TY - GEN
T1 - Global Optimization of Graph Filters with Multiple Shift Matrices
AU - Fan, Jie
AU - Tepedelenlioglu, Cihan
AU - Spanias, Andreas
N1 - Publisher Copyright:
© 2019 IEEE.
PY - 2019/11
Y1 - 2019/11
N2 - Using graphs to represent data sets that reside on irregular and complex structures can bring special advantages. Graph signal processing (DSPG) converts traditional DSP operators, such as time shift, linear filters and Fourier transform, from time and frequency domain to the graph domain. In machine learning applications, DSPG provides an approach for semi-supervised classification. Different from conventional graph-filter-based classifiers, we propose a new graph filter with multiple graph shift matrices that can provide better classification performance when the feature quality is uneven. To solve the resulting non-convex problem, a tight and efficient convex relaxation approach is introduced. Through a branch and bound optimization method, we can find the mapping from the optimum relaxed parameter set to original parameter set, which technically provides the globally optimum solution. Simulation experiments corroborate our results.
AB - Using graphs to represent data sets that reside on irregular and complex structures can bring special advantages. Graph signal processing (DSPG) converts traditional DSP operators, such as time shift, linear filters and Fourier transform, from time and frequency domain to the graph domain. In machine learning applications, DSPG provides an approach for semi-supervised classification. Different from conventional graph-filter-based classifiers, we propose a new graph filter with multiple graph shift matrices that can provide better classification performance when the feature quality is uneven. To solve the resulting non-convex problem, a tight and efficient convex relaxation approach is introduced. Through a branch and bound optimization method, we can find the mapping from the optimum relaxed parameter set to original parameter set, which technically provides the globally optimum solution. Simulation experiments corroborate our results.
KW - global optimization
KW - graph filter
KW - graph signal processing
KW - semi-supervised classification
UR - http://www.scopus.com/inward/record.url?scp=85083344236&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85083344236&partnerID=8YFLogxK
U2 - 10.1109/IEEECONF44664.2019.9048759
DO - 10.1109/IEEECONF44664.2019.9048759
M3 - Conference contribution
AN - SCOPUS:85083344236
T3 - Conference Record - Asilomar Conference on Signals, Systems and Computers
SP - 2082
EP - 2086
BT - Conference Record - 53rd Asilomar Conference on Circuits, Systems and Computers, ACSSC 2019
A2 - Matthews, Michael B.
PB - IEEE Computer Society
T2 - 53rd Asilomar Conference on Circuits, Systems and Computers, ACSSC 2019
Y2 - 3 November 2019 through 6 November 2019
ER -