Mary McGlohon Graduate Assistant Machine Learning Department Carnegie Mellon University 5000 Forbes Ave., Pittsburgh, PA 15213 Office: Wean 7110- 412.268.7670 Fax: 412.268.5576 Email: mmcgloho+www at cs dot cmu dot edu

I am a Ph.D. student in the Machine Learning Department at CMU's School of Computer Science. My advisors are Christos Faloutsos and Alan Montgomery. My current research involves finding the structure of information propagation using data from blogs and optimizing pricing strategies using these findings. My primary research goal is to efficiently discover patterns in real-world data and find useful applications of these patterns. I am receiving support from a NSF Graduate Research Fellowship.

What I'm up to, Summer 2008

- Interning at Microsoft Research/Live Labs, working with Leman Akoglu and Christos Faloutsos.

Research Interests

My field-related interests include the following:

• Knowledge discovery for the Web
• Link analysis
• Evolving networks
• Social networks
• Data mining applications to epidemiology and public health
• Data mining applications to marketing
• Anomaly detection in graphs

Conference Papers

• M. McGlohon, J. Leskovec, C. Faloutsos, N. Glance, and M. Hurst. Finding patterns in blog shapes and blog evolution. International Conference on Weblogs and Social Media. Boulder, Colo., March 2007. January 2007. [pdf] Tech Report (identical): CMU-ML-07-100.

• J. Leskovec, J, M. McGlohon, C. Faloutsos, N. Glance, and M. Hurst. Patterns of Cascading Behavior in Large Blog Graphs. Society of Industrial and Applied Mathematics- Data Mining. Minneapolis, Minn., April 2007. [pdf] Tech report (12 pgs): CMU-ML-06-113 [local]

• McGlohon, M. and S. Sen. Learning to cooperate in multi-agent systems by combining Q- learning and evolutionary strategy. World Conference on Lateral Computing, December 2004. [pdf] Best Student Paper Award.

Non-Refereed Publications

• M. McGlohon and Christos Faloutsos. ADAGE: A Software Package for Analyzing Graph Evolution. Tech Report. CMU-ML-07-112. [pdf][software]

Talks

• Tutorial at ICWSM, Graph Mining Techniques for Social Media Analysis [link to VIDEO] [gzipped pdfs]. (If you would like PPT version, please email me)
• Talk at Microsoft Live Labs over work on blogs: [ppt]
• Presentation in Analysis of Social Media seminar over graph mining techniques applied to blogs: [ppt]
• Talk at Sandia Labs over SDM07 work: [ppt]

Software

• ADAGE, Matlab add-in for analyzing evolving graphs. [tar.gz] See tech report for documentation: [pdf]

Data

I have compiled data from the U.S. Federal Election Commission, representing federal electoral campaign donations from individuals to committees and committees to candidates. A description, and the files, are located here (currently available to CMU only).

CV

Last updated February 2008: [pdf].

Graduate Coursework

In reverse chronological order. * indicates core MLD PhD curriculum.

• 36-835: Models and Methods for Networks, FA07
• 10-802: Analysis of Social Media, FA07
• 15-853: *Algorithms in the Real World, FA07
• 10-702: *Statistical Machine Learning, SP07
• 10-708: Probabilistic Graphical Models, FA06
• 36-728: Time Series Analysis, FA06Q1
• 36-720: Discrete Multivariate Analysis, FA06Q2
• 15-826: *Multimedia Databases and Data Mining, SP06
• 36-626: Mathematical Statistics II, SP06
• 10-701: *Machine Learning, FA05
• 10-705: *Intermediate Statistics, FA05

Other Activities

I have performed a number of service- and otherwise-related roles for SCS. They include:

• Databases group, lab manager. Page for Machines: (local)
• Dec/5 (TG's, various other SCS social events)
• Machine Learning Dept. Recruiting (Open House, maintaining contact with prospective students)
• SIGBOVIK 2007 programming committee

Links

• Tensor Toolbox- From Tamara Kolda and team at Sandia National Labs. A tool for multidimensional arrays in MATLAB. I've found sptensor, the sparse tensor functions, especially useful.

• BlogPulse- Several tools for blog tracking and analysis. Developed by Nielsen Buzzmetrics, some friends in industry. Also check out the blog of Matthew Hurst (now at MSR), Data Mining: Text Mining, Visualization and Social Media.

• Upcoming Conferences and Workshops- a handy reference for upcoming conference deadlines in AI, Data Mining, Machine Learning, and related fields.

Misc.

• M. McGlohon. Methods and Uses of Demoralizing Graphs. SIGBOVIK. Pittsburgh, Penn., April 2007. Paper: [pdf] Talk: [ppt]

• M. McGlohon. Fried Chicken Bucket Processes. SIGBOVIK. Pittsburgh, Penn., April 2007. [pdf]

My Erdös number: <=4

Paul Erdös - Noga M. Alon - Phillip B. Gibbons - Christos Faloutsos - Mary McGlohon