Technical Program

Click on the icon to view the manuscript on IEEE XPlore in the IEEE ICASSP 2020 Open Preview.

SS-L3: A Signal-Processing View of Graph Neural Networks

Session Type: Lecture
Time: Tuesday, 5 May, 11:30 - 13:30
Location: On-Demand
Virtual Session: View on Virtual Platform
Session Chair: Siheng Chen, Mitsubishi Electric Research Laboratories (MERL)
 
 SS-L3.1: GRAPH NEURAL NET USING ANALYTICAL GRAPH FILTERS AND TOPOLOGY OPTIMIZATION FOR IMAGE DENOISING
         Wengtai Su; National Tsing Hua University
         Gene Cheung; York University
         Richard P. Wildes; York University
         Chia-Wen Lin; National Tsing Hua University
 
 SS-L3.2: DEFENDING GRAPH CONVOLUTIONAL NETWORKS AGAINST ADVERSARIAL ATTACKS
         Vassilis N. Ioannidis; University of Minnesota
         Georgios B. Giannakis; University of Minnesota
 
 SS-L3.3: CONSTRAINED SPECTRAL CLUSTERING FOR DYNAMIC COMMUNITY DETECTION
         Abdullah Karaaslanli; Michigan State University
         Selin Aviyente; Michigan State University
 
 SS-L3.4: TOWARDS AN EFFICIENT AND GENERAL FRAMEWORK OF ROBUST TRAINING FOR GRAPH NEURAL NETWORKS
         Kaidi Xu; Northeastern University
         Sijia Liu; IBM Research
         Pin-Yu Chen; IBM Research
         Mengshu Sun; Northeastern University
         Caiwen Ding; University of Connecticut
         Bhavya Kailkhura; Lawrence Livermore National Laboratory
         Xue Lin; Northeastern University
 
 SS-L3.5: DEEP GEOMETRIC KNOWLEDGE DISTILLATION WITH GRAPHS
         Carlos Lassance; IMT Atlantique
         Myriam Bontonou; IMT Atlantique
         Ghouthi Boukli Hacene; IMT Atlantique
         Vincent Gripon; IMT Atlantique
         Jian Tang; HEC Montreal/Mila
         Antonio Ortega; University of Southern California
 
 SS-L3.6: ON THE CHOICE OF GRAPH NEURAL NETWORK ARCHITECTURES
         Clément Vignac; Ecole Polytechnique Fédérale de Lausanne (EPFL)
         Guillermo Ortiz-Jiménez; Ecole Polytechnique Fédérale de Lausanne (EPFL)
         Pascal Frossard; Ecole Polytechnique Fédérale de Lausanne (EPFL)