Methods and systems for training a machine learning method using high quality ground truth approximations
AssigneeAptiv Technologies AG
InventorsJittu Kurian, Jan Siegemund, Mirko Meuter
Life of this patent
Estimating the term…
Abstract
A computer-implemented method for training a machine-learning method comprises the following steps carried out by computer hardware components: determining measurement data from a first sensor; determining approximations of ground truths based on a second sensor; and training the machine-learning method based on the measurement data and the approximations of ground truths; wherein approximations of ground truths of lower-approximation quality have a lower effect on the training than approximations of ground truths of higher-approximation quality.
Claims20
A computer-implemented method for training a machine-learning model comprising:
- determining measurement data from a first sensor;
- determining approximations of ground truths based on data from a second sensor;
- and
- training the machine-learning model based on the measurement data and the approximations of ground truths, wherein a first group of the approximations of ground truths are of lower-approximation quality, the training including training the machine-learning model to a first extent based on the first group of approximations of ground truths of the lower-approximation quality and training the machine-learning model to a second extent based on a second group of the approximations of ground truths that are of higher-approximation quality than the first group, wherein the second extent is greater than the first extent.
A non-transitory computer-readable medium comprising instructions that, when executed, cause a processor of a computer system to:
- determine measurement data from a first sensor;
- determine approximations of ground truths based on data from a second sensor;
- and
- train a machine-learning model based on the measurement data and the approximations of ground truths, wherein a first group of the approximations of ground truths are of lower-approximation quality, the training of the machine-learning model including training the machine-learning model to a first extent based on the first group of approximations of ground truths of the lower-approximation quality and training the machine-learning model to a second extent based on a second group of the approximations of ground truths that are of higher-approximation quality than the first group, wherein the second extent is greater than the first extent.
A computer system configured to:
- determine measurement data from a first sensor;
- determine approximations of ground truths based on data from a second sensor;
- and
- train a machine-learning model based on the measurement data and the approximations of ground truths, wherein a first group of the approximations of ground truths are of lower-approximation quality, the training of the machine-learning model including training the machine-learning model to a first extent based on the first group of approximations of ground truths of the lower-approximation quality and training the machine-learning model to a second extent based on a second group of the approximations of ground truths that are of higher-approximation quality than the first group, wherein the second extent is greater than the first extent.
Dependencies are inferred from the claim text.