LIMITATIONS ON MEMORY LOADING IN AN ARTIFICIAL SIMULATION OF A CORTICAL NETWORK
The objective of this project were as follows:
1. research into biological background to the paper ” learning in networks of cortical neurons ” by shahaf and marom
2. set up an artificial neural network to simulate learning in an isolated cortical network
3. investigate memory loading in this simulation
4. if loading of single memories achieved, extend the study to ascertain the possible limitations on multiple memory loading in the artificial network.
In this project report, i have summarized the relevant area of the background research. these were essential for a successful solution to the problem . the method by which i achieved the objectives was to set up an artificial neural network. i configured the network to simulate the learning of an isolated cortical network. i then investigated memory loading in this model system. this involved through investigation into the different types of artificial neural networks and learning algorithms and then selection or creation of a simulation that was as close to the biological case as possible. this same technique was used to develop a learning algorithm that was a biologically plausible as possible. i then tested the solution to ascertain whether single memory loading was possible and modifications were made as required. once single memory loading had been achieved. i investigated multiple memory loading and attempted to load multiple memories.