Speaker: Nathaniel Lei Neligh, University of Tennessee Knoxville

Title: Re-Encoding With an Application to Human Memory

Abstract: We describe and characterize “re-encoding,” an informational tool which can improve the performance of noisy memory systems at the cost of introducing several quirks. Re-encoding is the process of periodically attempting to restore damaged or decayed information to its original state. This can help avoid the accumulation of errors which can ultimately lead to total information loss. We show that a form of re-encoding in the human mind could be responsible for several features of human memory including constructed memory, the poor correlation between memory accuracy and confidence, rehearsal effects, and specific spacing effects as well as the observed way that memories decay over time. We also consider the question of how to re-encode optimally.