Niels Leadholm's Projects
MATLAB code used to study coincidence-detecting networks that model feature binding, and what condictions cause them to fail.
TensorFlow implementation of CNNs exploring hierarchical feature binding for robustness.
Uses Brian spiking neural network simulator to explore the encoding of objects using delay-line computations.
Numenta published papers code and data
Experimental algorithms. Unsupported.
Numenta Platform for Intelligent Computing Tensorflow libraries
Numenta Platform for Intelligent Computing PyTorch libraries
C++ code for helper functions to enable creating simple, hierarchical polychronous neural network models to investigate the emergence of feature binding.
Python code for helper functions to enable creating simple, hierarchical polychronous neural network models to investigate the emergence of feature binding.
C++ code for creating and simulating polychronous neural networks.
MATLAB code used to study models in the predictive coding framework on a probabilistic task (see Bogacz 2017, 'A Tutorial on the Free-Energy...')
The following is a reading created from the rearrangement and modification of the words of Carl Sagan and Ann Druyan.
General Game Playing with Schema Networks