Details
| Originalsprache | Englisch |
|---|---|
| Veröffentlichungsnummer (amtliches Aktenzeichen) | US2022012636 |
| IPC | G06N 20/ 00 A I |
| Prioritätsdatum | 10 Juli 2020 |
| Publikationsstatus | Veröffentlicht - 13 Jan. 2022 |
Abstract
Computer-implemented method for creating a system, which is suitable for creating in an automated manner a machine learning system for computer vision. The method includes: providing predefined hyperparameters; determining an optimal parameterization of the hyperparameters using BOHB (Bayesian optimization (BO) and Hyperband (HB)) for a plurality of different training data sets; assessing all optimal parameterizations on all training data sets of the plurality of different training data sets with the aid of a normalized metric; creating a matrix, the matrix including the evaluated normalized metric for each parameterization and for each training data set; determining meta-features for each of the training data sets; optimizing a decision tree, which outputs as a function of the meta-features and of the matrix which of the optimal parameterization using BOHB is a suitable parameterization for the given meta-features.
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
Patent Nr.: US2022012636. Jan. 13, 2022.
Publikation: Schutzrecht/Patent › Patent
}
TY - PAT
T1 - METHOD AND DEVICE FOR CREATING A SYSTEM FOR THE AUTOMATED CREATION OF MACHINE LEARNING SYSTEMS
AU - Lindauer, Marius
AU - Zela, Arber
AU - Stoll, Danny Oliver
AU - Ferreira, Fabio
AU - Hutter, Frank
AU - Nierhoff, Thomas
PY - 2022/1/13
Y1 - 2022/1/13
N2 - Computer-implemented method for creating a system, which is suitable for creating in an automated manner a machine learning system for computer vision. The method includes: providing predefined hyperparameters; determining an optimal parameterization of the hyperparameters using BOHB (Bayesian optimization (BO) and Hyperband (HB)) for a plurality of different training data sets; assessing all optimal parameterizations on all training data sets of the plurality of different training data sets with the aid of a normalized metric; creating a matrix, the matrix including the evaluated normalized metric for each parameterization and for each training data set; determining meta-features for each of the training data sets; optimizing a decision tree, which outputs as a function of the meta-features and of the matrix which of the optimal parameterization using BOHB is a suitable parameterization for the given meta-features.
AB - Computer-implemented method for creating a system, which is suitable for creating in an automated manner a machine learning system for computer vision. The method includes: providing predefined hyperparameters; determining an optimal parameterization of the hyperparameters using BOHB (Bayesian optimization (BO) and Hyperband (HB)) for a plurality of different training data sets; assessing all optimal parameterizations on all training data sets of the plurality of different training data sets with the aid of a normalized metric; creating a matrix, the matrix including the evaluated normalized metric for each parameterization and for each training data set; determining meta-features for each of the training data sets; optimizing a decision tree, which outputs as a function of the meta-features and of the matrix which of the optimal parameterization using BOHB is a suitable parameterization for the given meta-features.
M3 - Patent
M1 - US2022012636
ER -