1. 1 Truth Validation and Veracity Analysis with Information Networks Jiawei Han Data Mining Group, Computer Science University of Illinois at Urbana-Champaign www.cs.uiuc.edu/~hanj May 22, 2015

2. 2 Outline TruthFinder: Tuth Validation by Information Network Analysis Beyond TruthFinder: Multiple Versions of Truth and Evolution of Truth Enhancing Truth Validation by InfoNet Analysis: The RankClus & NetClus Methodology Summary

3. 3 Motivation Why truth validation and veracity analysis? Information sharing Sharing trustable, quality information Identifying false information among many conflicting ones Information security Protecting trustable information and its sources Identifying which information providers are suspicious ones: frequently providing false information Tracing back suspicious information providers via information networks

4. 4 Truth Validation and Veracity Analysis by Information Network Analysis The trustworthiness problem of the web (according to a survey): 54% of Internet users trust news web sites most of time 26% for web sites that sell products 12% for blogs TruthFinder: Truth discovery on the Web by link analysis Among multiple conflict results, can we automatically identify which one is likely the true fact? Veracity (conformity to truth): Given a large amount of conflicting information about many objects, provided by multiple web sites (or other information providers), how to discover the true fact about each object? Xiaoxin Yin, Jiawei Han, Philip S. Yu, “Truth Discovery with Multiple Conflicting Information Providers on the Web”, TKDE’08

5. 5 Conflicting Information on the Web Different websites often provide conflicting info. on a subject, e.g., Authors of “Rapid Contextual Design ” Online StoreAuthors Powell’s booksHoltzblatt, Karen Barnes & NobleKaren Holtzblatt, Jessamyn Wendell, Shelley Wood A1 BooksKaren Holtzblatt, Jessamyn Burns Wendell, Shelley Wood Cornwall booksHoltzblatt-Karen, Wendell-Jessamyn Burns, Wood Mellon’s booksWendell, Jessamyn Lakeside booksWENDELL, JESSAMYNHOLTZBLATT, KARENWOOD, SHELLEY Blackwell onlineWendell, Jessamyn, Holtzblatt, Karen, Wood, Shelley

6. 6 Mapping It to Information Networks Each object may have a set of conflicting facts E.g., different author names for a book And each web site provides some facts How to find the true fact for each object? w1w1 f1f1 f2f2 f3f3 w2w2 w3w3 w4w4 f4f4 f5f5 Web sitesFacts o1o1 o2o2 Objects

7. 7 Basic Heuristics for Problem Solving 1. There is usually only one true fact for a property of an object 2. This true fact appears to be the same or similar on different web sites E.g., “Jennifer Widom” vs. “J. Widom” 3. The false facts on different web sites are less likely to be the same or similar False facts are often introduced by random factors 4. A web site that provides mostly true facts for many objects will likely provide true facts for other objects

8. 8 Mutual Consolidation between Confidence of Facts and Trustworthiness of Providers Confidence of facts ↔ Trustworthiness of web sites A fact has high confidence if it is provided by (many) trustworthy web sites A web site is trustworthy if it provides many facts with high confidence The TruthFinder mechanism, an overview: Initially, each web site is equally trustworthy Based on the above four heuristics, infer fact confidence from web site trustworthiness, and then backwards Repeat until achieving stable state

9. 9 Analogy to Authority-Hub Analysis Facts ↔ Authorities, Web sites ↔ Hubs Difference from authority-hub analysis Linear summation cannot be used A web site is trustable if it provides accurate facts, instead of many facts Confidence is the probability of being true Different facts of the same object influence each other w1w1 f1f1 Web sitesFacts HubsAuthorities High trustworthiness High confidence

10. 10 Inference on Trustworthness Inference of web site trustworthiness & fact confidence w1w1 f1f1 f2f2 w2w2 w3w3 w4w4 f4f4 Web sitesFacts o1o1 o2o2 Objects f3f3 True facts and trustable web sites will become apparent after some iterations

11. 11 Computation Model: t(w) and s(f) The trustworthiness of a web site w : t(w) Average confidence of facts it provides The confidence of a fact f : s(f) One minus the probability that all web sites providing f are wrong w1w1 f1f1 w2w2 t(w1)t(w1) t(w2)t(w2) s(f1)s(f1) Sum of fact confidence Set of facts provided by w Probability that w is wrong Set of websites providing f

12. 12 Experiments: Finding Truth of Facts Determining authors of books Dataset contains 1265 books listed on abebooks.com We analyze 100 random books (using book images) CaseVotingTruthFinderBarnes & Noble Correct718564 Miss author(s)1224 Incomplete names1856 Wrong first/middle names113 Has redundant names0223 Add incorrect names155 No information002

13. 13 Experiments: Trustable Info Providers Finding trustworthy information sources Most trustworthy bookstores found by TruthFinder vs. Top ranked bookstores by Google (query “bookstore”) Bookstoretrustworthiness#bookAccuracy TheSaintBookstore0.971280.959 MildredsBooks0.969101.0 Alphacraze.com0.968130.947 BookstoreGoogle rank#bookAccuracy Barnes & Noble1970.865 Powell’s books3420.654 TruthFinder Google

14. 14 Outline TruthFinder: Tuth Validation by Information Network Analysis Beyond TruthFinder: Multiple Versions of Truth and Evolution of Truth Enhancing Truth Validation by InfoNet Analysis: The RankClus & NetClus Methodology Summary

15. 15 Beyond TruthFinder: Extensions Limitations of TruthFinder: Only one version of truth But people may have different, contrasting opinions Not consider the time factor But truth may change with time, e.g., Obama’s status in 2008 and 2009 Needed Extensions Multiple versions of truth or opinions Evolution of truth Philosophy Truth is a relative, evolving, and dynamically changing judgment

16. 16 Multiple Versions of Truth Watch out of copy-cats! Copy-cat: Some information providers or even new agencies simply copy each other Falsity could be amplified by copy-cats How to judge copy-cats: Always copying in certain dimensional space Treat copy-cats as one instead of multiples w1w1 f1f1 f2f2 f3f3 w2w2 w3w3 w4w4 f4f4 f5f5 Web sitesFacts o1o1 o2o2 Objects Statements can be clustered into multiple centers False statements: still diverse, spread, and lack of converge Statements could be clustered based on different dimensional space (context), e.g., Java

17. 17 Transition/Evolution of Truth Truth is not static: It changes dynamically with time Associating different versions of truth with different time periods Clustering statements based on time durations Statements Identifying clusters (density-based clustering) Distinguishing time-based clusters from outliers Information providers Leaders, followers, and old-timers Information-network based ranking and clustering Powerful analysis by information network analysis

18. 18 Outline TruthFinder: Tuth Validation by Information Network Analysis Beyond TruthFinder: Multiple Versions of Truth and Evolution of Truth Enhancing Truth Validation by InfoNet Analysis: The RankClus & NetClus Methodology Summary

19. 19 Why RankClus? More meaningful cluster Within each cluster, ranking score for every object is available as well More meaningful ranking Ranking within a cluster is more meaningful than in the whole network Address the problem of clustering in heterogeneous networks No need to compute pair-wise similarity of objects Mapping each object into a low measure space What type of objects to be clustered: Target objects (specified by user) Clustering of target objects can induce a sub-network of the original network

20. 20 Algorithm Framework - Illustration Sub-Network Ranking Clustering

21. 21 Algorithm Framework - Summary Step 0. Initialization Randomly partition target objects into K clusters Step 1. Ranking Ranking for each sub-network induced from each cluster, which serves as feature for each cluster Step 2. Generating new measure space Estimate mixture model coefficients for each target object Step 3. Adjusting cluster Step 4. Repeat Step 1-3 until stable

22. 22 Focus on A Bi-type Network Case Conference-author network, links can exist between Conference (X) and author (Y) Author (Y) and author (Y) Use W to denote the links and there weights W =

23. 23 Step 1: Feature Extraction — Ranking Simple Ranking Proportional to degree counting for objects E.g., number of publications of authors Considers only immediate neighborhood in the network Authority Ranking Extension to HITS in weighted bi-type network Rules: Rule 1: Highly ranked authors publish many papers in highly ranked conferences Rule 2: Highly ranked conferences attract many papers from many highly ranked authors Rule 3: The rank of an author is enhanced if he or she co- authors with many authors or many highly ranked authors

24. 24 Rules in Authority Ranking Rule 1: Highly ranked authors publish many papers in highly ranked conferences Rule 2: Highly ranked conferences attract many papers from many highly ranked authors Rule 3: The rank of an author is enhanced if he or she co- authors with many authors or many highly ranked authors

25. 25 Example: Authority Ranking in the 2- Area Conference-Author Network Given the correct cluster, the ranking of authors are quite distinct from each other

26. 26 Example: 2-D Coefficients in the 2- Area Conference-Author Network The conferences are well separated in the new measure space Scatter plots of two conferences and component coefficients

27. 27 A Running Case Illustration for 2-Area Conf-Author Network Initially, ranking distributions are mixed together Two clusters of objects mixed together, but preserve similarity somehow Improved a little Two clusters are almost well separated Improved significantly Stable Well separated

