Improving Deep Neural Networks with Probabilistic Maxout Units

We present a probabilistic variant of the recently introduced maxout unit. The success of deep neural networks utilizing maxout can partly be attributed to favorable performance under dropout, when compared to rectified linear units. It however also depends on the fact that each maxout unit performs a pooling operation over a group of linear transformations and is thus partially invariant to changes in its input. Starting from this observation we ask the question: Can the desirable properties of maxout units be preserved while improving their invariance properties ? We argue that our probabilistic maxout (probout) units successfully achieve this balance. We quantitatively verify this claim and report classification performance matching or exceeding the current state of the art on three challenging image classification benchmarks (CIFAR-10, CIFAR-100 and SVHN).

[1]  Nitish Srivastava,et al.  Discriminative Transfer Learning with Tree-based Priors , 2013, NIPS.

[2]  Brendan J. Frey,et al.  Adaptive dropout for training deep neural networks , 2013, NIPS.

[3]  Rob Fergus,et al.  Visualizing and Understanding Convolutional Networks , 2013, ECCV.

[4]  Razvan Pascanu,et al.  Learned-norm pooling for deep neural networks , 2013, ArXiv.

[5]  Mario Fritz,et al.  Learnable Pooling Regions for Image Classification , 2013, ICLR.

[6]  Trevor Darrell,et al.  Beyond spatial pyramids: Receptive field learning for pooled image features , 2012, 2012 IEEE Conference on Computer Vision and Pattern Recognition.

[7]  Yoshua Bengio,et al.  Gradient-based learning applied to document recognition , 1998, Proc. IEEE.

[8]  Alex Krizhevsky,et al.  Learning Multiple Layers of Features from Tiny Images , 2009 .

[9]  Yoshua Bengio,et al.  Deep Sparse Rectifier Neural Networks , 2011, AISTATS.

[10]  Yoshua Bengio,et al.  Slow, Decorrelated Features for Pretraining Complex Cell-like Networks , 2009, NIPS.

[11]  Yoshua Bengio,et al.  Deep Generative Stochastic Networks Trainable by Backprop , 2013, ICML.

[12]  Rob Fergus,et al.  Stochastic Pooling for Regularization of Deep Convolutional Neural Networks , 2013, ICLR.

[13]  Yann LeCun,et al.  Regularization of Neural Networks using DropConnect , 2013, ICML.

[14]  Nitish Srivastava,et al.  Improving neural networks by preventing co-adaptation of feature detectors , 2012, ArXiv.

[15]  Yoshua Bengio,et al.  Estimating or Propagating Gradients Through Stochastic Neurons , 2013, ArXiv.

[16]  Jasper Snoek,et al.  Practical Bayesian Optimization of Machine Learning Algorithms , 2012, NIPS.

[17]  Nitish Srivastava,et al.  Improving Neural Networks with Dropout , 2013 .

[18]  R. Fergus,et al.  Learning invariant features through topographic filter maps , 2009, 2009 IEEE Conference on Computer Vision and Pattern Recognition.

[19]  Honglak Lee,et al.  Convolutional deep belief networks for scalable unsupervised learning of hierarchical representations , 2009, ICML '09.

[20]  Shenghuo Zhu,et al.  Deep Learning of Invariant Features via Simulated Fixations in Video , 2012, NIPS.

[21]  Peter Kulchyski.  and , 2015 .

[22]  Razvan Pascanu,et al.  Pylearn2: a machine learning research library , 2013, ArXiv.

[23]  Andrew Y. Ng,et al.  Reading Digits in Natural Images with Unsupervised Feature Learning , 2011 .

[24]  Geoffrey E. Hinton,et al.  Rectified Linear Units Improve Restricted Boltzmann Machines , 2010, ICML.