Search Papers | Poster Sessions | All Posters
Poster C26 in Poster Session C - Friday, August 9, 2024, 11:15 am – 1:15 pm, Johnson Ice Rink
Capacity of the Neuroidal Model for Shared Memory Representations
Patrick Perrine1 (), Chandradeep Chowdhury1, Mugizi Rwebangira1; 1California Polytechnic State University
A central question in neuroscience is how neuronal activity leads to higher level phenomena such as the formation of memories. The Neuroidal model was proposed as a general computational model for brain cognition. This model was later used to suggest how new memories might be created in the mammalian cortex (Valiant, 2005). This was one of many early quantitative theories of memory that used biologically plausible values for parameters such as number of neurons n, number of synapses per neuron d, and inverse synapse strength k. Yet, many fundamental questions remain about the properties of this model. For example, how many memories can be stored given a particular set of parameters? To better understand this question, we offer simple methods for theoretically and empirically evaluating the capacity of the Neuroidal model within specific contexts.
Keywords: neuroids memory capacity learning