Baseline Models for Solving the Inverse Scattering Problem (Under construction: code will be made PEP8 compliant by March 7th.)
This repository contains four baseline deterministic models and the U-ViT diffusion model for solving the wideband inverse scattering problem. The included models are:
- SwitchNet
- Wideband Butterfly Network
- Uncompressed Equivariant Model (EquiNet)
- Compressed Equivariant Model (B-EquiNet)
- U-ViT Diffusion Model
SwitchNet is described in SwitchNet: a neural network model for forward and inverse scattering problems.
Wideband Butterfly Network is detailed in Wide-band butterfly network: stable and efficient inversion via multi-frequency neural networks.
The Uncompressed and Compressed Equivariant Models are explained in Solving the wide-band inverse scattering problem via equivariant neural networks.
The deterministic models were implemented by Borong Zhang using code provided by the original authors and the Swirl-Dynamics repository, while the U-ViT Diffusion Model was implemented by Martin Guerra based on the Swirl-Dynamics probabilistic diffusion project.
Project Environment can be installed by
conda create -n isp_baseline python=3.11
conda activate isp_baseline
pip install git+https://github.com/borongzhang/ISP_baseline.git@main
pip install --upgrade "jax[cuda12_pip]" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
Demos for these models can be found in the examples folder.
We have made the datasets and the data generation scripts publicly available in the repository.