Inferring-TAPBrain

Code for training and inferring the TAP model brain.

Dependencies: Python 3.5.6 or newer, PyTorch 1.5.1 or newer.

Description of contents

code

This folder contains the main python code:

• rnnmodel.py : RNN model for the TAP brain
• tapdynamics.py : functions for generating the TAP dynamics
• particlefilter.py : particle filter and Q function required for particle-EM
• notebookutils.py : functions for loading a TAP brain, generating model parameters, etc
• utils.py : all other required utils

data/brains

This folder contains the parameters for a few example TAP brains.

data/pem-estimates

This folder contains particle-EM estimates for a few example TAP brains.

notebooks

• TrainTAPbrain.ipynb : example notebook that illustrates the training of a TAP brain model
• InferTAPbrain.ipynb : example notebook that illustrates inferring inference in a TAP brain model
• GreedyBackwardSearch.ipynb: example notebook that illustrates the greedy approach to find the best set of G and J parameters

scripts

• InferTAPbrain.py : script for inferring inference in a TAP brain model
• runscripts.sh: shell script for launching multiple inference jobs in parallel