Towards a neocortically-inspired ab initio cellular model of associative memory

Ab Initio Cellular Models;Associative Memory;Attractors;Complex Networks;Emergence;Information-Preserving;Multi-Scale Modeling;Neurorobotics;Nonlinear Dynamics;Real-World Coupling;Spiking Neural Networks;Spike-timing Dependent Plasticity;Spike-timing Dependent Plasticity (STDP) Learning;Topological Adaptation

Duda, Alexander Michael

关键词: Ab Initio Cellular Models;    Associative Memory;    Attractors;    Complex Networks;    Emergence;    Information-Preserving;    Multi-Scale Modeling;    Neurorobotics;    Nonlinear Dynamics;    Real-World Coupling;    Spiking Neural Networks;    Spike-timing Dependent Plasticity;    Spike-timing Dependent Plasticity (STDP) Learning;    Topological Adaptation;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/78735/DUDA-DISSERTATION-2015.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

We are interested in self-organization and adaptation in intelligent systems that are robustly coupled with the real world. Such systems have a variety of sensory inputs that provide access to the richness, complexity, and noise of real-world signals. Specifically, the systems we design and implement are ab initio (simulated) spiking neural networks (SNNs) with cellular resolution and complex network topologies that evolve according to spike-timing dependent plasticity (STDP). We desire to understand how external signals (like speech, vision, etc.)are encoded in the dynamics of such SNNs. In particular, we desire to identify and confirm the extent to which various network-level measurements are information-preserving and could be used as the basis of an associative memory. The dissertation details the relevant background and results of a series ofexperiments designed to accomplish this objective. The results provide encouraging empirical evidence that such a model can be used for encoding attractors with multi-sensory inputs and across sensory modalities, which both emphasize the potential of such a model for use as a multi-modal associative memory.