Reinforcement Learning in Spiking Neural Networks
Department
Computer Science
Major
Computer Science
Research Advisor
Taylor, Patrick
Advisor's Department
Computer Science
Funding Source
OURE
Abstract
Biological neurons communicate primarily via a spiking process. Recurrently connected spiking neural networks (RSNNs) more realistically model the brain, compared to their non-spiking counterparts. It is of great interest to discover a biologically realistic learning rule to achieve optimal levels of performance on machine learning tasks. Experimental data describe a phenomenon known as spike-timing-dependent-plasticity (STDP), which integrates local firing coincidences between neurons to learn. STDP is believed to underlie memory formation and storage within the brain. When a reward signal modulates STDP, it enables forming associative memories via operant conditioning. Neuromodulators like dopamine operate similarly in the brain. We employ processes like synaptic scaling to support R-STDP in large, unstructured RSNNs. Doing so produces an agent that achieves adequate performance on reinforcement learning tasks.
Biography
Darrien McKenzie is a junior Computer Science student at Missouri University of Science & Technology (MST) expecting to graduate with a bachelors in Spring 2023. Before he transferred to MST he gained industrial experience by working on Cerner's Data Intelligence team for almost two years. During Darrien's time at MST, he has engaged in undergraduate research for MST's Computational Neuroscience lab, headed by Dr. Taylor. His primary research interests involve reinforcement learning, neural networks, and automation. Darrien will begin working for Sandia National Laboratories in May 2022 under The Mathematics and Analytics Research Technical Internship for Advanced National Security (MARTIANS) program. After he acquires his bachelors, Darrien intends to pursue a PhD in Computer Science to further engage in artificial intelligence research.
Research Category
Sciences
Presentation Type
Oral Presentation
Document Type
Presentation
Location
Missouri Room
Presentation Date
14 Apr 2022, 2:00 pm - 2:30 pm
Reinforcement Learning in Spiking Neural Networks
Missouri Room
Biological neurons communicate primarily via a spiking process. Recurrently connected spiking neural networks (RSNNs) more realistically model the brain, compared to their non-spiking counterparts. It is of great interest to discover a biologically realistic learning rule to achieve optimal levels of performance on machine learning tasks. Experimental data describe a phenomenon known as spike-timing-dependent-plasticity (STDP), which integrates local firing coincidences between neurons to learn. STDP is believed to underlie memory formation and storage within the brain. When a reward signal modulates STDP, it enables forming associative memories via operant conditioning. Neuromodulators like dopamine operate similarly in the brain. We employ processes like synaptic scaling to support R-STDP in large, unstructured RSNNs. Doing so produces an agent that achieves adequate performance on reinforcement learning tasks.
