Neural Architecture Search for Inversion

TL;DR

This paper explores neural architecture search and loss function design to improve deep learning models for inversion problems, aiming to find optimal architectures and cost functions for better performance.

Contribution

It introduces a method to optimize neural architectures and loss functions specifically for inversion tasks, extending prior work with automated architecture search.

Findings

Identified more effective loss functions for inversion

Demonstrated neural architecture search improves inversion accuracy

Compared different neural network architectures for inversion

Abstract

Over the year, people have been using deep learning to tackle inversion problems, and we see the framework has been applied to build relationship between recording wavefield and velocity (Yang et al., 2016). Here we will extend the work from 2 perspectives, one is deriving a more appropriate loss function, as we now, pixel-2-pixel comparison might not be the best choice to characterize image structure, and we will elaborate on how to construct cost function to capture high level feature to enhance the model performance. Another dimension is searching for the more appropriate neural architecture, which is a subset of an even bigger picture, the automatic machine learning, or AutoML. There are several famous networks, U-net, ResNet (He et al., 2016) and DenseNet (Huang et al., 2017), and they achieve phenomenal results for certain problems, yet it's hard to argue they are the best for inversion problems without thoroughly searching within certain space. Here we will be showing our architecture search results for inversion.