28. 28 Time Complexity Analysis At each iteration, |E|: edges in network, m: number of target objects, K: number of clusters Ranking for sparse network ~O(|E|) Mixture model estimation ~O(K|E|+mK) Cluster adjustment ~O(mK^2) In all, linear to |E| ~O(K|E|)

29. 29 Case Study: Dataset: DBLP All the 2676 conferences and 20,000 authors with most publications, from the time period of year 1998 to year 2007. Both conference-author relationships and co-author relationships are used. K=15

30. 30 Beyond RankClus: A NetClus Model RankClus combines ranking and clustering successfully to analyze information networks A study on how ranking and clustering can mutually reinforce each other in information network analysis RankClus works well on bi-typed information networks Extension of bi-type network model to star-network model DBLP: Author - paper - conference - title (subject) AuthorConference Subject Paper

31. 31 NetClus: Database System Cluster database 0.0995511 databases 0.0708818 system 0.0678563 data 0.0214893 query 0.0133316 systems 0.0110413 queries 0.0090603 management 0.00850744 object 0.00837766 relational 0.0081175 processing 0.00745875 based 0.00736599 distributed 0.0068367 xml 0.00664958 oriented 0.00589557 design 0.00527672 web 0.00509167 information 0.0050518 model 0.00499396 efficient 0.00465707 Surajit Chaudhuri 0.00678065 Michael Stonebraker 0.00616469 Michael J. Carey 0.00545769 C. Mohan 0.00528346 David J. DeWitt 0.00491615 Hector Garcia-Molina 0.00453497 H. V. Jagadish 0.00434289 David B. Lomet 0.00397865 Raghu Ramakrishnan 0.0039278 Philip A. Bernstein 0.00376314 Joseph M. Hellerstein 0.00372064 Jeffrey F. Naughton 0.00363698 Yannis E. Ioannidis 0.00359853 Jennifer Widom 0.00351929 Per-?ke Larson 0.00334911 Rakesh Agrawal 0.00328274 Dan Suciu 0.00309047 Michael J. Franklin 0.00304099 Umeshwar Dayal 0.00290143 Abraham Silberschatz 0.00278185 VLDB 0.318495 SIGMOD Conf. 0.313903 ICDE 0.188746 PODS 0.107943 EDBT 0.0436849 Ranking authors in XML

32. 32 Outline TruthFinder: Tuth Validation by Information Network Analysis Beyond TruthFinder: Multiple Versions of Truth and Evolution of Truth Enhancing Truth Validation by InfoNet Analysis: The RankClus & NetClus Methodology Summary

33. 33 Summary Progress Highlights 3 PhD graduated in 2009 Currently over 20 Ph.D.s working on closely related projects Attract more funded projects: 3 NSFs, NASA, DHS, … Industry collaborations: Microsoft Research, IBM Research, Boeing, HP Labs, Yahoo!, Google, … Research papers published in 2008 & 2009: 8 journal papers and 53 conference papers, including KDD, NIPS, SIGMOD, VLDB, ICDM, SDM, ICDE, ECML/PKDD, SenSys, ICDCS, IJCAI, AAAI, Discovery Science, PAKDD, SSDBM, ACM Multimedia, EDBT, CIKM, … Truth validation by information network analysis: A promising direction: TruthFinder, iNextCube, and beyond Knowledge is power, but knowledge is hidden in massive links Integration of data mining with the project: Much more to be explored!

34. 34 Recent Publications Related to the Talk X. Yin, J. Han, and P. S. Yu, “Truth Discovery with Multiple Conflicting Information Providers on the Web”, TKDE’08 Y. Sun, J. Han, P. Zhao, Z. Yin, H. Cheng, T. Wu, “RankClus: Integrating Clustering with Ranking for Heterogeneous Information Network Analysis”, EDBT'09 Y. Sun, Y. Yu, and J. Han, “Ranking-Based Clustering of Heterogeneous Information Networks with Star Network Schema", KDD'09 Y. Sun, J. Han, J. Gao, and Y. Yu, “iTopicModel: Information Network- Integrated Topic Modeling", ICDM'09 J. Han, “Mining Heterogeneous Information Networks by Exploring the Power of Links", Discovery Science'09 (Invited Keynote Speech) M.-S. Kim and J. Han, “A Particle-and-Density Based Evolutionary Clustering Method for Dynamic Networks", VLDB'09 Y. Yu, C. Lin, Y. Sun, C. Chen, J. Han, B. Liao, T.Wu, C. Zhai, D. Zhang, and B. Zhao, “iNextCube: Information Network-Enhanced Text Cube", VLDB'09 (system demo).