Molecule Representation Learning: A Perspective from Topology, Geometry, and Textual Description
Valence Labs via YouTube
Google, IBM & Meta Certificates — 40% Off for a Limited Time
MIT Sloan: Lead AI Adoption Across Your Organization — Not Just Pilot It
Overview
Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
Explore molecule representation learning from the perspectives of topology, geometry, and textual description in this 55-minute conference talk by Shengchao Liu from Valence Labs. Delve into the core problem of AI for drug discovery, focusing on molecule pretraining techniques. Learn about three key papers: GraphMVP, GeoSSL, and MoleculeSTM. Discover the potential of zero-shot text-guided molecule editing and examine practical examples. Gain insights into molecular data structures, representations, and their applications in AI-driven drug discovery. Conclude with a comprehensive recap and engage in a Q&A session to deepen your understanding of this cutting-edge field.
Syllabus
- Intro
- Molecular Data Structure and Representation
- Paper #1 - GraphMVP
- Paper #2 - GeoSSL
- Paper #3 - MoleculeSTM
- Zero-shot Text-Guided Molecule Editing
- Molecule Editing Examples
- Recap
- Q+A
Taught by
Valence Labs