Neural approximation of Wasserstein distance via a universal architecture for symmetric and factorwise group invariant functions.
Samantha ChenYusu WangPublished in: NeurIPS (2023)
Keyphrases
- network architecture
- neural architecture
- neural network
- neural architectures
- pointwise
- distance function
- continuous functions
- distance measure
- error bounds
- associative memory
- real time
- bio inspired
- management system
- approximation error
- affine transformation
- approximation algorithms
- euclidean distance
- greedy algorithm
- convex functions
- special case
- absolute error