TY  - CPAPER
U1  - Konferenzveröffentlichung
A1  - Merker, Jochen
A1  - Schuldt, Gregor
T1  - An attempt to explain double descent in modern machine learning
T2  - 26. Interdisziplinäre Wissenschaftliche Konferenz Mittweida
N2  - This article aims to explain mathematically, why the so called double descent observed by Belkin et al., Reconciling modern machine-learning practice and the classical bias-variance trade-off, PNAS 116(32) (2019), p. 15849-15854, occurs on the way from the classical approximation regime of machine learning to the modern interpolation regime. We argue that this phenomenon may be explained by a decomposition of mean squared error plus complexity into bias, variance and an unavoidable irreducible error inherent to the problem. Further, in case of normally distributed output errors, we apply this decomposition to explain, why LASSO provides reliable predictors avoiding overfitting.
KW  - Maschinelles Lernen
KW  - Komplexität
KW  - bias-variance
KW  - double descent
Y1  - 2021
UN  - https://nbn-resolving.org/urn:nbn:de:bsz:mit1-opus4-122935
SN  - 1437-7624
SS  - 1437-7624
U6  - https://doi.org/10.48446/opus-12293
DO  - https://doi.org/10.48446/opus-12293
IS  - 002
SP  - 141
EP  - 144
S1  - 4
PB  - Hochschule Mittweida
CY  - Mittweida
ER  -