Implementation of our CVPR 2018 paper "Rethinking Feature Distribution for Loss Functions in Image Classification".
Paper authors: Weitao Wan, Yuanyi Zhong, Tianpeng Li, Jiansheng Chen.

We implemented it in Caffe. I also have a tensorflow implementation but it hasn't been fully tested yet. Now I'm rearranging the code to make it look neat and (hopefully) a bit more beautiful.

Code is written by Yuanyi Zhong and Weitao Wan.

We propose a large-margin Gaussian Mixture (L-GM) loss for deep neural networks in classification tasks. Different from the softmax cross-entropy loss, our proposal is established on the assumption that the deep features of the training set follow a Gaussian Mixture distribution. By involving a classification margin and a likelihood regularization, the L-GM loss facilitates both a high classification performance and an accurate modeling of the training feature distribution. As such, the L-GM loss is superior to the softmax loss and its major variants in the sense that besides classification, it can be readily used to distinguish abnormal inputs, such as the adversarial examples, based on their features' likelihood to the training feature distribution. Extensive experiments on various recognition benchmarks like MNIST, CIFAR, ImageNet and LFW, as well as on adversarial examples demonstrate the effectiveness of our proposal.

• Install this caffe
• Examples for CIFAR-100 in ./examples/cifar100

  ./train.sh 0 simple  # 0 is the GPU id, simple is the folder containing network definitions and solver

• Specify margin parameter α and likelihood weight λ, which is margin_mul and center_coef in the layer param, respectively.

    margin_mul {
      policy: STEPUP
      value: 0.1
      step: 5000 
      gamma: 2
      max: 0.3 
}

This specifies a gradually growing value for α, which is helpful for training.

• other indicators

update_sigma: false

Fix the variances to initial values (1.0).

isotropic: true

The variances of different dimensions are identical.

More contents under construction ......

We've described how the data is pre-processed in our paper. For example, the CIFAR-100 training data (32x32) is padded to 40x40 and then randomly cropped with a 32x32 window for training.
In the CIFAR-100 example, we use data in HDF5 format. You can choose other formats, changing the data layer accordingly.

If you find this work useful, please consider citing it.

@article{LGM2018,
  title={Rethinking Feature Distribution for Loss Functions in Image Classification},
  author={Wan, Weitao and Zhong, Yuanyi and Li, Tianpeng and Chen, Jiansheng},
  journal={arXiv preprint arXiv:1803.02988},
  year={2018}
}