/EMPJ

Everything needed to replicate the figures from the paper "Engineering recurrent neural networks from task-relevant manifolds and dynamics" (Pollock and Jazayeri, 2019)

Primary LanguageJupyter NotebookMIT LicenseMIT

Embedding Manifolds with Population-level Jacobians (EMPJ)

Everything needed to replicate the figures from the paper "Engineering recurrent neural networks from task-relevant manifolds and dynamics" (Pollock and Jazayeri, 2019)

The Jupyter notebook contains an implementation of the method and reproduces the figures from the paper.