Code for project "Deep Learning for Clustering" under lab course "Deep Learning for Computer Vision and Biomedicine" - TUM. Depends on numpy, theano, lasagne, scikit-learn, matplotlib.

• Mohd Yawar Nihal Siddiqui
• Elie Aljalbout
• Vladimir Golkov (Supervisor)

This repository is an implementation of the paper : Elie Aljalbout, Vladimir Golkov, Yawar Siddiqui, Daniel Cremers "Clustering with Deep Learning: Taxonomy and new methods"

• arxiv: https://arxiv.org/abs/1801.07648

Use the main script for training, visualizing clusters and/or reporting clustering metrics

python main.py <options>

We've implemented a NetworkBuilder class that can be used to quickly describe the architecture of an autoencoder through a json file. The json specification of the architecture is a dictionary with the following fields

Further, each item in the layers list is a dictionary with the following fields

Only the encoder part of the autoencoder needs to be specified, the decoder will be automatically generated by the class.

Example of a network description

{
    "name": "c-5-6_p_c-5-16_p_c-4-120",
    "use_batch_norm": 1,
    "batch_size": 100,
    "layers": [
      {
        "type": "Input",
        "output_shape":[1, 28, 28]
      },
      {
        "type": "Conv2D",
        "num_filters": 50,
        "filter_size": [5, 5],
        "non_linearity": "rectify"
      },
      {
        "type": "MaxPool2D*",
        "filter_size": [2, 2]
      },
      {
        "type": "Conv2D",
        "num_filters": 50,
        "filter_size": [5, 5],
        "non_linearity": "rectify"
      },
      {
        "type": "MaxPool2D*",
        "filter_size": [2, 2]
      },
      {
        "type": "Conv2D",
        "num_filters": 120,
        "filter_size": [4, 4],
        "non_linearity": "linear"
      }
    ]
  }

This would generate the network 50[5x5] 50[5x5]_bn max[2x2] 50[5x5] 50[5x5]_bn max[2x2] **120[4x4] 120[4x4]_bn **50[4x4] 50[4x4]_bn ups*[2x2] 50[5x5] 50[5x5]_bn ups*[2x2] 1[5x5]

We trained and tested the network on two datasets - MNIST and COIL20

Clustering was performed with two different loss functions -

• Loss = KL-Divergence(soft assignment distribution, target distribution) Autoencoder Reconstruction loss , where the target distribution is a distribution that improves cluster purity and puts more emphasis on data points assigned with a high confidence. For more details check out the DEC paper [1].
• Loss = k-Means loss Autoencoder Reconstruction loss

Note: The commit b34743114f68624b5371cd0d4c059b141422902f gives upto 0.96 accuracy and 0.92 NMI on the MNIST dataset. We will include it to the main branch once we can get better results with the COIL architecture