TITLE: Mining large graphs and streams using matrix and tensor tools

INSTRUCTOR:
        Christos Faloutsos, CMU
        Tammy Kolda, Sandia National lab
        Jimeng Sun, CMU

INTENDED DURATION: 3 hours

DESCRIPTION - OBJECTIVES

How can we find patterns in sensor streams (eg., a
sequence of temperatures, or water-pollutant measurements,
machine room measurements)?
How to mine internet traffic graph over time?
How to make the process incremental?
We review the state of the art in four related fields:
(a) numerical analysis and linear algebra (b) multi-linear/tensor analysis
(c) graph mining and (d) stream mining.
We will present both the theoretical results/algorithms and
case study on several real applications.


CONTENT AND OUTLINE

Note to evaluators: We are asking for a 3-hour duration, 
to include all parts. However, we could easily create a 2-hour version,
by omitting Part II below.


[Part I. Core]
Data model - Fundamental concepts (3'-5')
- Time series
- Matrices
- Tensors

Matrix analysis (12'-30')
- SVD, PCA and eigen-decomposition 
- Page-rank, HITS
- Independent Component Analysis (ICA)
- sparse decompositions: CUR
- Semi-discrete decomposition (SDD)
- Co-clustering

Tensor analysis (30'-35')
- Intro (15')
- Parafac (5')
- Tucker Model (5'-10')
	Tucker 1 and PCA; 
	Tucker 2 and Tensor PCA; 
	Tucker 3 and High-order SVD (HO-SVD)
- Other models (5'-10')
        Combination of PARAFAC and Tucker
        DEDICOM

[II. Extension - optional]
Nonnegativity (15'-20')
- Nonnegative matrix factorization
- Nonnegative tensor factorization

Miss values (15'- 20')
- Matrices
- Tensors

Stream mining (10'-15')
- Incremental PCA
- Dynamic tensor analysis
- Window-based tensor analysis

[III. Practice]
Software (10')
- Intro
- Issues
	Scalability, Accuracy, Sparsity

Case study (10'-15')
- sensor network, machine monitoring
- internet forensic computing
- social network analysis 
- web graph study 

WHO SHOULD ATTEND

Researchers that want to get up to speed with the major tools
in stream mining, graph mining. 
Also, practitioners who want a concise, intuitive
overview of the state of the art.


RELATED PREVIOUS TUTORIALS (if any)

None

ABOUT THE INSTRUCTORS


Christos Faloutsos is a Professor at Carnegie Mellon University.
    He has received the Presidential Young Investigator Award by
    the National Science Foundation (1989), seven ``best paper'' awards,
    and several teaching awards.
    He has served as a member of the executive committee of SIGKDD;
    he has published over 140 refereed articles, one monograph, 
    and holds five patents.  His research interests include data mining
    for streams and networks, fractals, indexing for 
    multimedia and bio-informatics data bases, and performance.

Tamara Kolda ............

Jimeng Sun ............