JOB TARGET

 

Seeking position utilizing my strong background in statistical and probabilistic methods to research challenging problems and applications in real-world domains.

 

EDUCATIONAL BACKGROUND

 

·         Ph.D. Graduated with distinction.  Computer Science Department, University of New Mexico.  July 2007.  (3.98 GPA)

·         M.S. in Computer Science, University of New Mexico, May 2002. (3.90 GPA). 

·         B.S. in Computer Science, Mathematics minor, cum laude, University of New Mexico, December 2000. (3.66 GPA)

 

RESEARCH INTERESTS

 

·        Learning in graphical models such as Bayesian networks, Markov random fields, chain graphs, factor graphs and deterministic finite automata.

·        Applications in challenging real-world domains such as image analysis, neuro and bioinformatics, network security, traffic flow, etc.

·        Extension of propositional graphical models to relational domains.

·        Sampling methods such as Markov chain Monte Carlo and particle filters

·        Spectral methods such as ICA, PCA and eigenvector decomposition.

·        A broad interest in statistical and probabilistic modeling methods in general

 

RESEARCH EXPERIENCE

 

·         Knowledge of many methods in Machine Learning, a challenging and mathematically rigorous field.  In depth knowledge of Bayesian networks.

·         Introduced a class-discriminative Bayesian network scoring function that is faster than other class-discriminative scores while still yielding networks with higher classification accuracy than commonly employed generative scores [6].

·         Introduced hierarchical methods for Bayesian network structure search which find higher scoring topologies [5] and parameters [4] than traditional search methods.

·         Proposed a novel approach for analyzing neuroimaging data with discrete random variables that is capable of modeling non-linear relationships missed by most neuroimaging techniques [9][10][11][12].

·         Introduced a method to elicit deterministic finite automata (DFAs) for modeling non-stationary HTTP requests.  ROC analysis revealed attack recognition superior to current methods [1].

·         Developed a virtual laboratory for multi-physics robotic simulation [2][7].

·         Demonstrated the use of stochastic learning automata for robotic control [2].

·         Developed a statistical visualization package for tracking damaged US roads [3].

PREVIOUS WORK EXPERIENCE

 

University of New Mexico                                                2000-present                Albuquerque, NM

Graduate student for professors Koon Chua (Civil Engineering), Mohammad Jamshidi (Electrical Engineering), Stephanie Forrest (Computer Science) and Terran Lane (Machine Learning).  Yielded publications with every department I worked with.

 

Science and Engineering Associates                          1998-2000         Albuquerque, NM

Graphical user interface developer and tester for Microsoft applications.

 

Daifuku:                                                                                1997-1998         Albuquerque, NM

Lead developer for database applications.

 

SOFTWARE DEVELOPMENT AND MISCELLANEOUS QUALIFICATIONS

 

·         Extensive knowledge of C++ and the standard template library (STL)

·         Java, Matlab, Perl, C

·         Win32 API programming and MS Development Studio IDEs

·         Microsoft Foundation Classes (MFC)

·         Familiarity with Scheme, Lisp, ML, Prolog, CORBA, OpenGL

·         Windows and Linux operating systems

·         Excellent literary composition and presentation skills

 

PUBLICATIONS

Peer Reviewed Journals

[1]       (2007) Discrete dynamic Bayesian Network Analysis of fMRI data.  Human Brain Mapping, to appear.  http://www3.interscience.wiley.com/cgi-bin/abstract/116838597/ABSTRACT

[2]       (2006) Modeling web requests with finite automata. Ingham, K., Somayaji, A., Burge, J., Forrest, S. To appear in Journal on Computer Networks.

[3]       (2002) V-Lab-a virtual laboratory for autonomous agents-SLA-based learning controllers.  El-Osery, A.I., Burge, J., Jamshidi, M., Saba, A., Fathi, M., Akbarzadeh-T, M.-R., Systems, Man and Cybernetics, Part B, IEEE Transactions on, V: 32 I: 6, Dec. pg. 791 -803

[4]       (2001) Virtual Environment for Transportation Data Management System. Chua, K. M., McKeen, G., Burge, J., Luger, G.  Journal of the Transportation Research Board, No. 1764, pg. 164-175.

Full Paper Reviewed Conferences

[5]       (2007) Learning Bayesian Network Structures with Shrinkage Parameter Estimates.  Burge, J., Terran, L.  The 18th European Conference on Machine Learning, Warsaw, Poland. (24% acceptance rate).

[6]       (2006) Improving Bayesian Network Structure Search with Random Variable Aggregation Hierarchies.  Burge, J., Lane, T.  17th European Conference on Machine Learning, Berlin, Germany  (20% acceptance rate).

[7]       (2005) Learning Class-Discriminative Dynamic Bayesian Networks.  (2005) Burge, J.  Lane, T.  International Conference on Machine Learning, Bonn, Germany (27% acceptance rate).

[8]       (2001) A Discrete Event Systems Approach to a Virtual Laboratory for Distributed Robotic Agents. (2001)  Burge, J., El-Osery, A., Fathi, M., Jamshidi, M., and Mallipeddi, S. IEEE Systems, Man and Cybernetics Conference, Tucson, Arizona.

Workshop Papers

[9]       (2005) Comprehensibility of Generative vs. Class Discriminative Dynamic Bayesian Multinets. Burge, J.  Lane, T.  AAAI.,  "Human Comprehensibility" workshop.

Abstract Reviewed Conferences

[10]   (2005)  Bayesian Network Analysis of Neuroanatomical Data.  Burge, J.  Lane, T.  First CS UNM Student Conference (CSUSC).  Albuquerque, New Mexico.

[11]   (2005)  Dynamic Bayesian Network Classification of fMRI Data Reveals Altered Functional Connectivity in Dementia.  Burge, J.  Lane, T.  Clark, V.  Organization of Human Brain Mapping.

[12]   (2004) Dynamic Bayesian Network Classification of fMRI Data Reveals Enhanced Amygdala Connectivity in Dementia.  Burge, J., Lane, T., Clark, V.P. Society for Neuroscience.

Invited Addresses

[13]   (2006) Discrete Bayesian Network Structure Search with an Application to fMRI Data.  Presented to the University of Kentucky, Lexington, Kentucky, Machine Learning Seminar.

Poster Presentations

[14]   (2004) Bayesian Classification of FMRI Data: Evidence for Altered Neural Networks in Dementia. Burge, J., Clark, V.P., Lane, T.  International Conference on Machine Learning.

[15]   (2003) Incorporating Shrinkage and DBNs into fMRI Classification. Burge, J., Lane, T.  International Conference on Machine Learning.

Technical Reports

[16]   (2007) Selecting Bayesian Network Parameterizations for Generating Simulated Data.  Burge, J., Terran, L. (Number pending).

[17]   (2005) Dynamic Bayesian Networks: Class Discriminative Structure Search with an Application to Functional Magnetic Resonance Imaging Data.  Burge, J.  TR-CS-2005-24, University of New Mexico.

[18]   (2004)  Evidence for Altered Neural Networks in Dementia. Burge, J., Clark, V.P., Lane, T., Link, H., Qiu, S.  TR-CS-2004-28, University of New Mexico

Community Involvement

Reviewer: Human Brain Mapping 2005, ICML 2005, 2007, AAAI 2007, Trans. Neural Systems and Rehab. Eng. 2007

Student scholarships: International Conference Machine Learning: 2003, 2004, 2005