Education:
Academic Background:
Image Processing: Segmentation, Classification, Features extraction, Feature Selection, Sparse Neural Networks.
Audio & Music Processing: Deep Learning Networks for timbre, pitch estimation, Audio Scene Classification, Music Tag Classification.
Industrial Experience:
May 2015- Sept 2015
Aug 2011- 2012
Dec 2010-April 2011
The Mathworks Inc., Natick, MA
Machine Learning Development Intern
- Developed C++ production code to implement distance metrics for heterogeneous mixtures including categorical & numerical data.
Siemens Energy Inc, Richland, Mississippi, USA
Technical Staff Member
- Developed Communication layer protocol based on ARM Processor for MJ-5 voltage regulators.
- Developed the firmware code for new product line on next generation control panel for step voltage regulators.
- Designed a frequency detection algorithm to be used in regulators.
Verizon Wireless, Irving, Texas, USA
Software Developer Intern
-Designed Web Service interface using XML, XDS’s, Axis and JAVA. Analyzed functional requirement and worked in the development of recommendation engine algorithms for FiOS project modules.
Honors and Scholarships:
1) Received 2013 IEEE Computational Intelligence Society Walter Karplus Summer Research Grant.
2) STEM Graduate Fellowship, Office of Graduate Studies, The University of Texas at Arlington.
3) Two times recipient of IEEE Computational Intelligence society-2007(IEEE-CIDM, Hawaii) and 2011( IEEE-FUZZ, Taiwan) for Outstanding student paper Travel Grant.
4) Member of Tau Beta Pi Engineering Honor Society.
5) Image Processing and Neural Networks lab Graduate Scholarship.
Reviewer:
1) Neurocomputing Journal.
2) Journal of Engineering Applications for Artificial Intelligence (EAAI) .
Research Experience:
Spring 2012- Current:
Summer 2013:
Summer 2011:
Fall 2009- 2011:
Summer 2008- 2009:
Summer 2005:
The University of Texas at Arlington, USA
Graduate Research Assistant, Image Processing, and Neural Networks Lab
Supervisor: Dr. Michael Manry
Designing second order deep learning autoencoder for Classification problems
Seoul National University, South Korea
Visiting Researcher, Music, and Audio Research Group
Supervisor: Dr. Kyogu Lee
Development of Deep Learning Networks for Music and audio applications.
- Worked for Music Tag classification, Audio Scene classification and Bird sound classification problem.
-Developed Multi-instance, Multi-Instance (MIML) classifiers.
- Technical Report: [pdf]
Ajou University, South Korea
Visiting Researcher, Multimedia Signal Processing Lab
Supervisor: Dr. Nojun Kwak
Development of L-1 based subspace methods and Viola-Jones face detection algorithm
- Worked for feature selection and vehicle detection and classification algorithm funded by Hyundai-Kia motors.
- Paper published: ( LINK )
The University of Texas at Arlington, USA
Graduate Research Assistant, Image Processing, and Neural Networks Lab
Supervisor: Dr. Michael Manry
Designing a modified Multilayer Perceptron (MLP) for Approximation and Classification problems
- Modify MLP with optimum hidden units and no local minima to solve real-life approximation and classification problems using Orthogonal Least Squares and second order learning factors. Developed in both C and MATLAB.
Construction and training of radial basis functions (RBF) networks with optimization of network parameters using second order algorithms (Master’s Thesis)
- Designed a family of orthogonal least squares based algorithm for RBF Neural Networks in C and MATLAB. Developed algorithms for optimizing RBF parameters and modified distance measure via Newton’s Method.
Indian Institute of Technology, Kanpur, India
Research Assistant, Neural Network Laboratory
Supervisor: Dr. Prem.K.Kalra
Audio Based Fault Diagnosis of Internal Combustion Engines
- Responsible for developing Data Acquisition software through LabVIEW. Designed Adaptive Filters as pre-processing with wavelet packet energy based feature extraction and later dynamic clustering. Developed a Neural Network classifier. Responsible for developing final JAVA based Software, installed at a leading Automobile Company in India.
Blind Source Separation of Audio signal in a reverberant environment
- Developed Novel Multistage Multi-resolution Algorithm in MATLAB and applied on real-time speech signals.
Indian Institute of Technology, Kanpur, India
Summer Intern, Neural Network Laboratory
Supervisor: Dr. Prem.K.Kalra
Counter-propagation Network: Application in Function Approximation, Classification, Geometrical Rotation, Image and Data Compression.
- Developed a novel counter propagation neural model to solve real classification and Image compression problems.
Based on this research, worked on the video surveillance application.
Graduate Teaching Experience:
Fall'14
Spring'14
Fall '13
Spring '13
Summer'10
Spring'10
EE 3417 Linear Systems (Undergraduate Course)
EE 5350 Digital Signal Processing (Graduate Course)
EE 3417 Linear Systems (Undergraduate Course)
EE-5356 Digital Image Processing (Graduate Course)
EE 3318 Discrete Signal and Systems (Undergraduate Course)
EE 5350 Digital Signal Processing (Graduate Course)
Software Releases:
Kanishka Tyagi, Dr.Cai Xun, Dr. M. Manry, “Multilayer Perceptron for Approximation/ Regression Analysis Networks”, copyright © 2010, by the IPNN Lab, The University of Texas. Available in both MATLAB and C versions.
Kanishka Tyagi, Dr.Cai Xun, Dr.M.Manry, “Multilayer Perceptron for Designing Classification Networks”, copyright © 2010, by the IPNN Lab, The University of Texas. Available in C version.
More Details: http://www-ee.uta.edu/eeweb/IP/new_software.html
Computer skills: