---

Telfeyan John
James Madison University

Subject Listing - Physics/Astronomy
Advisor: Dr. Niculescu Gabriel

Thursday, Oral Session 1, Presentation 4, Robinson Hall 217

USING NEURAL NETWORKS TO MODEL ELECTRON-NUCLEON INTERACTIONS

The Thomas Jefferson National Accelerator Facility (Jefferson Lab.), in Newport News, Virginia, has an extensive program of studying inclusive electron-nucleon interactions in the few GeV energy range. In this work we used multi-layer perceptrons, a particular form of artificial neural networks to analyze electron-proton and electron-deuteron inclusive scattering data. Even though neural networks simulate the activity of only a few neurons in the brain, they are powerful mathematical tools that are ideally suited to tackle complex, non-linear problems. With a set of test data, a user can train a neural network to find patterns, thus providing a representation of the data. This talk will outline the theory behind neural networks and highlight how we applied this technique in particle physics research.

Advisor: Dr. Niculescu Gabriel, Assistant Professor, Department of Physics, James Madison University, Harrisonburg, VA