Registration desk will be open at 5pm on June 1 and a small wellcome reception will be held. The following plenary talks will be presented in the conference: June 2: 9:30 - 10:15 * Prof. Paulo S. R. Diniz, Univ. Federal do Rio de Janeiro "Adaptive Filtering in Wireless Communications" June 2: 13:30 - 14:00 Prof. H. Krim, North Carolina State University "Smart Nonlinear Diffusion: A probabilistic Approach" June 3: 9:30 - 10:00 * Prof. Yih-Fang Huang, Notre Dame University "Nonlinear MMSE Multistage Interference Cancellation with Set-Membership Filtering" (Professors Diniz and Huang are IEEE CAS Society Distinguished Lecturers). June 3: 13:30 - 14:00 Prof. Georgios B. Giannakis, University of Minnesota "Generalized Multi-Carrier CDMA for Multiuser Wireless Communications" June 3: 14:00 - 14:30 Prof. Haldun Ozaktas, Bilkent University "The Fractional Fourier Transform and Time-Order Signal Representations" * (Professors Diniz and Huang are IEEE CAS Society Distinguished Lecturers). Abstracts: Nonlinear MMSE Multistage Interference Cancellation with Set-Membership Filtering", by Yih-Fang Huang Multistage interference cancellation techniques such as Parallel Interference Cancellation (PIC) and Successive Interference Cancellation (SIC) have been shown to be effective for mitigating multiple access interference (MAI) in CDMA wireless communications. This presentation will start with a discussion on nonlinear MMSE multiuser detection (MMSE-MUD), an alternative to the maximum likelihood (ML) or minimum probability of error (MPE) multiuser detection. It will then present a nonlinear MMSE-PIC as an iterative solution to the nonlinear MMSE-MUD, with much reduced complexity. A novel algorithm, whose complexity is only linearly (not exponentially) proportional to the number of users, will be presented for the MMSE-PIC. This algorithm employs set-membership filtering (SMF) for channel estimation. One of the main features of SMF is the selective update of channel parameter estimates. This renders substantial reduction in computational complexity. The presentation will coclude by showing examples that demonstrate significantly enhanced service capacity. ------------------------------------------------ Smart Nonlinear Diffusion: A probabilistic Approach, by H. Krim We revisit the nonlinear diffusion problem of signals/images which has been and remains of great research interest in image analysis and computer vision. We first reformulate the linear diffusion in a probabilistic setting and interpret as a result, the filtering as random motions of particles. With this approach, we are able to also interpret the well-known Perona-Malik nonlinear diffusion equation and explicate its asymptotic behavior (stable point) of total smoothing of the signal/image. In light of this development, we are able to propose a new technique which not only achieves remarkably improved performance but also obviates and hence solves the long standing of problem of evolution stopping time. We extend such an approach to include diffusion along level sets to be preserve and capture characteristic features in images which are typically impoortant for recognition, compression etc.. We demonstrate the performance via a number of substantiating examples.