- GearNet: A Simple, Effective Protein Structure Encoder
4
- Edge Message Passing
5
- GearNet-Edge
6
- Experimental Setup
7
- Geometric Pretraining
8
- Contrastive Learning: SimCLR
9
- Self-Prediction Baselines
10
- Experimental Setup
11
- Pretraining on Different Datasets
12
- Conclusion & Future Work
13
- Q+A
Description:
Explore protein representation learning through geometric structure pretraining in this 53-minute talk by Zuobai Zhang from Valence Labs. Dive into the world of protein function prediction and discover GearNet, a simple yet effective protein structure encoder. Learn about edge message passing, GearNet-Edge, and various experimental setups. Understand geometric pretraining techniques, including contrastive learning with SimCLR and self-prediction baselines. Examine the results of pretraining on different datasets and gain insights into future research directions. Conclude with a Q&A session to deepen your understanding of this innovative approach to protein representation learning.
Protein Representation Learning by Geometric Structure Pretraining