iGDL 2021: Israeli Geometric Deep Learning Workshop

Monday, October 11th, 2021 14:00-19:00 (Israel timezone), Virtual

14:00 - 14:10 pm	Opening Remarks
14:10 - 14:40 pm	Ron Kimmel Title: Understanding or Learning Geometry: Axiomatic, Unsupervised, and Semisupervised Approaches
14:40-15:30 pm 4 x (10+2 minutes)	Short talk session 1: Meshes and implicits Amir Barda (TAU) MeshCNN++ Benjamin Groiser (Technion) Human Correspondence Matching with Mesh Convolutional Networks Lior Yariv (WIS) Volume Rendering of Neural Implicit Surfaces Or Perel (TAU) SAPE: Spatially-Adaptive Progressive Encoding for Neural Optimization
15:30 - 16:00 pm	Ronen Basri Title: Equivariant structure from motion and more
16:00-16:10 pm	Short break
16:10 - 17:10 pm 5 x (10+2 minutes)	Short talk session 2: Point clouds Gal Metzer (TAU) Orienting Point clouds with Dipole Propagation Itai Lang (TAU) Geometric Adversarial Attacks and Defenses on 3D Point Clouds Noam Rotstein (Technion) Colored Point Cloud to Image Alignment Sigal Raab (TAU) FLEX: Parameter-free Multi-view 3D Human Motion Reconstruction Matan Atzmon (WIS) Frame averaging for Invariant and Equivariant Netwoek Design
17:10-17:40 pm	Amir Globserson Title: Modeling Entities in Language and Vision
17:40 - 18:50 pm 6 x (10+2 minutes)	Short talk session 3: Graphs Eli Meirom (NVIDIA) Controlling graph dynamics with reinforcement learning and graph neural networks Amit Boyarsky (Technion) Deep Matrix Completion On Graphs: A Spectral Approach Moshe Eliasof (BGU) PDE-GCN: Novel Architectures for Graph Neural Networks Motivated by Partial Differential Equations Eran Treister (BGU) Haar Wavelet Feature Compression for Quantized Graph Convolutional Networks Gilad Yehudai (NVIDIA/WIS) From Local Structures to Size Generalization in Graph Neural Networks Ben Finkelstein (Technion) Single-Node Attacks for Fooling Graph Neural Networks
18:50-19:00 pm	Concluding Remarks

Contact Info

E-mail: israel.gdl.2021@gmail.com