Research Work
Research Projects and Contribution
Theoretical Contribution
My major research contribution is the development
of "Adaptive Kernel Models for Modeling and Generalization". This work is
well regarded by the international research community and has been published
in the most prestigious journals and conference proceedings [1-10]. This
work is funded by the University Research Initiative Grant of AUD150K.
Research Projects
My recent research interest (last 3~5 years) has
been:
-
Theoretical development of
radial basis function network, probabilistic networks and support vector
machines. Mainly in the domain of nonlinear modeling and prediction.
[1]-[10] (supported by UTS Research Strength Grant of $150K)
-
Machine learning techniques for decision
support system: mainly in the areas of terrorist threat assessment
[11]-[17] (supported by ARC Discovery Grant of $168K)
-
Machine learning
techniques for image and signal processing [18]-[26] (supported by
ARC Linkage Grant of $240K)
Research Grants
I am a joint Chief Investigator for the research projects which attracted
total grants of A600K (for 2003-2005). Some awarded grants are listed below.
Invitations, Awards and
Scholarships
- Invited session chair at the IEEE International Symposium on Signal
Processing and Information Technology, Darmstadt, Germany 2003
- Elected technical committee member for International Association of
Pattern Recognition (IAPR)
- Organizing committee member for International Conference on
Computational Intelligence for Modeling, Control and Automation (CIMCA),
Brisbane, Australia, 2004
- Recipient of the Best paper prize at the International Conference on
Image and Vision Computing held at New Zealand (with Hatice Gunes), 2002
- Recipient of the Best Research Contribution Award at the Department of
Computer Systems at University of Technology, Sydney, 2002.
- Recipient of the Norman I PhD Scholarships at the Department of
Electrical Engineering, University of Sydney (2000-2001).
- Recipient of the First Prize Winner at the All Australian Mathematics
Competition
Research Interest and
Plans
- Incremental Support Vector Machine (SVM) - large margin classifiers
such as SVM is well known for classification. However incremental SVM
learning is still an open interesting problem [3, 5].
- Kernel Principal Component Analysis (KPCA) - kernel based learning is
a very important topic in machine learning research [5, 8, 10].
- Semiparametric Kernel Density Estimation (KDE) - KDE is inherently
overfitting. Some degree of regularization is required to achieve a stable
model [18]. The key problem is the bias-variance tradeoff.
- Smart multimedia search engines - current search engines are limited
to text based search only. Recent advances in multimedia tools (MPEG 7 and
21) pave many new exciting ways in which the multimedia information may be
manipulated to provide exciting multimedia search systems [12, 17].
Qualifications
- PhD Research in Information Engineering (UC)
- Master of Computer Engineering, University of Sydney (APAI Research
Scholarship)
- Bachelor of Electrical and Information Engineering (Honours),
University of WA.
Experience
- T. Jan, Combining analytic models in neural networks for robust
classification, IEEE International Journal on Neural Networks, Elsevier
Press, (submitted).
- T. Jan, An adjustable linear to nonlinear regression model, IEEE
Transaction on Neural Networks, (submitted).
- T. Jan, Incremental Probabilistic Neural Network, International
Journal on Neural Networks, Elsevier Press, (IEEE) (in press, 2004).
- T. Jan, Combination of Linear Model and Kernel Based Neural Network
Model for Modeling and Prediction, International Journal of
Neurocomputing, Elsevier Press, (in press).
- T. Jan, Modified Probabilistic Network with Embedded Local Linear
Models, Proc. of IEEE International Conference on Machine Learning for
Signal Processing, Sao Luis, Brazil, 2004. (accepted)
- T. Jan, Combining analytical model with neural networks, Proc. of IEEE
International Symposium on Signal Processing and Information Technology
(IEEE-ISSPIT), Darmstadt, Germany, December 14-17, 2003.
- T. Jan, Robust Short Term Prediction using Combination of Linear
Regression and Modified Probabilistic Neural Network, Proc. of IEEE
International Joint Conference on Neural Networks (IEEE-IJCNN), pp.
2478-2481, Portland, Oregon, July, 2003.
- T. Jan, Short term prediction using adjustable general regression
model, Proc. of 4th International Conference on Intelligent Data
Engineering and Automated Learning, 21st-23rd March, 2003, Hong Kong.
