An FPGA-based eigenfilter using fast Hebbian learning
Refereed conference paper presented and published in conference proceedings

CUHK Authors
Author(s) no longer affiliated with CUHK

Full Text

Times Cited
Web of Science0WOS source URL (as at 13/08/2020) Click here for the latest count

Other information
AbstractWe present a high-gain, multiple learning/decay rate, "cooling off' annealing strategy to a modified Generalized Hebbian Algorithm (GHA) that gives good approximate solution within one training epoch, and with fast convergence to accurate principal components within a few more epochs. A novel bit-shifting normalization procedure is shown to bound the weight vector norm effectively and eliminates the need for performing division. This leads to an FPGA-based computational framework using only fixed point arithmetic instead of more complicated floating point design. Simulation results on Xilinx DSP System Generator tool indicate the practicality of the approach, where real-time eigenfilter can be readily implemented on field programmable gate arrays with limited resources.
All Author(s) ListLam KP, Mak ST
Name of ConferenceIEEE International Conference on Acoustics, Speech, and Signal Processing
Start Date of Conference06/04/2003
End Date of Conference10/04/2003
Place of ConferenceHONG KONG
Country/Region of ConferenceChina
Pages765 - 768
LanguagesEnglish-United Kingdom
Web of Science Subject CategoriesAcoustics; Computer Science; Computer Science, Artificial Intelligence; Engineering; Engineering, Biomedical; Engineering, Electrical & Electronic

Last updated on 2020-14-08 at 03:56