ML p(r)ior | Unsupervised Regenerative Learning of Hierarchical Features in Spiking Deep Networks for Object Recognition

Unsupervised Regenerative Learning of Hierarchical Features in Spiking Deep Networks for Object Recognition

2016-02-03
1602.01510 | cs.NE
We present a spike-based unsupervised regenerative learning scheme to train Spiking Deep Networks (SpikeCNN) for object recognition problems using biologically realistic leaky integrate-and-fire neurons. The training methodology is based on the Auto-Encoder learning model wherein the hierarchical network is trained layer wise using the encoder-decoder principle. Regenerative learning uses spike-timing information and inherent latencies to update the weights and learn representative levels for each convolutional layer in an unsupervised manner. The features learnt from the final layer in the hierarchy are then fed to an output layer. The output layer is trained with supervision by showing a fraction of the labeled training dataset and performs the overall classification of the input. Our proposed methodology yields 0.92%/29.84% classification error on MNIST/CIFAR10 datasets which is comparable with state-of-the-art results. The proposed methodology also introduces sparsity in the hierarchical feature representations on account of event-based coding resulting in computationally efficient learning.
PDF

Highlights - Most important sentences from the article

Login to like/save this paper, take notes and configure your recommendations

Related Articles

2015-12-10

Deeper neural networks are more difficult to train. We present a residual learning framework to ease… show more
PDF

Highlights - Most important sentences from the article

2016-08-22

We propose local binary convolution (LBC), an efficient alternative to convolutional layers in stand… show more
PDF

Highlights - Most important sentences from the article

2016-08-31

The success of CNNs in various applications is accompanied by a significant increase in the computat… show more
PDF

Highlights - Most important sentences from the article

2012-07-03

When a large feedforward neural network is trained on a small training set, it typically performs po… show more
PDF

Highlights - Most important sentences from the article

2016-11-16

This paper proposes a new method, that we call VisualBackProp, for visualizing which sets of pixels … show more
PDF

Highlights - Most important sentences from the article

2015-11-21

Convolutional Neural Networks spread through computer vision like a wildfire, impacting almost all v… show more
PDF

Highlights - Most important sentences from the article

2016-06-27
1606.08165 | cs.NE

Gradient descent training techniques are remarkably successful in training analog-valued artificial … show more
PDF

Highlights - Most important sentences from the article

2016-06-23
1606.07326 | cs.CV

Deep learning using multi-layer neural networks (NNs) architecture manifests superb power in modern … show more
PDF

Highlights - Most important sentences from the article

2019-01-28

Spiking neural networks are nature's versatile solution to fault-tolerant and energy efficient signa… show more
PDF

Highlights - Most important sentences from the article

2017-06-08

We present a general framework for training deep neural networks without backpropagation. This subst… show more
PDF

Highlights - Most important sentences from the article

2018-11-27

A growing body of work underlines striking similarities between biological neural networks and recur… show more
PDF

Highlights - Most important sentences from the article

2017-09-13

Following their success in Computer Vision and other areas, deep learning techniques have recently b… show more
PDF

Highlights - Most important sentences from the article

2018-05-21
1805.07866 | cs.NE

Spiking neural networks (SNNs) are positioned to enable spatio-temporal information processing and u… show more
PDF

Highlights - Most important sentences from the article

2018-03-31

The primate visual system has inspired the development of deep artificial neural networks, which hav… show more
PDF

Highlights - Most important sentences from the article

2018-09-05
1810.08646 | cs.NE

Configuring deep Spiking Neural Networks (SNNs) is an exciting research avenue for low power spike e… show more
PDF

Highlights - Most important sentences from the article

2019-04-12

In recent years, Spiking Neural Networks (SNNs) have demonstrated great successes in completing vari… show more
PDF