- T. Jan, Robust Short Term Model Prediction using Combination of Linear
Model and General Regression Neural Network, Proc. of 3rd International
Conference on Computational Intelligence for Modeling, Control and
Automation, pp. 463-471, Vienna, Austria, 12th-14th February, 2003.
- T. Jan, A. Zaknich, An Adjustable Model for Linear to Nonlinear
Regression, Proc. of IEEE International Joint Conference on Neural
Networks (IEEE-IJCNN), pp. 846-850, Washington, USA, 10-16th July, 1999.
- T. Jan, Neural network based threat assessment for terrorist
activities, IEEE International Joint Conference on Neural Networks
(IEEE-IJCNN), Budapest, Hungary, 25-29 July 2004 (accepted, in press)
- T. Jan, M. Piccardi, T. Hintz, Neural Network Classifiers for Video
Surveillance, Proc. of IEEE International Workshop on Neural Network for
Signal Processing (IEEE-NNSP), pp. 729-783, Toulouse, France, September
17-19, 2003.
- M. Piccardi, and T. Jan, Recent Advances in Computer Vision, The
Industrial Physicist, vol. 9, no. 1, pp. 18-21, Feb/Mar. 2003, American
Institutes of Physics, (ISSN: 1082-1848).
- T. Jan, T. Hintz, Efficient Surveillance Image Modeling using
Combination of Linear Model and Modified Probabilistic Neural Network,
Proc. of 3rd International Conference on Computational Intelligence for
Modeling, Control and Automation, pp. 338-345, Vienna, Austria, 12th-14th
February, 2003.
- T. Jan, M. Piccardi, T. Hintz, Detection of Suspicious Behavior using
Modified Probabilistic Neural Network, Proc. of 3rd International
Conference on Computational Intelligence for Modeling, Control and
Automation, pp. 376-385, Vienna, Austria, 12th-14th February, 2003.
- T. Jan, M. Piccardi, T. Hintz, Suspicious User Activity Detection with
General Regression Neural Networks, Proc. of International Conference on
Image and Vision Computing, pp. 237-242, Auckland, New Zealand, 23rd-26th
November, 2002.
- T. Jan, Incremental Radial Basis Function Network for Robust Learning,
Proceedings of International Conference on Neural Information Processing
(ICONIP), IEEE, Shanghai, China, 14 - 18 November, 2001.
- T. Jan, Effective video object encoder using locality-enhanced support
vector machines, Proc. of IEEE International Conference on Multimedia
Signal Processing, Siena, Italy, 2004 (accepted and in press)
- T. Jan, Video Objects Encoding using Region of Interest based Neural
Network Classifiers, Proc. of IEEE International Symposium on Signal
Processing and Information Technology (IEEE-ISSPIT), Darmstadt, Germany,
14-17 December, 2003.
- T. Jan, Semiparametric kernel density estimation for statistical image
modeling, Journal of Image and Vision Computing, Elsevier Press,
(submitted).
- T. Jan, Efficient Image Modeling using Combination of Linear
Regression and General Regression Neural Network, Proc. of International
Conference on Image and Vision Computing, pp. 19-21, Auckland, New
Zealand, 23rd-26th November, 2002.
- (with H. Gunes), T. Jan, Facial Image Analysis for Plastic Surgery,
Proc. of IEEE International Conference on Systems, Man and Cybernetics,
2004.
- (with H. Gunes), T. Jan, Multimodal human computer interface, Proc.
of Asia Pacific Conference on Human Computer Interface, New Zealand,
2004.
- H. Gunes, M. Piccardi, T. Jan, Facial Proportion Analysis using
Supervised Learning Classifiers, SPIE International Proceedings on Optical
Imaging, San Jose, Nevada, USA, 2003. (in press)
- H. Gunes, M. Piccardi, T. Jan, Automatic Classification of Female
Faces using Learning Algorithms, Proc. of International Conference on
Image and Vision Computing, pp. 169-173, Auckland, New Zealand, 23rd-26th
November, 2002.
- T. Jan, A. Zaknich, Separation of Signals with Overlapping Spectra
using Signal Characterization and Hyperspace Filtering, Proc. of IEEE
International Symposium on Adaptive Systems for Signal Processing,
Communications, and Control, pp. 327-332, Lake Louise, Alberta, Canada,
1-4th October, 2000.
|