Finite Neuron Method - Lecture 2

Explore a comprehensive lecture on the finite neuron method in this 1 hour 25 minute conference talk by Jinchao Xu at the Centre International de Rencontres Mathématiques in Marseille, France. Delve into topics such as electrical entropy, the curse of dimensionality, neural network functions, and hierarchical basis. Examine approximation properties, isotropic spaces, deep networks, and the gradient neuron method. Investigate the frequency principle, ReLU activation, accuracy, rate of convergence, and the multigrade method. Access this video and other mathematical talks on CIRM's Audiovisual Mathematics Library, featuring chapter markers, keywords, abstracts, bibliographies, and multi-criteria search functionality for an enhanced learning experience.

Electrical entropy
Curse of dimensionality
Neural network functions
Hierarchical basis
Approximation properties
Conclusion
Isotropic spaces
Deep networks
Gradient neuron method
Frequency principle
Relu
Accuracy
Rate of convergence
Multigrade method