Vertaling van "trained machine learning models" in Japans
We konden deze vermelding niet vinden. Er worden benaderende resultaten weergegeven. Controleer je spelling of stel voor deze term aan het woordenboek toe te voegen.
訓練済み機械学習モデル
This monitoring is divided into some patterns in accordance with the method of collecting correct answers compared with inference results of trained machine learning models required for calculating the accuracy.
In cases where it is found that trained machine learning models do not satisfy the specifications after going through training, the process may return to this step to adjust the design of datasets.
As for "secrecy", it is pointed out in academic research results that a part of training data can be estimated based on the content of trained machine learning models under certain conditions, and this could pose a future challenge from the viewpoint of protection of personal information and privacy.
Moreover, technologies to retrain machine learning models which play the core role of updating trained machine learning models and technologies to create additional training data used therein will be presented.
Therefore, internal qualities 1) can sufficiently respond to changes in external environments and 2) prevent the quality from deteriorating due to changes in trained machine learning models made for such response.
Moreover, a literature [Lindstrom 2011] proposes a method of detecting data changes by using confidence scores associated with output from trained machine learning models without using labels (CDBD).
Overfitting Trained machine learning models are adapted excessively to training data and it becomes impossible for them to return desired output in response to input data other than training data.
This concept means that internal qualities can fully respond to changes in operational environments outside the system and that any change in trained machine learning models do not cause unnecessary deterioration of quality.
Compared to the standard training method, it mat be able to improve the resistance of trained machine learning models against adversarial examples.
通常の訓練と比較して、機械学習モデルの敵対的データに対する耐性を向上させる ことができる。
7.6 (skipped) The technologies with regard to the stability of trained machine learning models are described in Section 7.5 together with those regarding to the accuracy.
It includes software scripts needed to convert TensorFlow files and accelerator FPGA IP to run trained machine learning models as well as tutorials and examples.
7.5 Correctness and stability of trained machine learning models Checking the correctness and stability of trained machine learning models corresponds to "unit test" of software testing in the conventional software development.
Er worden alleen voorbeelden gegeven om u te helpen het woord of de woordcombinatie waarop u hebt gezocht, te vertalen. Deze worden niet door ons geselecteerd of gevalideerd en kunnen ongepaste taal bevatten. Wij vragen u melding te maken van voorbeelden die dienen te worden aangepast of verwijderd. Vertalingen met grof of informeel taalgebruik worden meestal rood of oranje gemarkeerd.