Fast Modifications of the SpikeProp Algorithm

McKennoch, S.; Liu, Dingding; Bushnell, L.G.

doi:10.1109/ijcnn.2006.1716646

Cited by 34 publications

(50 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Although the gradient descent learning method has been tested with various benchmark problems [5], all other studies on SpikeProp ( [7], [8], [9], [10], [11], [12]) used the XOR problem or the Iris dataset with identical network structures as in the original paper on SpikeProp [5]. This gives little insight into the capabilities or limitations of a spiking neural network trained with SpikeProp.…”

Section: Reference Start Timementioning

confidence: 99%

See 1 more Smart Citation

Reference time in SpikeProp

Sporea

Grüning

2011

The 2011 International Joint Conference on Neural Networks

View full text Add to dashboard Cite

show abstract

Section: Reference Start Timementioning

confidence: 99%

“…16 (0) 16 (0) 16 (0) 10 (1) f 8 10 (1) 10 (1) 10 (1) 16 (0) f 9 16 (0) 10 (1) 10 (1) 16 (0) f 10 10 (1) 16 (0) 10 (1) 16 (0) f 11 16 (0) 16 (0) 10 (1) 16 (0) f 12 10 (1) 10 (1) …”

unclassified

Reference time in SpikeProp

Sporea

Grüning

2011

The 2011 International Joint Conference on Neural Networks

View full text Add to dashboard Cite

show abstract

“…However, there were several issues this algorithm needed to address, such as slow convergence especially for large datasets and the problem of non-firing (silent) neurons. Subsequently, several methods have been developed to improve SpikeProp [11][12][13][14][15]. These gradient based algorithms are computationally powerful but are often regarded as non-biologically plausible because they require a non-local spread of error signals from one synapse to another.…”

Section: Introductionmentioning

confidence: 99%

SpikeTemp: An Enhanced Rank-Order-Based Learning Approach for Spiking Neural Networks With Adaptive Structure

Wang

Belatreche

Maguire

et al. 2017

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

The authors would like to thank the reviewers for taking the time to review our manuscript and for helping us improve our work. We appreciate the reviewer's very positive comments and we are glad to learn that the reviewers are now satisfied with our work.We would also like to thank the Associate Editor for the comment regarding the referencing style. We have thoroughly checked all references and made sure that all of the references now follow the IEEE referencing style as per the referencing style note provided by the Associate Editor. The proposed SpikeTemp algorithm is demonstrated on several benchmark datasets and on an image recognition task. The results show that SpikeTemp can achieve better classification performance and is much faster than the existing rank-order based learning approach. In addition, the number of output neurons is much smaller when the square cosine encoding scheme is employed. Furthermore, SpikeTemp is benchmarked against a selection of existing machine learning algorithms and the results demonstrate the ability of SpikeTemp to classify different datasets after just one presentation of the training samples with comparable classification performance.

show abstract

“…This is referred to as the 'silent neuron' problem and has been discussed in detail in [8]. Several extensions to SpikeProp have been proposed which improve its convergence characteristics by adding a momentum term [9,10] and by using adaptive learning rates [11]. Other variations of SpikeProp have also been proposed which are capable of learning with multiple spikes [12][13][14].…”

Section: Introductionmentioning

confidence: 99%