full-FORCE Learning Tutorial

Presented on 8/7/19 @ MIT for the Harvard-MIT Theoretical and Computational Neuroscience Journal Club

Date: August 7, 2019

Authors: Eli Pollock (MIT)

Instructions

1. The code uses Python 3, which can be downloaded here

2. You'll also need Jupyter to run the interactive notebooks.

Background reading

Recurrent neural networks (RNNs) are a popular kind of model in computational neuroscience. A number of algorithms now exist for training RNNs. This tutorial focuses on a type of training algorithm originally introduced in Sussillo and Abbott, 2009 and expanded upon in DePasquale et al., 2018.