CV

Education

  • PhD in Systems Engineering (Computer Science), December 2010

    • Oakland University, Rochester, MI
    • Dissertation: An Induction Method for Multilabel Decision Trees with Application to Multimedia Tagging and Other Problems
    • Advisor: Ishwar K. Sethi
  • MS in Computer Science, April 2005

    • Oakland University, Rochester, MI
  • BS in Math and Computer Science, August 2002, Summa Cum Laude

    • Eastern Michigan University, Ypsilanti, MI
    • Harvey Mudd College, Claremont, CA (1998-2000)

Awards and Honors

  • Michigan Space Grant Consortium Graduate Fellowship, 5/03-4/04
  • 1st place, Oakland University Programming Competition, Fall 2002
  • Addison Wesley Award for Outstanding Senior in Computer Science, Eastern Michigan University, 2002

Research Experience

  • Dissertation Research, Oakland University, Rochester, MI, 5/07-8/10
    • An Induction Method for Multilabel Decision Tree

      Developed an induction method for constructing multi-label decision trees from weighted labels, such as tag clouds, or from hierarchical labels. The proposed performed noticeably better on several datasets of varying media types than other pre-existing multilabel approaches for a variety of performance and loss measures. Furthermore, the proposed multilabel method was notably faster than a single label SVM approach, while performing similary in terms of classification ability.

      • A.Ma, I.K.Sethi, and N.Patel, “Multilabel Classification Method for Multimedia Tagging,” Intl. Journal of Multimedia Data Engineering and Management (IJMDEM), Volume 1, Issue 3, 2010.
      • A.Ma and I.K.Sethi, “Iterative Split Adjustment for Building Multilabel Decision Trees,” 2nd Intl. Conf. on Bioinformatics and Computational Biology (BICoB), Honolulu, Hawaii, Mar. 2010.
      • A.Ma, I.K.Sethi, and N.Patel, “Multimedia Content Tagging using Multilabel Decision Tree,” 5th IEEE Intl. Workshop on Multimedia Information Processing and Retrieval (MIPR), San Diego, CA, Dec. 2009.
  • Student Assistant (Internship), Quality Assurance/Quality Control, Production Sequencing Department, Production Genomics Facility, Department of Energy – Joint Genome Institute, Walnut Creek, CA 6/06-12/06

      Developed methods for feature extraction and analysis of 4-channel signal, DNA sequence data. Implemented a Naïve Bayesian active learning classifier for detection of production problems at JGI. Developed a controlled error rate approach to counteract differences between training set distribution and actual production distribution.

      At the time, JGI was one of only six large scale sequencing centers in the world with over 70 DNA sequencing instruments continually running; all trace production was automatically analyzed and summarized. Feature extraction methods and classifiers were prototyped in Matlab and rewritten in C++ for speed; a set of shell and perl scripts were written to automatically update a web based summary report.

  • Research Assistant, Oakland University, Rochester, MI, 8/02-5/06
    • Distributed K-Median Clustering”, 12/05-5/07

      Developed approximate methods to compute the spatial median of a set of samples. Used the methods to perform K-median clustering for vector data and K-clustroid clustering for non-vector data in a data distributed environment. Test application involved image clustering.

      • A.Ma and I.K.Sethi, “Region Based Image Clustering using Distributed K-Median Clustering,” Intl. Conf. on Distributed Multimedia Systems (DMS), Redwood City, CA, Sept. 2007.
      • A.Ma and I.K.Sethi, “Distributed K-Median Clustering with Application to Image Clustering,” Intl. Workshop on Pattern Recognition in Information Systems (PRIS), Funchal, Madeira, Portugal, June 2007.
    • Confidence Based Classifier Design for Image Segmentation”, 12/04-5/06

      Expanding and applying active learning and confidence based classifiers for image segmentation in digital photographs and in medical areas such as brain tumor segmentation and chroma analysis for quantitative immunohistochemistry. Developed in Matlab with interfacing to external programs.

      • A.Ma, N.Patel, M.Li, and I.K.Sethi, “Confidence Based Active Learning for Whole Object Image Segmentation,” Intl. Workshop on Multimedia Content Representation, Classification, and Security (MCRCS), Istanbul, Turkey, Sept. 2006.
      • N.Patel, A.Ma, R.Shah, and I.K.Sethi, “Chroma Analysis for Quantitative Immunohistochemistry using Active Learning,” SPIE Symposium on Medical Imaging - Image Processing, San Diego, CA, Feb. 2006.
    • Indexing and Retrieval of Satellite Images”, 8/03-12/04

      Developed a local association, shape-based method for content based indexing and retrieval of infrared satellite images. On average, retrieval from a database of 1965 shapes took 0.7 seconds on a P4 PC. Retrieval speed was determined to be more dependent on the degree of representation than on the size of the database. Developed in a combination of Matlab and C. Partially supported by a graduate fellowship from the Michigan Space Grant Consortium.

      • A.Ma and I.K.Sethi, “Local Shape Association Based Retrieval of Infrared Satellite Images,” IEEE Intl. Workshop on Multimedia Information Processing and Retrieval (MIPR), Irvine, CA, Dec. 2005.
      • Aiyesha Ma. “Indexing and Retrieval of Satellite Imaging”. Thesis, Master of Science in Computer Science and Engineering, Oakland University, Rochester, MI. 2005.
    • Predicting Post-Operative Risks in Cardiac Surgery”, 1/03-8/03

      Predicted 7 different post-operative risks using a Näive Bayesian model and a multi-feature-split decision tree. Developed a combined classifier that resulted in improved classification accuracy. The combined classifier predicted post-operative death with a C-statistic of 0.78; an improvement over published C-statistics ranging from 0.65 to 0.75. Developed in Matlab. This work was a joint project with Mandala Sciences, Inc. and Health Data Research, Inc. and supported with a grant from the National Institute of Aging.

