This repository contains code for the paper Learning by Association - A versatile semi-supervised training method for neural networks (CVPR 2017) and the follow-up works Associative Domain Adaptation (ICCV 2017) and [Associative Deep Clustering - Training a classification network with no labels (GCPR 2018)]

It is implemented with TensorFlow. Please refer to the TensorFlow documentation for further information.

The core functions are implemented in semisup/backend.py. The files train.py and eval.py demonstrate how to use them. A quick example is contained in mnist_train_eval.py.

To run unsupervised (clustering) mode, use the train_unsup2.py script. For reference see also our paper. An example command with hyperparameters will be added soon.

In order to reproduce the results from the paper, please use the architectures and pipelines from the {stl10,svhn,synth}_tools.py. They are loaded automatically by setting the flag package in {train,eval}.py accordingly.

Before you get started, please make sure to add the following to your ~/.bashrc:

export PYTHONPATH=/path/to/learning_by_association:$PYTHONPATH

Copy the file semisup/tools/data_dirs.py.template to semisup/tools/data_dirs.py, adapt the paths and .gitignore this file.

If you use the code, please cite the paper "Learning by Association - A versatile semi-supervised training method for neural networks" or "Associative Domain Adaptation":

@string{cvpr="IEEE Conference on Computer Vision and Pattern Recognition (CVPR)"}
@InProceedings{haeusser-cvpr-17,
  author = 	 "P. Haeusser and A. Mordvintsev and D. Cremers",
  title = 	 "Learning by Association - A versatile semi-supervised training method for neural networks",
  booktitle = cvpr,
  year = 	 "2017",
}

@string{iccv="IEEE International Conference on Computer Vision (ICCV)"}
@InProceedings{haeusser-iccv-17,
  author = 	 "P. Haeusser and T. Frerix and A. Mordvintsev and D. Cremers",
  title = 	 "Associative Domain Adaptation",
  booktitle = iccv,
  year = 	 "2017",
}

For questions please contact Philip Haeusser ([email protected]).