Structure learning in convolutional neural networks
The present disclosure provides an improved approach to implement structure learning of neural networks by exploiting correlations in the data/problem the networks aim to solve. A greedy approach is described that finds bottlenecks of information gain from the bottom convolutional layers all the way to the fully connected layers. Rather than simply making the architecture deeper, additional computation and capacitance is only added where it is required.
1. A method implemented with a processor, comprising:
creating a neural network comprising a plurality of layers;
identifying an extraneous layer from the plurality of layers;
removing the extraneous layer from the neural network; and
identifying and removing other extraneous layers from the plurality of layers until no other extraneous layer is identified,
wherein the neural network undergoes both vertical splitting and horizontal splitting.
2. The method of claim 1 , wherein an all-or-nothing highway network is employed to identify the extraneous layer in the neural network to be removed.
3. The method of claim 2 , further comprising associating a penalty with use of the extraneous layer.
4. The method of claim 3 , wherein a size of the penalty corresponds to an amount of computational power used by the extraneous layer.
5. The method of claim 4 , wherein a size of the penalty is varied.
6. The method of claim 3 , wherein the penalty is set to 0 when the processor is part of a cloud computing platform.
7. The method of claim 2 , wherein the all-or-nothing highway network generates a binary decision to not use the extraneous layer.
8. The method of claim 2 , further comprising the all-or-nothing highway network introducing a mixing matrix to determine how to transform a skip connection corresponding to the extraneous layer.
9. The method of claim 8 , further comprising parameterizing the mixing matrix by a scalar value.
10. The method of claim 9 , wherein when the scalar value is zero, an input to the extraneous layer is passed onto an output from the extraneous layer, thereby skipping the extraneous layer before removing the extraneous layer form the neural network.
11. The method of claim 2 , further comprising using identity mapping to transform a skip connection corresponding to the extraneous layer.
12. The method of claim 1 , further comprising updating a model for the neural network to obtain an updated model, wherein the extraneous layer is removed from the model to obtain the updated model.
13. The method of claim 1 , wherein a plurality of loss layers are added to the neural network.
14. The method of claim 13 , further comprising generating predictions at one of the loss layers, and converting the predictions to one or more confusion matrices forming a tensor T,
wherein a structure of T is analyzed to modify and augment an existing structure of the neural network both in terms of depth and breadth.
15. The method of claim 1 , wherein K-way Bifurcation is performed to implement the horizontal splitting.
16. The method of claim 1 , wherein the neural network is employed to classify images captured for a virtual realty or augmented reality system.
17. The method of claim 1 , wherein each layer of the neural network is addressed independently, and a given layer of the neural network undergoes splitting by performing a greedy choice to split the given layer which provides a best improvement on a training loss.
18. A system, comprising:
a processor;
a memory for holding programmable code; and
wherein the programmable code includes instructions for creating a neural network comprising a plurality of layers; identifying an extraneous layer from the plurality of layers; removing the extraneous layer from the neural network; and identifying and removing other extraneous layers from the plurality of layers until no other extraneous layer is identified,
wherein the neural network undergoes both vertical splitting and horizontal splitting.
19. A computer program product embodied on a non-transitory computer readable medium, the non-transitory computer readable medium having stored thereon a sequence of instructions which, when executed by a processor causes the processor to execute a method comprising:
creating a neural network comprising a plurality of layers;
identifying an extraneous layer from the plurality of layers;
removing the extraneous layer from the neural network; and
identifying and removing other extraneous layers from the plurality of layers until no other extraneous layer is identified,
wherein the neural network undergoes both vertical splitting and horizontal splitting.