Piecewise-linear recurrent neural network (`PLRNN`)

[MartinuzziFrancesco]

Paper

https://doi.org/10.1371/journal.pcbi.1007263

Equations

$$\begin{aligned} \mathbf{h}(t) &= \mathbf{A}\,\mathbf{h}(t-1) + \mathbf{W}\,\phi\!\big(\mathbf{h}(t-1)\big) + \mathbf{C}\,\mathbf{x}(t) + \mathbf{b} + \boldsymbol{\epsilon}(t), \\\ \boldsymbol{\epsilon}(t) &\sim \mathcal{N}\!\big(\mathbf{0},\,\boldsymbol{\Sigma}\big). \end{aligned}$$

Official implementation

https://github.com/DurstewitzLab/PLRNN_SSM