Conferences and Seminars


  • October 22-26, 2007, Stability Problems for Stochastic Models
  • Joint Israeli-Russian Workshop on Stochastic Models: Theory and Application, April 5, 2006,
  • Algorithmic Techniques for Data Mining Workshop,  May, 2005


"Some New Pathways of Research at IAM of METU andCollaboration Proposed --- MARS --- SVM with Infinitely Many Kernels, Coding Theory and Cryptography Indicated
Prof. Gerhard-Wilhelm Weber Institute of Applied Mathematics,Middle East Technical University,Ankara, Turkey,and Faculty of Economics, Business and Law,University of Siegen,Germany
Date and Place 
24.01.08, 12:00-13:00,  Room M208
By this presentation, some of the recent studies at Institute of Applied Mathematics will become briefly introduced which wehope to be of some interest for the precious colleagues fromComputer Science and Computer Engineering at Technion.Firstly and mainly, we discuss modern spline regression with the help of continuous optimization, especially, multivariate adaptive regression spline (MARS) of J. Friedman. This is a tool from statistics, important in classification and regression, with applicability in many areas of finance, science and technology,very useful in high dimensions and shows a great promise for fitting nonlinear multivariate functions.
The MARS algorithm for estimating the model function consists of two subalgorithms.
As an alternative to the backward stepwise algorithm, we propose to construct a penalized residual sum of squares for a Tikhonov regularization problem which we treat using continuous optimization, considered  to become a complementary technology and alternative to the back-ward stepwise algorithm. Especially, we employ conic quadratic programming(CQP), permitting the use of interior point methods. The new approach will be analyzed, computationally applied and compared with the classical one of MARS. Conic programming has a great tradition in Israel!Secondly, still remaining in statistical learning and computationalstatistics, we will sketch some new pathways of research initiated by us in Machine Learning. It is characterized by an infinite number of kernels in the optimization problem called Support Vector Machine. Today, we are employing both infinite and semi-infinite optimization here.Thirdly, a short introduction about the research of IAM Department of Cryptography will be given. The colleagues there and all the ones related with aforementioned research projects would be very honoured and glad if closer scientific exchange could be  possible on all those various challenging fields of science.By this presentation a special thank will be given to the preciouscolleagues from Israel and ORSIS for all their rich and uniquecontributions to scientific progress, collaboration andfriendship - also in the name of EUROPT, EURO Working Groupon Continuous Optimization.