Code for Z. Monfared & D. Durstewitz (2020), Transformation of ReLU-based recurrent neural networks from discrete-time to continuous-time, Proceedings of the 37th International Conference on Machine Learning (ICML). (c) Monfared & Durstewitz, CC-Attrib. Lic.
Notes on files:
- Files starting 'FlowPLRNN...' run the discrete & cont. time models and produce the actual figs. from the paper for the 4 examples (vdP, add. prob., Lorenz, fMRI)
- Files starting 'contPLRNN...' contain the continuous-time conversion of discrete PLRNN
- Files starting 'ffcPLRNN...' produce flow fields
- .mat files contain data or actual discrete-PLRNN parameters from Koppe et al. (2019), see folder PLRNN_SSM
Git stats
1 commits 1 branch 0 tags