Harish Srinivasan

HARISH SRINIVASAN

hs32@cedar.buffalo.edu

Ph: 716 310 0758

INTERESTS

Machine learning concepts such as Condition Random Fields, Dirichlet processes with Infinite mixture models, Gaussian Processes, Statistical learning.
Application of machine learning in Information Extraction tasks.

WORK EXPERIENCE

Research Scientist at Janya Inc (Mar '08-Present)
Research Assistant at CEDAR (Dec '03-Aug '04,Dec '04-Aug '05,Dec '05-Aug '06, Dec '07-Present) (Center of Excellence for Document Analysis and Recognition) SUNY Buffalo. CEDAR-FOX: (Material/Apparatus for Verifying/Analyzing Handwritten/Bio-metric samples. U.S. Patent). Developed a software named CEDAR-FOX in working with a team at CEDAR, SUNY Buffalo that provides for a number of applications such as Handwriting verification and identification, legibility and reading analysis, signature verification, search engines for handwritten documents, word and character recognition. The various modules involve extensive statistical modeling, machine learning and implementation of a variety of parametric and non-parametric classifiers. (C, VC++, MFC).
FINGERPRINTS: Working on fingerprint verification, individuality models and partial, missing and noisy finger print analysis using statistical machine learning techniques
Graduate Teaching Assistant: Machine Learning (Aug’04 – Dec ‘04, Aug '05- Dec ‘05,Aug '06- Dec '06) Assisted students with design algorithms for projects involving machine learning concepts.
Marfic Technologies (May – Aug’03): Developed a front end web interface using PERL CGI to link with backend database using MySQL queries

EDUCATION

PhD., Computer Science & Engineering, University at Buffalo, State University of New York, February’08
M.S in Electrical Engineering, University at Buffalo, State University of New York, February ’04 (GPA:3.5)
B.Tech in Information Technology, Madras University, India, Sept ’02 (GPA:3.8)

Machine learning approaches for person identification and verification: (C3I Technologies conference proceedings for the March 2005 Defense & Security Symposium Orlando.)
Comparison of Classification Methods for Fingerprint Verification for Variable Number of Minutiae (Biometric Technology for Human Identification III: SPIE Defense and Security Symposium April 17-22, 2006)
Spotting Words in Handwritten Arabic Documents. (Document Recognition and Retrieval XIII SPIE, San Jose, CA January 2006).

Biometric and Forensic Aspects of Digital Document Processing (Digital Document Processing, B. B. Chaudhuri (ed.), Springer, 2006.)
Automatic Scoring of Handwritten Essays using Latent Semantic Analysis (Document Analysis Systems Springer, Nelson, New Zealand February 2006).
Signature verification using the Kolmogorov-Smirnov statistic: (Proc. Int. Graphonomics Society Conference Salerno, Italy, June 2005, pp. 152-156)
A Search Engine for Handwritten Documents: (Document Recognition and Retrieval XII SPIE, San Jose, CA January 21, 2005.)
Content Based Retrieval of Handwritten Documents: (Knowledge-Based Computer Systems (KBCS) Hyderabad, India December 21, 2004.)

PROJECTS

Machine Learning: Machine learning strategies for classifying E-Mail Spam using Naïve Bayes, Bayesian Neural network classifiers. (MATLAB)
Pattern recognition: Non-parametric classifiers for handwritten character recognition using decision trees, SVM’s (Support Vector Machine) and KNN (K- Nearest Neighbor). ( C,C++ )
Information retrieval: Indexing and retrieval of a large collection of text documents using probabilistic as well as a vector space method ( C++)

SKILLS

Harish Srinivasan - Resume