Abstract:Gravitational wave science is a pioneering field with rapidly evolving data analysis methodology currently assimilating and inventing deep learning techniques. The bulk of the sophisticated flagship searches of the field rely on the time-tested matched filtering principle within their core. In this paper, we make a key observation on the relationship between the emerging deep learning and the traditional techniques: matched filtering is formally equivalent to a particular neural network . This means that a neural network can be constructed analytically to exactly implement matched filtering and can be further trained on data or boosted with additional complexity for improved performance. Moreover, we show that the proposed neural network architecture can outperform matched filtering, both with or without knowledge of a prior on the parameter distribution. When a prior is given, the proposed neural network can approach the statistically optimal performance. We also propose and investigate two different neural network architectures MNet-Shallow and MNet-Deep , both of which implement matched filtering at initialization and can be trained on data. MNet-Shallow has a simpler structure, while MNet-Deep is more flexible and can deal with a wider range of distributions. Our theoretical findings are corroborated by experiments using real LIGO data and synthetic injections, where our proposed methods significantly outperform matched filtering at false positive rates above 5×10−3% . The fundamental equivalence between matched filtering and neural networks allows us to define a "complexity standard candle" to characterize the relative complexity of the different approaches to gravitational wave signal searches in a common framework. Additionally, it also provides a glimpse of an intriguing symmetry that could provide clues on interpretability, namely how neural networks approach the problem of finding signals in overwhelming noise. Finally, our results suggest new perspectives on the role of deep learning in gravitational wave detection. https://doi.org/10.1103/PhysRevD.105.043006 © 2022 American Physical Society

Fast Wiener filtering of CMB maps with Neural Networks

Enhancing CMB map reconstruction and power spectrum estimation with convolutional neural networks

Filtering Enhanced Tomographic PIV Reconstruction Based on Deep Neural Networks

Preconditioner-free Wiener filtering with a dense noise matrix

Generalized approach to matched filtering using neural networks

Fast Bayesian gravitational wave parameter estimation using convolutional neural networks

On the Importance of Neural Wiener Filter for Resource Efficient Multichannel Speech Enhancement

Novel neural-network architecture for continuous gravitational waves

AI-assisted Reconstruction of Cosmic Velocity Field from Redshift-Space Spatial Distribution of Halos

Deep Neural Networks to Enable Real-time Multimessenger Astrophysics

Constraining primordial non-Gaussianity using Neural Networks

Wiener Reconstruction of The Large Scale Structure

Neural Fourier Filter Bank

Cosmological model discrimination with Deep Learning

Characterizing seismic isolation using convolutional neural networks and Wiener filters

Foreground removal from WMAP 7yr polarization maps using an MLP neural network

Window function convolution with deep neural network models

Neural network reconstruction of density and velocity fields from the 2MASS Redshift Survey