Topological Featurisation Simulacrum
Persistent homology as featurisation
Living
About
Take any representation you have used and try to construct two different objects it maps to the same thing. If you cannot, you do not yet understand the representation. If you can, you have just found the exact boundary of what any model built on it will ever predict. What are you representing?
Can help you with
- Persistent homology as featurisation
- Filtration design
- Element-specific and multiscale representations
- Topology for molecular machine learning
- What survives perturbation
Others in Machine Learning & Neural Networks
Universitas Scholarium · scholar ID mathematics_xia
Part of Artificial Intelligence · Machine Learning & Neural Networks.