Research Assistant Professor
Institute of Molecular Biology and Biotechnology
Foundation for Research and Technology, Hellas (FO.R.T.H.)
GR 711 10 Heraklion, Crete
Tel: (+30) 2810 391139
Fax: (+30) 2810 391101
Dept. of Biomedical Engineering
University of Southern California
Los Angeles, CA 90089
Tel: (213) 740-3397
Fax: (213) 740-0343
I am a Greek Cypriot, born in Vatily, Famagusta on August 6, 1974. I received my B.S. in Mathematics from the University of Cyprus in May 1996, the M.S. degree in Biomedical Engineering in May 1998 and the Ph.D. degree in Biomedical Engineering in July 2000, both from the University of Southern California. I'm currently a Research Assistant Professor in the Institute of Molecular Biology and Biotechnology at FO.R.T.H., and head of the Computational Biology Laboratory.
Modeling/Simulations of Biological Systems: Development of simulation algorithms for biological system modeling and especially the functions of learning and memory storage in the brain under normal and pathogenic conditions (e.g. stress, aging, drug abuse, degenerative diseases) via the use of:
1. machine learning approaches (e.g. artificial neural networks, polynomial classifiers, probabilistic classifiers, etc.). Applications of methods for visual object recognition, declarative memory deficits in Alzheimer's and other degenerative diseases.
2. detailed biophysical models of CA1 neurons using the NEURON simulation environment. Analysis of computational properties of pyramidal neurons and prediction of learning and memory deficits under pathological conditions.
3. theoretical analysis and abstract mathematical modeling (e.g. analysis of the effects of active dendritic mechanisms, structural plasticity and dendritic morphology on the memory capacity of pyramidal neurons).
Bioinformatics: Development of innovative computer methods and tools in bioinformatics and computational genomics.
1. Mathematical analysis and numerical processing of images with gene expression data. Statistical and related methods for noise reduction, image sharpening, normalization etc.
2. Development of mathematical techniques for optimal of data normalization in microarray experiments.
3. Design and development of theoretical as well as computational methods for processing of gene expression data. Analysis and development of supervised and unsupervised algorithms for gene classification and clustering mainly focusing in search for genes/regulatory pathways with predictive pathogenic power and pharmaceutical applications.
Design of experimental studies for testing the role of nonlinear dendritic integration to a cells' computational/learning capacity both in vitro and in vivo.
Design of experimental protocols for optimal acquisition of gene expression data, e.g. selection of suitable control genes for better normalization and noise reduction.
Development of mathematical models and statistical methods for decision-making problems, with applications in artificial intelligence, robotics and medical informatics.
Statistical analysis and qualitative comparison of the learning capacity of linear and nonlinear classifiers, especially when used to model biological neurons.
Neural networks and related statistical techniques.
Poirazi, P. Brannon, T. & Mel, B.W. Online Supplement: About the Model. In, Neuron , vol. 37, March 2003. [PDF]
Poirazi, P. Brannon, T. & Mel, B.W.Pyramidal Neuron as 2-Layer Neural Network. In, Neuron , vol 37, pg. 989-999, March 2003. [PDF]
Poirazi, P. Neocleous, C., Pattichis, C. & Schizas, C. A brain-derived heterogeneous neural architecture for medical data classification. Submitted for publication (2002).
Poirazi, P. & Mel, B.W. Impact of Active Dendrites and Structural Plasticity on the Storage Capacity of Neural Tissue. In Neuron, vol. 29, pg. 1-20, 2001. [More]
Poirazi, P. & Mel, B.W. Choice and Value Flexibility Jointly Contribute to the Capacity of a Subsampled Quadratic Classifier. In Neural Computation, vol. 12, num. 15, pg. 1189-1205, 2000 [More]
Poirazi, P. & Mel, B.W. Towards the memory capacity of neurons with active dendrites. In Neurocomputing vol. 26-27, pg. 237-245, 1999. [More]
Poirazi, P., C. Neocleous, C. Pattichis & Schizas C. A Biologically Inspired Neural Network Composed of Dissimilar Single Neuron Models. In Proc. of the 23rd Annual International Conference of the IEEE Engineering in Medicine and Biology Society , Istanbul, Turkey, Oct. 2001.
Poirazi, P. & Mel, B.W. Memory Capacity of Linear vs. Nonlinear Models of Dendritic Integration. In Advances in Neural Information Processing Systems (NIPS - 12), Eds. S. A. Solla, T. K. Leen, K. R. Miller, vol. 12, pg. 157-163, MIT Press, 2000. [More]
Poirazi, P. & Mel, B.W. Sublinear vs. Superlinear Synaptic Integration? Tales of a Dublicitous Active Current. In Proc. of the 7th Joint Symposium on Neural Computation , UCSD, vol. 10, pg. 88-95, May 20 2000. [More]
Poirazi, P. & Mel, B.W. Effects of Morphology on the Memory Capacity of Neurons with Active Dendrites. Society for Neuroscience Abstracts , vol. 25, Part II, pg. 2258, 1999. [More]
Poirazi, P. & Mel, B.W. Effects of Morphology on the Memory Capacity of Neurons with Active Dendrites Proc. of the 6th Joint Symposium on Neural Computation, UCSD, vol. 9, pg. 104-110, May 22 1999. [More]
Poirazi, P. & Mel, B.W. The capacity of subsampled-quadratic classifiers: why neurons with active dendrites may win big. Proc. of the 5th Joint Symposium on Neural Computation, UCSD, vol. 8, pg. 123-129, May 16 1998. [More]
Poirazi, P. & Mel, B.W. Why active dendrites can remember more. Society for Neuroscience Abstracts ,vol. 24, Part I, pg. 329, 1998. [More]Poirazi, P. & Mel, B.W. Memory Capacity of neurons with active dendrites. Abstract in Proc. of the 4th Joint Symposium on Neural Computation, UCSD, vol. 7, pg. 166, May 17 1997. [More]