Analysis of Liquid Ensembles for Enhancing the Performance and Accuracy of Liquid State Machines
| Autor: | Gopalakrishnan Srinivasan, Priyadarshini Panda, Parami Wijesinghe, Kaushik Roy |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2019 |
| Předmět: |
discriminant ratio
Liquid state machine Computer science liquid state machines ensembles 02 engineering and technology lcsh:RC321-571 03 medical and health sciences 0302 clinical medicine 0202 electrical engineering electronic engineering information engineering separation property Representation (mathematics) lcsh:Neurosciences. Biological psychiatry. Neuropsychiatry Separation property Original Research Spiking neural network Artificial neural network General Neuroscience Process (computing) Range (mathematics) Recurrent neural network approximation property spiking neural networks 020201 artificial intelligence & image processing Algorithm 030217 neurology & neurosurgery Neuroscience |
| Zdroj: | Frontiers in Neuroscience, Vol 13 (2019) Frontiers in Neuroscience |
| DOI: | 10.3389/fnins.2019.00504/full |
| Popis: | Liquid state machine (LSM), a bio-inspired computing model consisting of the input sparsely connected to a randomly interlinked reservoir (or liquid) of spiking neurons followed by a readout layer, finds utility in a range of applications varying from robot control and sequence generation to action, speech, and image recognition. LSMs stand out among other Recurrent Neural Network (RNN) architectures due to their simplistic structure and lower training complexity. Plethora of recent efforts have been focused towards mimicking certain characteristics of biological systems to enhance the performance of modern artificial neural networks. It has been shown that biological neurons are more likely to be connected to other neurons in the close proximity, and tend to be disconnected as the neurons are spatially far apart. Inspired by this, we propose a group of locally connected neuron reservoirs, or an ensemble of liquids approach, for LSMs. We analyze how the segmentation of a single large liquid to create an ensemble of multiple smaller liquids affects the latency and accuracy of an LSM. In our analysis, we quantify the ability of the proposed ensemble approach to provide an improved representation of the input using the Separation Property (SP) and Approximation Property (AP). Our results illustrate that the ensemble approach enhances class discrimination (quantified as the ratio between the SP and AP), leading to better accuracy in speech and image recognition tasks, when compared to a single large liquid. Furthermore, we obtain performance benefits in terms of improved inference time and reduced memory requirements, due to lowered number of connections and the freedom to parallelize the liquid evaluation process. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|