Academic CS in Production Code

Fractional Max-Pooling

Benjamin Graham
2014
21 references

Abstract

Convolutional networks almost always incorporate some form of spatial pooling, and very often it is alpha times alpha max-pooling with alpha=2. Max-pooling act on the hidden layers of the network, reducing their size by an integer multiplicative factor alpha. The amazing by-product of discarding 75% of your data is that you build into the network a degree of invariance with respect to translations and elastic distortions. However, if you simply alternate convolutional layers with max-pooling layers, performance is limited due to the rapid reduction in spatial size, and the disjoint nature of the pooling regions. We have formulated a fractional version of max-pooling where alpha is allowed to take non-integer values. Our version of max-pooling is stochastic as there are lots of different ways of constructing suitable pooling regions. We find that our form of fractional max-pooling reduces overfitting on a variety of datasets: for instance, we improve on the state-of-the art for CIFAR-100 without even using dropout.

View Paper PDF DOI
🤖 Machine Learning
2 repositories
17 references

Code References

â–¶ pytorch/pytorch
2 files
â–¶ torch/nn/functional.py
2
http://arxiv.org/abs/1412.6071
http://arxiv.org/abs/1412.6071
â–¶ torch/nn/modules/pooling.py
2
https://arxiv.org/abs/1412.6071
https://arxiv.org/abs/1412.6071
â–¶ tensorflow/tensorflow
4 files
â–¶ tensorflow/core/kernels/fractional_pool_common.cc
1
// [1]: https://arxiv.org/abs/1412.6071
â–¶ tensorflow/core/kernels/fractional_pool_common.h
1
// http://arxiv.org/abs/1412.6071
â–¶ tensorflow/go/op/wrappers.go
3
// Graham, Fractional Max-Pooling](http://arxiv.org/abs/1412.6071) for
// Graham, Fractional Max-Pooling](http://arxiv.org/abs/1412.6071) for
// [Benjamin Graham, Fractional Max-Pooling](http://arxiv.org/abs/1412.6071)
â–¶ tensorflow/python/ops/nn_ops.py
8
[Graham, 2015](https://arxiv.org/abs/1412.6071)
([pdf](https://arxiv.org/pdf/1412.6071.pdf))
[Graham, 2015](https://arxiv.org/abs/1412.6071)
([pdf](https://arxiv.org/pdf/1412.6071.pdf))
[Graham, 2015](https://arxiv.org/abs/1412.6071)
([pdf](https://arxiv.org/pdf/1412.6071.pdf))
[Graham, 2015](https://arxiv.org/abs/1412.6071)
([pdf](https://arxiv.org/pdf/1412.6071.pdf))
Link copied to clipboard!