Research at daenet - New paper published today at ICPRAM 2021 - Vienna
Today we have presented our new paper in the field of computational intelligence at the internation conference for pattern recognition and models in Vienna in the category machine learning theory and models.
I research in the field of Hierarchical Temporal Memory In this work, I have found how to overcame the instability of the Spatial Pooler algorithm. This algorithm emulates the spatial learning cortical algorithm of the neo-cortex.
We have created the the algorithm that controlls the brain's homeostatic plasticity inside of the Spatial Pooler. This improves the performance and stability of the existing algorithm and shows how lesseons learned from biology can be applied to computational sciences.
Download paper from out open source project neocortexapi.