Kymatio: Scattering Transforms in Python
By Mathieu Andreux, Tomás Angles, Georgios Exarchakis, Roberto F. Leonarduzzi, G. Rochette, L. Thiry, J. Zarka, Stéphane Mallat, Joakim Andén, Eugene Belilovsky and 8 more. 2018
205 citations recorded by the source corpus.
The wavelet scattering transform is an invariant signal representation suitable for many signal processing and machine learning applications. We present the Kymatio software package, an easy-to-use, high-performance Python implementation of the scattering transform in 1D, 2D, and 3D that is compatible with modern deep learning frameworks. All transforms may be executed on a GPU (in addition to CPU), offering a considerable speed up over CPU implementations. The package also has a small memory footprint, resulting inefficient memory usage. The source code, documentation, and examples are available undera BSD license at this https URL
Subfields: Artificial Intelligence, Environmental Engineering, Oceanography
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Lenka Zdeborov'a
- Florent Krzakala
- Dan Boneh
- Florian Tramèr
- Sebastian Goldt
- Cédric Gerbelot
- Bruno Loureiro
- Hugo Cui
In Environmental Engineering
Citations in the corpus
- Differentially Private Learning Needs Better Features (or Much More Data) (2020, methodology): “We use the ScatterNet implementation from Kymatio (Andreux et al., 2020), and the DP-SGD implementation in opacus (pytorch/opacus, 2020) (formerly called pytorch-dp).”
- Capturing the learning curves of generic features maps for realistic data sets with a teacher-student model (2021, methodology): “The different curves compare the performance of no feature map (blue), random feature map with σ = erf and Gaussian projection (orange), the scattering transform with parameters J = 3, L = 8 [18] (green) and of the limiting kernel of the RF map above [56] (red).”