1 1 𝑤 2 𝑤 1 True True 𝐻 1 Username: Ortiz_Brandy Nation: USA 1 𝑤 1 Gender: female Network 2 2 𝑤 1 𝑤 2 Local 网络一致性 局部一致性 matching consistency 1 1 𝑤 2 𝑤 3 Username: 2 3 @ortizbrandy 𝑤 1 𝑤 1 Nation: USA Gender: female 𝐻 2 𝐻 3 2 2 3 3 𝑤 2 𝑤 3 𝑤 2 𝑤 3 12

  13. Local vs. Global consistency • Network matching: matching users’ ego networks 1 1 𝑤 2 𝑤 1 True True 𝐻 1 Username: Ortiz_Brandy Nation: USA 1 𝑤 1 Gender: female Network 2 2 𝑤 1 𝑤 2 Local 网络一致性 局部一致性 matching consistency 1 1 𝑤 2 1 𝑤 3 1 𝑤 1 𝑤 2 x True False Username: 2 3 @ortizbrandy 𝑤 1 𝑤 1 2 2 Nation: USA 𝑤 2 𝑤 1 Gender: female 𝐻 2 𝐻 3 2 2 3 3 𝑤 2 𝑤 3 𝑤 2 𝑤 3 13

  14. Local vs. Global consistency • Network matching: matching users’ ego networks 1 1 𝑤 2 𝑤 1 True True 𝐻 1 Username: Ortiz_Brandy Nation: USA 1 𝑤 1 Gender: female Network 2 2 𝑤 1 𝑤 2 Local 网络一致性 局部一致性 matching consistency 1 1 𝑤 2 1 𝑤 3 1 𝑤 1 𝑤 2 x True False Username: 2 3 @ortizbrandy 𝑤 1 𝑤 1 2 2 Nation: USA 𝑤 2 𝑤 1 Gender: female 𝐻 2 𝐻 3 1 1 𝑤 2 2 2 3 3 𝑤 1 𝑤 2 𝑤 3 𝑤 2 𝑤 3 x False False x 2 2 𝑤 2 𝑤 1 14

  15. Local vs. Global consistency • Network matching: matching users’ ego networks 1 1 𝑤 2 𝑤 1 True True 𝐻 1 Username: Ortiz_Brandy Nation: USA 1 𝑤 1 Gender: female Network 2 2 𝑤 1 𝑤 2 Local 网络一致性 局部一致性 matching consistency 1 1 𝑤 2 1 𝑤 3 1 𝑤 1 𝑤 2 x True False Username: 2 3 @ortizbrandy 𝑤 1 𝑤 1 2 2 Nation: USA 𝑤 2 𝑤 1 Gender: female 𝐻 2 𝐻 3 1 1 𝑤 2 2 2 3 3 𝑤 1 𝑤 2 𝑤 3 𝑤 2 𝑤 3 x False False x 2 2 𝑤 2 𝑤 1 15

Recommend


More recommend