Learnable wavelet neural networks for cosmological inference
This work addresses the need for efficient and interpretable methods in cosmological data analysis, particularly for expensive simulations, but it is incremental as it builds on existing scattering transform techniques.
The authors tackled the problem of cosmological inference and marginalization over astrophysical effects by applying learnable wavelet neural networks, finding that these scattering architectures outperform CNNs, especially with small training data samples, and also developed an interpretable lightweight model.
Convolutional neural networks (CNNs) have been shown to both extract more information than the traditional two-point statistics from cosmological fields, and marginalise over astrophysical effects extremely well. However, CNNs require large amounts of training data, which is potentially problematic in the domain of expensive cosmological simulations, and it is difficult to interpret the network. In this work we apply the learnable scattering transform, a kind of convolutional neural network that uses trainable wavelets as filters, to the problem of cosmological inference and marginalisation over astrophysical effects. We present two models based on the scattering transform, one constructed for performance, and one constructed for interpretability, and perform a comparison with a CNN. We find that scattering architectures are able to outperform a CNN, significantly in the case of small training data samples. Additionally we present a lightweight scattering network that is highly interpretable.