accelbrainbase.observabledata._mxnet.adversarialmodel package

Submodules

accelbrainbase.observabledata._mxnet.adversarialmodel.discriminative_model module

class accelbrainbase.observabledata._mxnet.adversarialmodel.discriminative_model.DiscriminativeModel(model, initializer=None, learning_rate=1e-05, optimizer_name='SGD', hybridize_flag=True, scale=1.0, ctx=gpu(0), **kwargs)

Bases: mxnet.gluon.block.HybridBlock, accelbrainbase.observabledata._mxnet.adversarial_model.AdversarialModel

Discriminative model, which discriminates true from fake, in the Generative Adversarial Networks(GANs).

The Generative Adversarial Networks(GANs) (Goodfellow et al., 2014) framework establishes a min-max adversarial game between two neural networks – a generative model, G, and a discriminative model, D. The discriminator model, D(x), is a neural network that computes the probability that a observed data point x in data space is a sample from the data distribution (positive samples) that we are trying to model, rather than a sample from our generative model (negative samples).

Concurrently, the generator uses a function G(z) that maps samples z from the prior p(z) to the data space. G(z) is trained to maximally confuse the discriminator into believing that samples it generates come from the data distribution. The generator is trained by leveraging the gradient of D(x) w.r.t. x, and using that to modify its parameters.

References

  • Gauthier, J. (2014). Conditional generative adversarial nets for convolutional face generation. Class Project for Stanford CS231N: Convolutional Neural Networks for Visual Recognition, Winter semester, 2014(5), 2.
  • Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., … & Bengio, Y. (2014). Generative adversarial nets. In Advances in neural information processing systems (pp. 2672-2680).
  • Makhzani, A., Shlens, J., Jaitly, N., Goodfellow, I., & Frey, B. (2015). Adversarial autoencoders. arXiv preprint arXiv:1511.05644.
  • Mirza, M., & Osindero, S. (2014). Conditional generative adversarial nets. arXiv preprint arXiv:1411.1784.
  • Salimans, T., Goodfellow, I., Zaremba, W., Cheung, V., Radford, A., & Chen, X. (2016). Improved techniques for training gans. In Advances in neural information processing systems (pp. 2234-2242).
  • Zhao, J., Mathieu, M., & LeCun, Y. (2016). Energy-based generative adversarial network. arXiv preprint arXiv:1609.03126.
  • Warde-Farley, D., & Bengio, Y. (2016). Improving generative adversarial networks with denoising feature matching.
collect_params(select=None)

Overrided collect_params in mxnet.gluon.HybridBlok.

forward_propagation(F, x)

Hybrid forward with Gluon API.

Parameters:
  • F – mxnet.ndarray or mxnet.symbol.
  • x – mxnet.ndarray of observed data points.
Returns:

mxnet.ndarray or mxnet.symbol of inferenced feature points.
get_model()

getter for mxnet.gluon.hybrid.hybridblock.HybridBlock.

hybrid_forward(F, x)

Hybrid forward with Gluon API.

Parameters:
  • F – mxnet.ndarray or mxnet.symbol.
  • x – mxnet.ndarray of observed data points.
Returns:

mxnet.ndarray or mxnet.symbol of inferenced feature points.
inference(observed_arr)

Draw samples from the fake distribution.

Parameters:observed_arr – mxnet.ndarray or mxnet.symbol of observed data points.
Returns:Tuple of `mxnet.ndarray`s.
model

getter for mxnet.gluon.hybrid.hybridblock.HybridBlock.

set_model(value)

setter for mxnet.gluon.hybrid.hybridblock.HybridBlock.

accelbrainbase.observabledata._mxnet.adversarialmodel.generative_model module

class accelbrainbase.observabledata._mxnet.adversarialmodel.generative_model.GenerativeModel(noise_sampler, model, initializer=None, condition_sampler=None, conditonal_dim=1, learning_rate=1e-05, optimizer_name='SGD', hybridize_flag=True, scale=1.0, ctx=gpu(0), **kwargs)

Bases: mxnet.gluon.block.HybridBlock, accelbrainbase.observabledata._mxnet.adversarial_model.AdversarialModel

Generative model, which draws samples from the fake distribution, in the Generative Adversarial Networks(GANs).

The Generative Adversarial Networks(GANs) (Goodfellow et al., 2014) framework establishes a min-max adversarial game between two neural networks – a generative model, G, and a discriminative model, D. The discriminator model, D(x), is a neural network that computes the probability that a observed data point x in data space is a sample from the data distribution (positive samples) that we are trying to model, rather than a sample from our generative model (negative samples).

Concurrently, the generator uses a function G(z) that maps samples z from the prior p(z) to the data space. G(z) is trained to maximally confuse the discriminator into believing that samples it generates come from the data distribution. The generator is trained by leveraging the gradient of D(x) w.r.t. x, and using that to modify its parameters.

References

  • Gauthier, J. (2014). Conditional generative adversarial nets for convolutional face generation. Class Project for Stanford CS231N: Convolutional Neural Networks for Visual Recognition, Winter semester, 2014(5), 2.
  • Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., … & Bengio, Y. (2014). Generative adversarial nets. In Advances in neural information processing systems (pp. 2672-2680).
  • Makhzani, A., Shlens, J., Jaitly, N., Goodfellow, I., & Frey, B. (2015). Adversarial autoencoders. arXiv preprint arXiv:1511.05644.
  • Mirza, M., & Osindero, S. (2014). Conditional generative adversarial nets. arXiv preprint arXiv:1411.1784.
  • Salimans, T., Goodfellow, I., Zaremba, W., Cheung, V., Radford, A., & Chen, X. (2016). Improved techniques for training gans. In Advances in neural information processing systems (pp. 2234-2242).
  • Zhao, J., Mathieu, M., & LeCun, Y. (2016). Energy-based generative adversarial network. arXiv preprint arXiv:1609.03126.
  • Warde-Farley, D., & Bengio, Y. (2016). Improving generative adversarial networks with denoising feature matching.
collect_params(select=None)

Overrided collect_params in mxnet.gluon.HybridBlok.

conditon_sampler

getter for ConditionSampler of sampler to draw conditons from user-defined distributions.

draw()

Draw samples from the fake distribution.

Returns:Tuple of `mxnet.ndarray`s.
get_condition_sampler()

getter for ConditionSampler of sampler to draw conditons from user-defined distributions.

get_model()

getter for mxnet.gluon.hybrid.hybridblock.HybridBlock.

get_noise_sampler()

getter

model

getter for mxnet.gluon.hybrid.hybridblock.HybridBlock.

noise_sampler

getter

set_condition_sampler(value)

setter for ConditionSampler of sampler to draw conditons from user-defined distributions.

set_model(value)

setter for mxnet.gluon.hybrid.hybridblock.HybridBlock.

set_noise_sampler(value)

setter

Module contents