TY - JOUR
T1 - The Eighty Five Percent Rule for optimal learning
AU - Wilson, Robert C.
AU - Shenhav, Amitai
AU - Straccia, Mark
AU - Cohen, Jonathan D.
N1 - Funding Information:
This project was made possible through the support of a grant from the John Templeton Foundation to J.D.C., a Center of Biomedical Research Excellence grant P20GM103645 from the National Institute of General Medical Sciences to A.S., and National Institute on Aging grant R56 AG061888 to R.C.W. The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of the funders.
Publisher Copyright:
© 2019, The Author(s).
PY - 2019/12/1
Y1 - 2019/12/1
N2 - Researchers and educators have long wrestled with the question of how best to teach their clients be they humans, non-human animals or machines. Here, we examine the role of a single variable, the difficulty of training, on the rate of learning. In many situations we find that there is a sweet spot in which training is neither too easy nor too hard, and where learning progresses most quickly. We derive conditions for this sweet spot for a broad class of learning algorithms in the context of binary classification tasks. For all of these stochastic gradient-descent based learning algorithms, we find that the optimal error rate for training is around 15.87% or, conversely, that the optimal training accuracy is about 85%. We demonstrate the efficacy of this ‘Eighty Five Percent Rule’ for artificial neural networks used in AI and biologically plausible neural networks thought to describe animal learning.
AB - Researchers and educators have long wrestled with the question of how best to teach their clients be they humans, non-human animals or machines. Here, we examine the role of a single variable, the difficulty of training, on the rate of learning. In many situations we find that there is a sweet spot in which training is neither too easy nor too hard, and where learning progresses most quickly. We derive conditions for this sweet spot for a broad class of learning algorithms in the context of binary classification tasks. For all of these stochastic gradient-descent based learning algorithms, we find that the optimal error rate for training is around 15.87% or, conversely, that the optimal training accuracy is about 85%. We demonstrate the efficacy of this ‘Eighty Five Percent Rule’ for artificial neural networks used in AI and biologically plausible neural networks thought to describe animal learning.
UR - http://www.scopus.com/inward/record.url?scp=85074812807&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85074812807&partnerID=8YFLogxK
U2 - 10.1038/s41467-019-12552-4
DO - 10.1038/s41467-019-12552-4
M3 - Article
C2 - 31690723
AN - SCOPUS:85074812807
SN - 2041-1723
VL - 10
JO - Nature communications
JF - Nature communications
IS - 1
M1 - 4646
ER -