A carregar...
Publicação
On The Topological Expressive Power of Neural Networks
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 374.18 KB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
We propose a topological description of neural network expressive power. We adopt the topology of the space of decision boundaries realized by a neural architecture as a measure of its intrinsic expressive power. By sampling a large number of neural architectures with different sizes and design, we show how such measure of expressive power depends on the properties of the architectures, like depth, width and other related quantities.
Descrição
Petri, G., & Leitão, A. (2020). On The Topological Expressive Power of Neural Networks. Poster session presented at Topological Data Analysis and Beyond. , . https://openreview.net/forum?id=I44kJPuvqPD
Palavras-chave
Neural Networks Expressive Power Decision Boundary Classification