      • A.Ma, I.K.Sethi, M.H.Schwartz, and U.S.Page,“Using Multiple Classifiers to Improve Classification of Post Operative Complications in Cardiac Surgery,” Medinfo 2004, San Francisco, CA, Sept. 2004.
    • Analysis of Glass Coating Process”, 8/02-12/02

      Assisted in database cleaning, data preprocessing, and the collection of statistical information for a data mining project to predict glass coating quality. Software included SQL Server, MS Access, SPSS, and SAS. Work performed for and supported by Guardian Industries.

  • Student Researcher, UnCoRe Research Experience for Undergraduates, Oakland University, Rochester, MI, 5/02-7/02
    • Hausdorff Based Vector Quantization for Binary Images

      Worked with a partner to develop a method for comparing binary line drawings according to shape. The method was able to distinguish between images of man-made objects and images of natural objects. This project was conducted under a research experience for undergraduates program sponsored by the National Science Foundation.

      • A.Ma, R.Mukhopadhyay, and I.K.Sethi, “Hausdorff Metric Based Vector Quantization of Binary Images,” Information and Knowledge Engineering Conf., Las Vegas, NV, July 2003.
      • R.Mukhopadhyay, A.Ma, and I.K.Sethi, “Pathfinder Networks for Content Based Image Retrieval Based On Automated Shape Feature Discovery,” Sixth IEEE Intl. Symposium on Multimedia Software Engineering (ISMSE 2004), Miami, FL, Dec. 2004.

Teaching Experience

  • Teaching Assistant, Oakland University, Rochester, MI, 1/07-5/07

      Teaching Assistant for online section of Introduction to Computer Programming using Excel. Responsible for holding regular office hours, answering student questions, and grading homework and exams.

  • Math Teacher, Clonlara High School, Ann Arbor, MI, 8/00-5/02

      Responsibilities included helping students learn topics such as algebra and geometry, designing the math curriculum, choosing which books to use, and creating a learning plan for each student.

  • Supplemental Instructor, Eastern Michigan University, Ypsilanti, MI, 8/01-4/02

      Responsible for helping students improve their grades in an introductory level geography course, by preparing worksheets and leading study sessions. Was able to help students improve their test grades as much as 40%.

Other Experience

  • System Verification and Test Engineer, Contract, Pacific Biosciences, Menlo Park, CA, 6/11-12/11

      Worked with mechanical engineers to verify hardware changes to the RS (DNA sequencing instrument) and wrote tests to characterize whether instrument functions were within acceptable tolerances. Worked with software engineers to validate and verify new software features. Performed verification tests, and reported hardware and software bugs.

      Performed triaging of lab instrument failures. Worked with manufacturing and field service to troubleshoot manufacturing floor and remote instrument problems.

      Refactored test suite, and modified PyUnit variant framework to support differences in real and simulated instruments. Wrote and modified tests within test suite to support new field service requirements and use cases. Instrument tests and scripts written in Python.

  • Program Coordinator, UnCoRe Research Experience for Undergraduates, Oakland University, Rochester, MI, May-July, 2003, 2004, 2005

      Responsible for organization and planning of events over the course of the summer. Maintained and revamped website. Often led discussions during weekly coffee hours and assisted the students as needed.

      In 2005, helped advise one project. Student's abstract was accepted to the National Conference for Undergraduate Research. Program funded by the National Science Foundation.

Selected Class Research Projects

  • Video Object Segmentation using MPEG Motion Compensation Vectors, Advanced Visual Computing (Group Project)
  • MPEG Video Retrieval Using Motion Information, Advanced Visual Computing (Group Project) (SPIE Abstract)
  • Indexing and Retrieval of Offline Handwritten Notes, Information Retrieval
  • Visualization of Lace-like Materials, Advanced Computer Graphics. (Project Paper)
  • Predicting High Risk Loan Customers using a Multifeature Split Decision Tree, Data Mining
  • Composition of 4-part Music Harmony using a Genetic Algorithm, Heuristics
  • Review of Parallel Data Mining Algorithms for Distributed Machines, Parallel and Distributed Processing (No project, report and presentation only. Final Report)

Service and Volunteer Activities

  • ACM-W, Oakland University Chapter Founding President 4/03-10/04

  • ACM, Oakland University Chapter Executive Board 9/03-4/05

  • ACM, OU Team Coach for East Central North American Regional Programming Competition 2003

  • Mark Twain Elementary, Saturday Math Program Volunteer, 2003

  • Golden Key Honor Society, Eastern Michigan University Chapter Webmaster 3/01-4/02

  • SWE, Harvey Mudd College Section Co-President 8/99-5/00

Affiliations

  • Society for Women Engineers (SWE), 8/98-6/07
  • Golden Key International Honour Society, 3/01-present
  • Association for Computing Machinery (ACM), 6/02-6/07, 1/11-present
  • Association for Computing Machinery - Women in Computing (ACM-W), 4/03-4/06