A Spiking Network that Learns to Extract Spike Signatures from Speech Signals

Spiking neural networks (SNNs) with adaptive synapses reflect core properties of biological neural networks. Speech recognition, as an application involving audio coding and dynamic learning, provides a good test problem to study SNN functionality. We present a novel and efficient method that learns to convert a speech signal into a spike train signature. The signature is distinguishable from signatures for other speech signals, representing different words, thereby enabling digit recognition and discrimination in devices that use only spiking neurons. The method uses a small SNN consisting of Izhikevich neurons equipped with spike timing dependent plasticity (STDP) and biologically realistic synapses. This approach introduces an efficient, fast, and multi-speaker strategy without error-feedback training, although it does require supervised training. The new simulation results produce discriminative spike train patterns for spoken digits in which highly correlated spike trains belong to the same category and low correlated patterns belong to different categories. The signatures can be used as a set of feature maps for classification. Our experiments compare two simple classifiers for spoken digit recognition in both clean and noisy environments.