Machine-learning training service for synthetic data
Various embodiments, methods and systems for implementing a distributed computing system machine-learning training service are provided. Initially a machine learning model is accessed. A plurality of synthetic data assets are accessed, where a synthetic data asset is associated with asset-variation parameters that are programmable for machine-learning. The machine learning model is retrained using the plurality of synthetic data assets. The machine-learning training service is further configured for executing real-time calls to generate an on-the-fly-generated synthetic data asset such that the on-the-fly-generated synthetic data asset is rendered in real-time to preclude pre-rendering and storing the on-the-fly-generated synthetic data asset. The machine-learning training service further supports hybrid-based machine learning training, where the machine learning model is trained based on a combination of the plurality of synthetic data assets, a plurality of non-synthetic data assets, and synthetic data asset metadata associated with the plurality of synthetic data assets.
1. A computer system comprising:
one or more hardware processors and memory configured to provide computer program instructions to the one or more hardware processors;
a machine-learning training service including a training feedback interface comprising interface elements for accessing and reviewing deficiencies identified in training datasets, the machine-learning training service is configured to utilize the one or more hardware processors to:
access a machine learning model;
access a plurality of synthetic data assets, wherein a synthetic data asset is associated with asset-variation parameters that are programmable for machine-learning; and
retrain the machine learning model using the plurality of synthetic data assets.
2. The system of claim 1 , wherein the machine-learning training service executes real-time calls to generate an on-the-fly-generated synthetic data asset for training the machine learning model.
3. The system of claim 2 , wherein the on-the-fly-generated synthetic data asset is rendered in real-time to preclude pre-rendering and storing the on-the-fly-generated synthetic data asset.
4. The system of claim 2 , wherein the machine-learning training service automatically discards the on-the-the on-the-fly-generated synthetic data upon completion of training based on the on-the-fly-generated synthetic data asset.
5. The system of claim 1 , wherein the plurality of synthetic data assets are generated to retrain the machine learning model to overcome a known bias associated with the machine learning model.
6. The system of claim 1 , wherein the machine-learning training service supports hybrid-based machine learning training, wherein the machine learning model is trained based on a combination of the plurality of synthetic data assets, a plurality of non-synthetic data assets; and synthetic data asset metadata associated with the plurality of synthetic data assets.
7. The system of claim 1 , wherein retraining the machine learning model using the plurality of synthetic data assets comprises identifying irrelevant asset-variation parameters; dropping the irrelevant asset-variation parameters; identifying new asset-variation parameters; and adding new asset-variation parameters.
8. One or more computer storage media storing computer-useable instructions that, when used by one or more computing devices, cause the one or more computing devices to perform operations comprising:
accessing a machine learning model associated with a machine-learning training service including a training feedback interface comprising interface elements for accessing and reviewing deficiencies identified in training datasets associated with the machine learning model;
accessing a plurality of synthetic data assets, wherein a synthetic data asset is associated with asset-variation parameters that are programmable for machine-learning; and
retraining the machine learning model using the plurality of synthetic data assets.
9. The media of claim 8 , further comprising executing real-time calls to generate an on-the-fly-generated synthetic data asset.
10. The media of claim 9 , wherein the on-the-fly-generated synthetic data asset is rendered in real-time to preclude pre-rendering and storing the on-the-fly-generated synthetic data asset.
11. The media of claim 8 , wherein the plurality of synthetic data assets are generated to retrain the machine learning model to overcome a known bias associated with the machine learning model.
12. The media of claim 8 , further comprising retraining the machine learning model using the plurality of synthetic data assets and a plurality of non-synthetic data assets for generating a hybrid-based machine learning model.
13. The media of claim 8 , wherein retraining the machine learning model using the plurality of synthetic data assets further comprises training the machine learning model using synthetic data asset metadata associated with the plurality of synthetic data assets.
14. The media of claim 8 , wherein retraining the machine learning model using the plurality of synthetic data assets comprises identifying irrelevant asset-variation parameters; dropping the irrelevant asset-variation parameters; identifying new asset-variation parameters; and adding new asset-variation parameters.
15. A method , the method comprising:
accessing a machine learning model associated with a machine-learning training service including a training feedback interface comprising interface elements for accessing and reviewing deficiencies identified in training datasets associated with the machine learning model;
accessing a plurality of synthetic data assets, wherein a synthetic data asset is associated with asset-variation parameters that are programmable for machine-learning; and
retraining the machine learning model using the plurality of synthetic data assets.
16. The method of claim 15 , further comprising executing real-time calls to generate an on-the-fly-generated synthetic data asset.
17. The method of claim 16 , wherein the on-the-fly-generated synthetic data asset is rendered in real-time to preclude pre-rendering and storing the on-the-fly-generated synthetic data asset.
18. The method of claim 15 , wherein the plurality of synthetic data assets are generated to retrain the machine learning model to overcome a known bias associated with the machine learning model.
19. The method of claim 15 , wherein retraining the machine learning model using the plurality of synthetic data assets further comprises:
using the plurality of synthetic data assets and a plurality of non-synthetic data assets for generating a hybrid-based machine learning model; and
training the machine learning model using synthetic data asset metadata associated with the plurality of synthetic data assets.
20. The method of claim 15 , wherein retraining the machine learning model using the plurality of synthetic data assets comprises identifying irrelevant asset-variation parameters; dropping the irrelevant asset-variation parameters; identifying new asset-variation parameters; and adding new asset-variation parameters.