1. Spotting Culprits in Epidemics: How many and Which ones? B. Aditya Prakash Virginia Tech Jilles Vreeken University of Antwerp Christos Faloutsos Carnegie Mellon University IEEE ICDM Brussels December 11, 2012

2. Contagions Social collaboration Information Diffusion Viral Marketing Epidemiology and Public Health Cyber Security Human mobility Games and Virtual Worlds Ecology Localized effects: riots…

3. Virus Propagation Susceptible-Infected (SI) Model [AJPH 2007] CDC data: Visualization of the first 35 tuberculosis (TB) patients and their 1039 contacts Diseases over contact networks Prakash, Vreeken, Faloutsos 2012 β

4. Outline Motivation---Introduction Problem Definition Intuition MDL Experiments Conclusion Prakash, Vreeken, Faloutsos 2012

5. Culprits: Problem definition 2-d grid Q: Who started it? Prakash, Vreeken, Faloutsos 2012

6. Culprits: Problem definition Prakash, Vreeken, Faloutsos 2012 Prior work: [Lappas et al. 2010, Shah et al. 2011] 2-d grid Q: Who started it?

7. Outline Motivation---Introduction Problem Definition Intuition MDL Experiments Conclusion Prakash, Vreeken, Faloutsos 2012

8. Culprits: Exoneration Prakash, Vreeken, Faloutsos 2012

9. Culprits: Exoneration Prakash, Vreeken, Faloutsos 2012

10. Who are the culprits Two-part solution – use MDL for number of seeds – for a given number: exoneration = centrality + penalty Running time = – linear! (in edges and nodes) Prakash, Vreeken, Faloutsos 2012 NetSleuth

11. Outline Motivation---Introduction Problem Definition Intuition MDL – Construction – Opitimization Experiments Conclusion Prakash, Vreeken, Faloutsos 2012

12. Modeling using MDL Minimum Description Length Principle == Induction by compression Related to Bayesian approaches MDL = Model + Data Model – Scoring the seed-set Number of possible |S|- sized sets En-coding integer |S| Prakash, Vreeken, Faloutsos 2012

13. Modeling using MDL Data: Propagation Ripples Original Graph Infected Snapshot Ripple R2 Ripple R1 Prakash, Vreeken, Faloutsos 2012

14. Modeling using MDL Ripple cost Total MDL cost How the ‘frontier’ advances How long is the ripple Ripple R Prakash, Vreeken, Faloutsos 2012

15. Outline Motivation---Introduction Problem Definition Intuition MDL – Construction – Opitimization Experiments Conclusion Prakash, Vreeken, Faloutsos 2012

16. How to optimize the score? Two-step process – Given k, quickly identify high-quality set – Given these nodes, optimize the ripple R Prakash, Vreeken, Faloutsos 2012

17. Optimizing the score High-quality k-seed-set – Exoneration Best single seed: – Smallest eigenvector of Laplacian sub-matrix – Analyze a Constrained SI epidemic Exonerate neighbors Repeat Prakash, Vreeken, Faloutsos 2012

18. Optimizing the score Optimizing R – Get the MLE ripple! Finally use MDL score to tell us the best set NetSleuth: Linear running time in nodes and edges Prakash, Vreeken, Faloutsos 2012 Ripple R

19. Outline Motivation---Introduction Problem Definition Intuition MDL Experiments Conclusion Prakash, Vreeken, Faloutsos 2012

20. Experiments Evaluation functions: – MDL based – Overlap based (JD == Jaccard distance) Closer to 1 the better Prakash, Vreeken, Faloutsos 2012 How far are they?

21. Experiments: # of Seeds One SeedTwo Seeds Three Seeds

22. Experiments: Quality (MDL and JD) Prakash, Vreeken, Faloutsos 2012 Ideal = 1 One SeedTwo Seeds Three Seeds

23. Experiments: Quality (Jaccard Scores) Prakash, Vreeken, Faloutsos 2012 Closer to diagonal, the better True NetSleuth One SeedTwo Seeds Three Seeds

24. Experiments: Scalability Prakash, Vreeken, Faloutsos 2012

25. Outline Motivation---Introduction Problem Definition Intuition MDL Experiments Conclusion Prakash, Vreeken, Faloutsos 2012

26. Conclusion Given: Graph and Infections Find: Best ‘Culprits’ Two-part solution – use MDL for number of seeds – for a given number: exoneration = centrality + penalty NetSleuth: – Linear running time in nodes and edges Prakash, Vreeken, Faloutsos 2012

27. B. Aditya Prakash http://www.cs.vt.edu/~badityap Any Questions? Prakash, Vreeken, Faloutsos 2012