This repository contains the code and data used to obtain the results reported in the following paper:
Orhan AE, Ma WJ (2017) Efficient probabilistic inference in generic neural networks trained with non-probabilistic feedback. Nature Communications. [arXiv:1601.03060]
The bulk of the code is written in Theano + Lasagne. Each folder contains code and data pertaining to a particular model type or experiment.
The code for generating the results reported in Figure 9 (qamar2013) is written in Matlab and uses some routines from the Matlab Statistics and Optimization Toolboxes. Some of the files are written to be run on a local computer cluster. You may need to modify them according to your needs.