Power Transform Revisited: Numerically Stable, and Federated

Installation

1. Create a Python environment:

conda create -n powertf python=3.11
conda activate powertf

2. Clone this repository:

git clone https://github.com/xuefeng-xu/powertf.git && cd powertf

3. Install dependencies:

pip install -e .

Reproduction

For training experiments, see train/README.md.

For numerical experiments, see numerical/README.md.

For federated experiments, see federated/README.md.

Plots are saved under ./img/*.pdf