Skip to content


Repository files navigation


PyPI version Downloads CC BY-NC-SA 4.0

ConDo Adapter performs Confounded Domain Adaptation, which corrects for batch effects while conditioning on confounding variables. We hope it sparks joy as you clean up your data!

Using and citing this toolbox

If you use this toolbox in your research and find it useful, please cite ConDo using the following reference to our arXiv preprint:

In Bibtex format:

  doi = {10.48550/ARXIV.2203.12720},
  url = {},
  author = {McCarter, Calvin},
  title = {Towards Backwards-Compatible Data with Confounded Domain Adaptation},
  publisher = {arXiv},
  year = {2022},


Installation from pip

You can install the toolbox through PyPI with:

pip install condo

Note: If you have issues with importing torchmin, you may need to install from source, as shown below. Or you can try re-installing pytorch-minimize from source.

Installation from source

After cloning this repo, install the dependencies on the command-line via:

pip install -r requirements.txt

In this directory, run

pip install -e .


Import ConDo and create the adapter:

import condo
condoer = condo.ConDoAdapter()

Try using it:

import numpy as np
X_T = np.sort(np.random.uniform(0, 8, size=(100, 1)))
X_S = np.sort(np.random.uniform(4, 8, size=(100, 1)))
Y_T = np.random.normal(4 * X_T + 1, 1 * X_T + 1)
Y_Strue = np.random.normal(4 * X_S + 1, 1 * X_S + 1)
Y_S = 5 * Y_Strue + 2, Y_T, X_S, X_T)
Y_S2T = condoer.transform(Y_S)
print(f"before ConDo: {np.mean((Y_S - Y_Strue) ** 2):.3f}")
print(f"after ConDo:  {np.mean((Y_S2T - Y_Strue) ** 2):.3f}")

More thorough examples are provided in the examples directory.



In this directory run


Code formatting

The Uncompromising Code Formatter: Black
black {source_file_or_directory}

Install it into pre-commit hook to always commit well-formatted code:
pre-commit install

License Information

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.