Theoretical-Neuroscience-Group / synaptic_filter

Code for running the synaptic filter: https://arxiv.org/abs/2008.03198

DOI: 10.5281/zenodo.3970145

(OLD) Generating and plotting of figures for "Learning as filtering" manuscript.

To plot, provide one of the following figure keys (str) as -f argument: fig1d, fig2a, fig2b, fig2c, fig2d, fig2e, fig3, fig4, figS1, figS2, figS3, figS4, figS5

To generate new figure-data, use the production keys (str) as -p arguments: fig1d, fig2_dim, fig2_beta, fig2_dim_pf, fig2_beta_pf, fig2_eta, fig2d, fig2e, fig3, fig4, figS4, figS5

Command line example for generating fig1d data and plot:

python main.py -f fig1d -p fig1d

The plot is saved as ./figures/fig1d.pdf The data (a pandas data frame) is stored as ./pkl_data/fig1d/fig1d.pkl

Further details:

This file contains 2 simulation environments, one for the biological &
one for the performance oriented simulations. Parameters are set in
three layers. Lower layers have priority.
1. default parameters apply to all simulations
2. simulation type parameters apply either to bio- or performance sims
3. for each figure, specpfic parameters can be selected

Plotting parameters (labels, line color ect) must be tuned directly in
the function "plt_manuscript_figures" in the file "./util/util.py"