Introduction to Geometric Deep Learning 수료증
모집인원999명
학습기간2024-03-01 ~ 2024-12-31
|
강의시간 |
강의내용 |
실습여부 |
|
1 |
- Why CNNs, MLPs, RNNs are insufficient for non-Euclidean data - Neural network for graphs & sets - Euclidean transformations, invariance, and equivariance |
|
|
2 |
- Invariant geometric GNNs (SchNet, DimeNet, and SphereNet) - Simple equivariant geometric GNNs (EGNN and NequIP) |
O |
|
3 |
- Local frame-based geometric GNNs (ClofNet and LEFTNet) - Frame averaging for geometric GNNs (Frame averaging, FAENet) |
O |
|
4 |
- Steerable features, rreducible representations, Wigner-D matrix, spherical harmonics, Clebsch-Gordan tensor product - Steerable geometric GNNs (Tensor field network, SE(3)-Transformer) |
|
|
5 |
- Steerable geometric GNNs (Equiformer, MACE, eSCN, EquifomerV2) |
O |
강의 소개 및 개요입니다.
|
성명 |
안성수 |
소속기관 |
POSTECH |
|
과목명 |
Introduction to Geometric Deep Learning |
강의시간 |
5 |
|
학습목표 |
Geometric deep learning (GDL) aims to develop models capable of handling structured and non-Euclidean data, such as geometric graphs. In this course, we study recent GDL models, with the focus on geometric graph neural networks for molecules. |
||
강의 선수과목 및 준비사항입니다.
|
선수과목 |
graph neural network |
|
참고자료 |
|
|
준비사항 |
실습에 참여할 경우 실험을, 위한 GPU 세팅 |